From theday79 at gmail.com Tue Aug 16 00:04:43 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 00:04:43 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: Message-ID: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in "cufft.h". It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang -------------- next part -------------- An HTML attachment was scrubbed... URL: From andreasg at phys.au.dk Tue Aug 16 04:21:04 2016 From: andreasg at phys.au.dk (Andreas Gravgaard Andersen) Date: Tue, 16 Aug 2016 10:21:04 +0200 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" wrote: > Hi all, > > > > I used to use RTK r1.03 without any problem. > > However, today I adapted rtkfdk to my own code and ran into the following > compile error messages.. > > > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan2d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan3d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecR2C referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecC2R referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftSetCompatibilityMode > referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const > &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@ > @YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftDestroy referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > > > I found that this error was caused by the following line in my code: > > > > #if CUDA_FOUND > > feldkamp = FDKCUDAType::New(); ? here! > > SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer > ())); > > #else > > .. > > I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, > cufftExecR2C, ..) are defined in ?cufft.h?. > > > > It is very weird to me because I did the exactly same thing in my previous > project and there was no problem at that time. > > Maybe I did something wrong but have no idea now. > > > > Any helps/suggestions will be highly appreciated! > > > > Regards, > > Yang > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cmake_mod.zip Type: application/zip Size: 7846 bytes Desc: not available URL: From theday79 at gmail.com Tue Aug 16 08:34:15 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 08:34:15 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: <001801d1f7ba$80901900$81b04b00$@gmail.com> Hi Andreas, Thanks a lot for your help! Now I can recall what happened in my previous project. I should have linked ?cufft.lib? manually in the CMakeLists file. Now it works fine. I believe this issue happens only in the old version of RTK and may have been fixed in the Nightly version. (I have been too lazy to move on to a newer version, though.) Thanks. Yang (To RTK staff: I?m sorry that I forgot to change the title even though this is a separate thread.) From: andreasga22 at gmail.com [mailto:andreasga22 at gmail.com] On Behalf Of Andreas Gravgaard Andersen Sent: Tuesday, August 16, 2016 4:21 AM To: Yang-Kyun Park Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Geometric phantom example on wiki Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" > wrote: Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in ?cufft.h?. It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Fri Aug 19 02:59:38 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Fri, 19 Aug 2016 06:59:38 +0000 Subject: [Rtk-users] Contribute to RTK? Message-ID: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics -------------- next part -------------- An HTML attachment was scrubbed... URL: From K.Niepel at physik.uni-muenchen.de Tue Aug 23 10:43:06 2016 From: K.Niepel at physik.uni-muenchen.de (K.Niepel) Date: Tue, 23 Aug 2016 16:43:06 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk Message-ID: <57BC60FA.40600@physik.uni-muenchen.de> Hello there, I am currently trying out the rtkfourdfdk application on a moving phantom data set with the number of phase bins varying from 15 down to one. However, there is no noticeable blurring for lower fewer bins as with the 3D rtkfdk app - actually, the first frame seems to be exactly the same for any total number of frames. The only setting i changed is the --frames argument. Am I missing something? Thanks in advance, Kati // -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 05:04:13 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 11:04:13 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk In-Reply-To: <57BC60FA.40600@physik.uni-muenchen.de> References: <57BC60FA.40600@physik.uni-muenchen.de> Message-ID: Hi, Quickly checking the code, the reason is pretty simple: rtkfourdfdk uses rtkSelectOneProjectionPerCycleImageFilter.h, which selects the closest projection to a given phase in each respiratory cycle. So setting the number of frames will not increase the number of frames per phase (and, therefore, blur). I think it's quite a common way of sorting in the literature, see for example [Sonke et al, Med Phys, 2005]: "For each breathing cycle, the projection closest to each breathing phase is selected." The reason is explained in [Dietrich et al, PMB, 2006] or in a simulation work I did in the past. Simon On Tue, Aug 23, 2016 at 4:43 PM, K.Niepel wrote: > Hello there, > > I am currently trying out the rtkfourdfdk application on a moving phantom > data set with the number of phase bins varying from 15 down to one. > However, there is no noticeable blurring for lower fewer bins as with the > 3D rtkfdk app - actually, the first frame seems to be exactly the same for > any total number of frames. > > The only setting i changed is the --frames argument. Am I missing > something? > > Thanks in advance, > Kati > > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 09:58:11 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 15:58:11 +0200 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: Message-ID: Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page ) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh < andy.shieh at sydney.edu.au> wrote: > Hi Simon, > > > > First of all, thank you and all the contributors for such a great toolkit. > I've been a RTK user for many years and it has been extremely helpful to my > research. > > > > I've been developing my programs using RTK for quite a while but haven't > really been active in this community. > > I have implemented the PICCS method in RTK (still needs to be tidied up a > bit though) and would like to contribute to the repository if it's not yet > available. > > It is probably not developed with the best computation efficiency, but > hopefully would still be useful to users who wish to test out the PICCS > method. > > > > If you think this could be useful to RTK, could you please let me know > what would be the best way to contribute my codes to RTK? > > > > Many thanks, > > Andy > > > > *DR CHUN-CHIEN (ANDY) SHIEH* | Research Associate > > Radiation Physics Laboratory | Sydney Medical School > > > > *THE UNIVERSITY OF SYDNEY* > Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 > * T* +61 2 9036 5012 > *E* *andy.shieh at sydney.edu.au * | *W* > *http://sydney.edu.au/medicine/radiation-physics* > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Wed Aug 24 17:58:28 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Wed, 24 Aug 2016 21:58:28 +0000 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: , Message-ID: <7587051e-cbb3-4865-99de-9b6ffe5bf201@sydney.edu.au> Thanks Simon. I will have a look and start working on it! Cheers, Andy Sent from Nine ________________________________ From: Simon Rit Sent: 24 Aug 2016 23:58 To: Chun-Chien (Andy) Shieh Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Contribute to RTK? Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh > wrote: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From theday79 at gmail.com Tue Aug 16 00:04:43 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 00:04:43 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: Message-ID: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in "cufft.h". It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang -------------- next part -------------- An HTML attachment was scrubbed... URL: From andreasg at phys.au.dk Tue Aug 16 04:21:04 2016 From: andreasg at phys.au.dk (Andreas Gravgaard Andersen) Date: Tue, 16 Aug 2016 10:21:04 +0200 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" wrote: > Hi all, > > > > I used to use RTK r1.03 without any problem. > > However, today I adapted rtkfdk to my own code and ran into the following > compile error messages.. > > > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan2d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan3d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecR2C referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecC2R referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftSetCompatibilityMode > referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const > &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@ > @YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftDestroy referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > > > I found that this error was caused by the following line in my code: > > > > #if CUDA_FOUND > > feldkamp = FDKCUDAType::New(); ? here! > > SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer > ())); > > #else > > .. > > I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, > cufftExecR2C, ..) are defined in ?cufft.h?. > > > > It is very weird to me because I did the exactly same thing in my previous > project and there was no problem at that time. > > Maybe I did something wrong but have no idea now. > > > > Any helps/suggestions will be highly appreciated! > > > > Regards, > > Yang > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cmake_mod.zip Type: application/zip Size: 7846 bytes Desc: not available URL: From theday79 at gmail.com Tue Aug 16 08:34:15 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 08:34:15 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: <001801d1f7ba$80901900$81b04b00$@gmail.com> Hi Andreas, Thanks a lot for your help! Now I can recall what happened in my previous project. I should have linked ?cufft.lib? manually in the CMakeLists file. Now it works fine. I believe this issue happens only in the old version of RTK and may have been fixed in the Nightly version. (I have been too lazy to move on to a newer version, though.) Thanks. Yang (To RTK staff: I?m sorry that I forgot to change the title even though this is a separate thread.) From: andreasga22 at gmail.com [mailto:andreasga22 at gmail.com] On Behalf Of Andreas Gravgaard Andersen Sent: Tuesday, August 16, 2016 4:21 AM To: Yang-Kyun Park Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Geometric phantom example on wiki Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" > wrote: Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in ?cufft.h?. It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Fri Aug 19 02:59:38 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Fri, 19 Aug 2016 06:59:38 +0000 Subject: [Rtk-users] Contribute to RTK? Message-ID: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics -------------- next part -------------- An HTML attachment was scrubbed... URL: From K.Niepel at physik.uni-muenchen.de Tue Aug 23 10:43:06 2016 From: K.Niepel at physik.uni-muenchen.de (K.Niepel) Date: Tue, 23 Aug 2016 16:43:06 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk Message-ID: <57BC60FA.40600@physik.uni-muenchen.de> Hello there, I am currently trying out the rtkfourdfdk application on a moving phantom data set with the number of phase bins varying from 15 down to one. However, there is no noticeable blurring for lower fewer bins as with the 3D rtkfdk app - actually, the first frame seems to be exactly the same for any total number of frames. The only setting i changed is the --frames argument. Am I missing something? Thanks in advance, Kati // -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 05:04:13 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 11:04:13 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk In-Reply-To: <57BC60FA.40600@physik.uni-muenchen.de> References: <57BC60FA.40600@physik.uni-muenchen.de> Message-ID: Hi, Quickly checking the code, the reason is pretty simple: rtkfourdfdk uses rtkSelectOneProjectionPerCycleImageFilter.h, which selects the closest projection to a given phase in each respiratory cycle. So setting the number of frames will not increase the number of frames per phase (and, therefore, blur). I think it's quite a common way of sorting in the literature, see for example [Sonke et al, Med Phys, 2005]: "For each breathing cycle, the projection closest to each breathing phase is selected." The reason is explained in [Dietrich et al, PMB, 2006] or in a simulation work I did in the past. Simon On Tue, Aug 23, 2016 at 4:43 PM, K.Niepel wrote: > Hello there, > > I am currently trying out the rtkfourdfdk application on a moving phantom > data set with the number of phase bins varying from 15 down to one. > However, there is no noticeable blurring for lower fewer bins as with the > 3D rtkfdk app - actually, the first frame seems to be exactly the same for > any total number of frames. > > The only setting i changed is the --frames argument. Am I missing > something? > > Thanks in advance, > Kati > > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 09:58:11 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 15:58:11 +0200 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: Message-ID: Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page ) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh < andy.shieh at sydney.edu.au> wrote: > Hi Simon, > > > > First of all, thank you and all the contributors for such a great toolkit. > I've been a RTK user for many years and it has been extremely helpful to my > research. > > > > I've been developing my programs using RTK for quite a while but haven't > really been active in this community. > > I have implemented the PICCS method in RTK (still needs to be tidied up a > bit though) and would like to contribute to the repository if it's not yet > available. > > It is probably not developed with the best computation efficiency, but > hopefully would still be useful to users who wish to test out the PICCS > method. > > > > If you think this could be useful to RTK, could you please let me know > what would be the best way to contribute my codes to RTK? > > > > Many thanks, > > Andy > > > > *DR CHUN-CHIEN (ANDY) SHIEH* | Research Associate > > Radiation Physics Laboratory | Sydney Medical School > > > > *THE UNIVERSITY OF SYDNEY* > Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 > * T* +61 2 9036 5012 > *E* *andy.shieh at sydney.edu.au * | *W* > *http://sydney.edu.au/medicine/radiation-physics* > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Wed Aug 24 17:58:28 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Wed, 24 Aug 2016 21:58:28 +0000 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: , Message-ID: <7587051e-cbb3-4865-99de-9b6ffe5bf201@sydney.edu.au> Thanks Simon. I will have a look and start working on it! Cheers, Andy Sent from Nine ________________________________ From: Simon Rit Sent: 24 Aug 2016 23:58 To: Chun-Chien (Andy) Shieh Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Contribute to RTK? Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh > wrote: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From theday79 at gmail.com Tue Aug 16 00:04:43 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 00:04:43 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: Message-ID: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in "cufft.h". It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang -------------- next part -------------- An HTML attachment was scrubbed... URL: From andreasg at phys.au.dk Tue Aug 16 04:21:04 2016 From: andreasg at phys.au.dk (Andreas Gravgaard Andersen) Date: Tue, 16 Aug 2016 10:21:04 +0200 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" wrote: > Hi all, > > > > I used to use RTK r1.03 without any problem. > > However, today I adapted rtkfdk to my own code and ran into the following > compile error messages.. > > > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan2d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan3d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecR2C referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecC2R referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftSetCompatibilityMode > referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const > &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@ > @YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftDestroy referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > > > I found that this error was caused by the following line in my code: > > > > #if CUDA_FOUND > > feldkamp = FDKCUDAType::New(); ? here! > > SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer > ())); > > #else > > .. > > I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, > cufftExecR2C, ..) are defined in ?cufft.h?. > > > > It is very weird to me because I did the exactly same thing in my previous > project and there was no problem at that time. > > Maybe I did something wrong but have no idea now. > > > > Any helps/suggestions will be highly appreciated! > > > > Regards, > > Yang > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cmake_mod.zip Type: application/zip Size: 7846 bytes Desc: not available URL: From theday79 at gmail.com Tue Aug 16 08:34:15 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 08:34:15 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: <001801d1f7ba$80901900$81b04b00$@gmail.com> Hi Andreas, Thanks a lot for your help! Now I can recall what happened in my previous project. I should have linked ?cufft.lib? manually in the CMakeLists file. Now it works fine. I believe this issue happens only in the old version of RTK and may have been fixed in the Nightly version. (I have been too lazy to move on to a newer version, though.) Thanks. Yang (To RTK staff: I?m sorry that I forgot to change the title even though this is a separate thread.) From: andreasga22 at gmail.com [mailto:andreasga22 at gmail.com] On Behalf Of Andreas Gravgaard Andersen Sent: Tuesday, August 16, 2016 4:21 AM To: Yang-Kyun Park Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Geometric phantom example on wiki Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" > wrote: Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in ?cufft.h?. It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Fri Aug 19 02:59:38 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Fri, 19 Aug 2016 06:59:38 +0000 Subject: [Rtk-users] Contribute to RTK? Message-ID: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics -------------- next part -------------- An HTML attachment was scrubbed... URL: From K.Niepel at physik.uni-muenchen.de Tue Aug 23 10:43:06 2016 From: K.Niepel at physik.uni-muenchen.de (K.Niepel) Date: Tue, 23 Aug 2016 16:43:06 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk Message-ID: <57BC60FA.40600@physik.uni-muenchen.de> Hello there, I am currently trying out the rtkfourdfdk application on a moving phantom data set with the number of phase bins varying from 15 down to one. However, there is no noticeable blurring for lower fewer bins as with the 3D rtkfdk app - actually, the first frame seems to be exactly the same for any total number of frames. The only setting i changed is the --frames argument. Am I missing something? Thanks in advance, Kati // -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 05:04:13 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 11:04:13 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk In-Reply-To: <57BC60FA.40600@physik.uni-muenchen.de> References: <57BC60FA.40600@physik.uni-muenchen.de> Message-ID: Hi, Quickly checking the code, the reason is pretty simple: rtkfourdfdk uses rtkSelectOneProjectionPerCycleImageFilter.h, which selects the closest projection to a given phase in each respiratory cycle. So setting the number of frames will not increase the number of frames per phase (and, therefore, blur). I think it's quite a common way of sorting in the literature, see for example [Sonke et al, Med Phys, 2005]: "For each breathing cycle, the projection closest to each breathing phase is selected." The reason is explained in [Dietrich et al, PMB, 2006] or in a simulation work I did in the past. Simon On Tue, Aug 23, 2016 at 4:43 PM, K.Niepel wrote: > Hello there, > > I am currently trying out the rtkfourdfdk application on a moving phantom > data set with the number of phase bins varying from 15 down to one. > However, there is no noticeable blurring for lower fewer bins as with the > 3D rtkfdk app - actually, the first frame seems to be exactly the same for > any total number of frames. > > The only setting i changed is the --frames argument. Am I missing > something? > > Thanks in advance, > Kati > > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 09:58:11 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 15:58:11 +0200 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: Message-ID: Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page ) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh < andy.shieh at sydney.edu.au> wrote: > Hi Simon, > > > > First of all, thank you and all the contributors for such a great toolkit. > I've been a RTK user for many years and it has been extremely helpful to my > research. > > > > I've been developing my programs using RTK for quite a while but haven't > really been active in this community. > > I have implemented the PICCS method in RTK (still needs to be tidied up a > bit though) and would like to contribute to the repository if it's not yet > available. > > It is probably not developed with the best computation efficiency, but > hopefully would still be useful to users who wish to test out the PICCS > method. > > > > If you think this could be useful to RTK, could you please let me know > what would be the best way to contribute my codes to RTK? > > > > Many thanks, > > Andy > > > > *DR CHUN-CHIEN (ANDY) SHIEH* | Research Associate > > Radiation Physics Laboratory | Sydney Medical School > > > > *THE UNIVERSITY OF SYDNEY* > Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 > * T* +61 2 9036 5012 > *E* *andy.shieh at sydney.edu.au * | *W* > *http://sydney.edu.au/medicine/radiation-physics* > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Wed Aug 24 17:58:28 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Wed, 24 Aug 2016 21:58:28 +0000 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: , Message-ID: <7587051e-cbb3-4865-99de-9b6ffe5bf201@sydney.edu.au> Thanks Simon. I will have a look and start working on it! Cheers, Andy Sent from Nine ________________________________ From: Simon Rit Sent: 24 Aug 2016 23:58 To: Chun-Chien (Andy) Shieh Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Contribute to RTK? Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh > wrote: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From theday79 at gmail.com Tue Aug 16 00:04:43 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 00:04:43 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: Message-ID: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in "cufft.h". It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang -------------- next part -------------- An HTML attachment was scrubbed... URL: From andreasg at phys.au.dk Tue Aug 16 04:21:04 2016 From: andreasg at phys.au.dk (Andreas Gravgaard Andersen) Date: Tue, 16 Aug 2016 10:21:04 +0200 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" wrote: > Hi all, > > > > I used to use RTK r1.03 without any problem. > > However, today I adapted rtkfdk to my own code and ran into the following > compile error messages.. > > > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan2d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan3d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecR2C referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecC2R referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftSetCompatibilityMode > referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const > &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@ > @YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftDestroy referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > > > I found that this error was caused by the following line in my code: > > > > #if CUDA_FOUND > > feldkamp = FDKCUDAType::New(); ? here! > > SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer > ())); > > #else > > .. > > I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, > cufftExecR2C, ..) are defined in ?cufft.h?. > > > > It is very weird to me because I did the exactly same thing in my previous > project and there was no problem at that time. > > Maybe I did something wrong but have no idea now. > > > > Any helps/suggestions will be highly appreciated! > > > > Regards, > > Yang > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cmake_mod.zip Type: application/zip Size: 7846 bytes Desc: not available URL: From theday79 at gmail.com Tue Aug 16 08:34:15 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 08:34:15 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: <001801d1f7ba$80901900$81b04b00$@gmail.com> Hi Andreas, Thanks a lot for your help! Now I can recall what happened in my previous project. I should have linked ?cufft.lib? manually in the CMakeLists file. Now it works fine. I believe this issue happens only in the old version of RTK and may have been fixed in the Nightly version. (I have been too lazy to move on to a newer version, though.) Thanks. Yang (To RTK staff: I?m sorry that I forgot to change the title even though this is a separate thread.) From: andreasga22 at gmail.com [mailto:andreasga22 at gmail.com] On Behalf Of Andreas Gravgaard Andersen Sent: Tuesday, August 16, 2016 4:21 AM To: Yang-Kyun Park Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Geometric phantom example on wiki Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" > wrote: Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in ?cufft.h?. It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Fri Aug 19 02:59:38 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Fri, 19 Aug 2016 06:59:38 +0000 Subject: [Rtk-users] Contribute to RTK? Message-ID: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics -------------- next part -------------- An HTML attachment was scrubbed... URL: From K.Niepel at physik.uni-muenchen.de Tue Aug 23 10:43:06 2016 From: K.Niepel at physik.uni-muenchen.de (K.Niepel) Date: Tue, 23 Aug 2016 16:43:06 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk Message-ID: <57BC60FA.40600@physik.uni-muenchen.de> Hello there, I am currently trying out the rtkfourdfdk application on a moving phantom data set with the number of phase bins varying from 15 down to one. However, there is no noticeable blurring for lower fewer bins as with the 3D rtkfdk app - actually, the first frame seems to be exactly the same for any total number of frames. The only setting i changed is the --frames argument. Am I missing something? Thanks in advance, Kati // -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 05:04:13 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 11:04:13 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk In-Reply-To: <57BC60FA.40600@physik.uni-muenchen.de> References: <57BC60FA.40600@physik.uni-muenchen.de> Message-ID: Hi, Quickly checking the code, the reason is pretty simple: rtkfourdfdk uses rtkSelectOneProjectionPerCycleImageFilter.h, which selects the closest projection to a given phase in each respiratory cycle. So setting the number of frames will not increase the number of frames per phase (and, therefore, blur). I think it's quite a common way of sorting in the literature, see for example [Sonke et al, Med Phys, 2005]: "For each breathing cycle, the projection closest to each breathing phase is selected." The reason is explained in [Dietrich et al, PMB, 2006] or in a simulation work I did in the past. Simon On Tue, Aug 23, 2016 at 4:43 PM, K.Niepel wrote: > Hello there, > > I am currently trying out the rtkfourdfdk application on a moving phantom > data set with the number of phase bins varying from 15 down to one. > However, there is no noticeable blurring for lower fewer bins as with the > 3D rtkfdk app - actually, the first frame seems to be exactly the same for > any total number of frames. > > The only setting i changed is the --frames argument. Am I missing > something? > > Thanks in advance, > Kati > > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 09:58:11 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 15:58:11 +0200 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: Message-ID: Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page ) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh < andy.shieh at sydney.edu.au> wrote: > Hi Simon, > > > > First of all, thank you and all the contributors for such a great toolkit. > I've been a RTK user for many years and it has been extremely helpful to my > research. > > > > I've been developing my programs using RTK for quite a while but haven't > really been active in this community. > > I have implemented the PICCS method in RTK (still needs to be tidied up a > bit though) and would like to contribute to the repository if it's not yet > available. > > It is probably not developed with the best computation efficiency, but > hopefully would still be useful to users who wish to test out the PICCS > method. > > > > If you think this could be useful to RTK, could you please let me know > what would be the best way to contribute my codes to RTK? > > > > Many thanks, > > Andy > > > > *DR CHUN-CHIEN (ANDY) SHIEH* | Research Associate > > Radiation Physics Laboratory | Sydney Medical School > > > > *THE UNIVERSITY OF SYDNEY* > Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 > * T* +61 2 9036 5012 > *E* *andy.shieh at sydney.edu.au * | *W* > *http://sydney.edu.au/medicine/radiation-physics* > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Wed Aug 24 17:58:28 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Wed, 24 Aug 2016 21:58:28 +0000 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: , Message-ID: <7587051e-cbb3-4865-99de-9b6ffe5bf201@sydney.edu.au> Thanks Simon. I will have a look and start working on it! Cheers, Andy Sent from Nine ________________________________ From: Simon Rit Sent: 24 Aug 2016 23:58 To: Chun-Chien (Andy) Shieh Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Contribute to RTK? Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh > wrote: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From theday79 at gmail.com Tue Aug 16 00:04:43 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 00:04:43 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: Message-ID: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in "cufft.h". It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang -------------- next part -------------- An HTML attachment was scrubbed... URL: From andreasg at phys.au.dk Tue Aug 16 04:21:04 2016 From: andreasg at phys.au.dk (Andreas Gravgaard Andersen) Date: Tue, 16 Aug 2016 10:21:04 +0200 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" wrote: > Hi all, > > > > I used to use RTK r1.03 without any problem. > > However, today I adapted rtkfdk to my own code and ran into the following > compile error messages.. > > > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan2d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan3d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecR2C referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecC2R referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftSetCompatibilityMode > referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const > &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@ > @YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftDestroy referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > > > I found that this error was caused by the following line in my code: > > > > #if CUDA_FOUND > > feldkamp = FDKCUDAType::New(); ? here! > > SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer > ())); > > #else > > .. > > I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, > cufftExecR2C, ..) are defined in ?cufft.h?. > > > > It is very weird to me because I did the exactly same thing in my previous > project and there was no problem at that time. > > Maybe I did something wrong but have no idea now. > > > > Any helps/suggestions will be highly appreciated! > > > > Regards, > > Yang > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cmake_mod.zip Type: application/zip Size: 7846 bytes Desc: not available URL: From theday79 at gmail.com Tue Aug 16 08:34:15 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 08:34:15 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: <001801d1f7ba$80901900$81b04b00$@gmail.com> Hi Andreas, Thanks a lot for your help! Now I can recall what happened in my previous project. I should have linked ?cufft.lib? manually in the CMakeLists file. Now it works fine. I believe this issue happens only in the old version of RTK and may have been fixed in the Nightly version. (I have been too lazy to move on to a newer version, though.) Thanks. Yang (To RTK staff: I?m sorry that I forgot to change the title even though this is a separate thread.) From: andreasga22 at gmail.com [mailto:andreasga22 at gmail.com] On Behalf Of Andreas Gravgaard Andersen Sent: Tuesday, August 16, 2016 4:21 AM To: Yang-Kyun Park Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Geometric phantom example on wiki Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" > wrote: Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in ?cufft.h?. It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Fri Aug 19 02:59:38 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Fri, 19 Aug 2016 06:59:38 +0000 Subject: [Rtk-users] Contribute to RTK? Message-ID: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics -------------- next part -------------- An HTML attachment was scrubbed... URL: From K.Niepel at physik.uni-muenchen.de Tue Aug 23 10:43:06 2016 From: K.Niepel at physik.uni-muenchen.de (K.Niepel) Date: Tue, 23 Aug 2016 16:43:06 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk Message-ID: <57BC60FA.40600@physik.uni-muenchen.de> Hello there, I am currently trying out the rtkfourdfdk application on a moving phantom data set with the number of phase bins varying from 15 down to one. However, there is no noticeable blurring for lower fewer bins as with the 3D rtkfdk app - actually, the first frame seems to be exactly the same for any total number of frames. The only setting i changed is the --frames argument. Am I missing something? Thanks in advance, Kati // -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 05:04:13 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 11:04:13 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk In-Reply-To: <57BC60FA.40600@physik.uni-muenchen.de> References: <57BC60FA.40600@physik.uni-muenchen.de> Message-ID: Hi, Quickly checking the code, the reason is pretty simple: rtkfourdfdk uses rtkSelectOneProjectionPerCycleImageFilter.h, which selects the closest projection to a given phase in each respiratory cycle. So setting the number of frames will not increase the number of frames per phase (and, therefore, blur). I think it's quite a common way of sorting in the literature, see for example [Sonke et al, Med Phys, 2005]: "For each breathing cycle, the projection closest to each breathing phase is selected." The reason is explained in [Dietrich et al, PMB, 2006] or in a simulation work I did in the past. Simon On Tue, Aug 23, 2016 at 4:43 PM, K.Niepel wrote: > Hello there, > > I am currently trying out the rtkfourdfdk application on a moving phantom > data set with the number of phase bins varying from 15 down to one. > However, there is no noticeable blurring for lower fewer bins as with the > 3D rtkfdk app - actually, the first frame seems to be exactly the same for > any total number of frames. > > The only setting i changed is the --frames argument. Am I missing > something? > > Thanks in advance, > Kati > > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 09:58:11 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 15:58:11 +0200 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: Message-ID: Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page ) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh < andy.shieh at sydney.edu.au> wrote: > Hi Simon, > > > > First of all, thank you and all the contributors for such a great toolkit. > I've been a RTK user for many years and it has been extremely helpful to my > research. > > > > I've been developing my programs using RTK for quite a while but haven't > really been active in this community. > > I have implemented the PICCS method in RTK (still needs to be tidied up a > bit though) and would like to contribute to the repository if it's not yet > available. > > It is probably not developed with the best computation efficiency, but > hopefully would still be useful to users who wish to test out the PICCS > method. > > > > If you think this could be useful to RTK, could you please let me know > what would be the best way to contribute my codes to RTK? > > > > Many thanks, > > Andy > > > > *DR CHUN-CHIEN (ANDY) SHIEH* | Research Associate > > Radiation Physics Laboratory | Sydney Medical School > > > > *THE UNIVERSITY OF SYDNEY* > Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 > * T* +61 2 9036 5012 > *E* *andy.shieh at sydney.edu.au * | *W* > *http://sydney.edu.au/medicine/radiation-physics* > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Wed Aug 24 17:58:28 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Wed, 24 Aug 2016 21:58:28 +0000 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: , Message-ID: <7587051e-cbb3-4865-99de-9b6ffe5bf201@sydney.edu.au> Thanks Simon. I will have a look and start working on it! Cheers, Andy Sent from Nine ________________________________ From: Simon Rit Sent: 24 Aug 2016 23:58 To: Chun-Chien (Andy) Shieh Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Contribute to RTK? Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh > wrote: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From theday79 at gmail.com Tue Aug 16 00:04:43 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 00:04:43 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: Message-ID: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in "cufft.h". It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang -------------- next part -------------- An HTML attachment was scrubbed... URL: From andreasg at phys.au.dk Tue Aug 16 04:21:04 2016 From: andreasg at phys.au.dk (Andreas Gravgaard Andersen) Date: Tue, 16 Aug 2016 10:21:04 +0200 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" wrote: > Hi all, > > > > I used to use RTK r1.03 without any problem. > > However, today I adapted rtkfdk to my own code and ran into the following > compile error messages.. > > > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan2d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan3d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecR2C referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecC2R referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftSetCompatibilityMode > referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const > &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@ > @YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftDestroy referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > > > I found that this error was caused by the following line in my code: > > > > #if CUDA_FOUND > > feldkamp = FDKCUDAType::New(); ? here! > > SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer > ())); > > #else > > .. > > I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, > cufftExecR2C, ..) are defined in ?cufft.h?. > > > > It is very weird to me because I did the exactly same thing in my previous > project and there was no problem at that time. > > Maybe I did something wrong but have no idea now. > > > > Any helps/suggestions will be highly appreciated! > > > > Regards, > > Yang > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cmake_mod.zip Type: application/zip Size: 7846 bytes Desc: not available URL: From theday79 at gmail.com Tue Aug 16 08:34:15 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 08:34:15 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: <001801d1f7ba$80901900$81b04b00$@gmail.com> Hi Andreas, Thanks a lot for your help! Now I can recall what happened in my previous project. I should have linked ?cufft.lib? manually in the CMakeLists file. Now it works fine. I believe this issue happens only in the old version of RTK and may have been fixed in the Nightly version. (I have been too lazy to move on to a newer version, though.) Thanks. Yang (To RTK staff: I?m sorry that I forgot to change the title even though this is a separate thread.) From: andreasga22 at gmail.com [mailto:andreasga22 at gmail.com] On Behalf Of Andreas Gravgaard Andersen Sent: Tuesday, August 16, 2016 4:21 AM To: Yang-Kyun Park Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Geometric phantom example on wiki Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" > wrote: Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in ?cufft.h?. It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Fri Aug 19 02:59:38 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Fri, 19 Aug 2016 06:59:38 +0000 Subject: [Rtk-users] Contribute to RTK? Message-ID: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics -------------- next part -------------- An HTML attachment was scrubbed... URL: From K.Niepel at physik.uni-muenchen.de Tue Aug 23 10:43:06 2016 From: K.Niepel at physik.uni-muenchen.de (K.Niepel) Date: Tue, 23 Aug 2016 16:43:06 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk Message-ID: <57BC60FA.40600@physik.uni-muenchen.de> Hello there, I am currently trying out the rtkfourdfdk application on a moving phantom data set with the number of phase bins varying from 15 down to one. However, there is no noticeable blurring for lower fewer bins as with the 3D rtkfdk app - actually, the first frame seems to be exactly the same for any total number of frames. The only setting i changed is the --frames argument. Am I missing something? Thanks in advance, Kati // -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 05:04:13 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 11:04:13 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk In-Reply-To: <57BC60FA.40600@physik.uni-muenchen.de> References: <57BC60FA.40600@physik.uni-muenchen.de> Message-ID: Hi, Quickly checking the code, the reason is pretty simple: rtkfourdfdk uses rtkSelectOneProjectionPerCycleImageFilter.h, which selects the closest projection to a given phase in each respiratory cycle. So setting the number of frames will not increase the number of frames per phase (and, therefore, blur). I think it's quite a common way of sorting in the literature, see for example [Sonke et al, Med Phys, 2005]: "For each breathing cycle, the projection closest to each breathing phase is selected." The reason is explained in [Dietrich et al, PMB, 2006] or in a simulation work I did in the past. Simon On Tue, Aug 23, 2016 at 4:43 PM, K.Niepel wrote: > Hello there, > > I am currently trying out the rtkfourdfdk application on a moving phantom > data set with the number of phase bins varying from 15 down to one. > However, there is no noticeable blurring for lower fewer bins as with the > 3D rtkfdk app - actually, the first frame seems to be exactly the same for > any total number of frames. > > The only setting i changed is the --frames argument. Am I missing > something? > > Thanks in advance, > Kati > > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 09:58:11 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 15:58:11 +0200 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: Message-ID: Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page ) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh < andy.shieh at sydney.edu.au> wrote: > Hi Simon, > > > > First of all, thank you and all the contributors for such a great toolkit. > I've been a RTK user for many years and it has been extremely helpful to my > research. > > > > I've been developing my programs using RTK for quite a while but haven't > really been active in this community. > > I have implemented the PICCS method in RTK (still needs to be tidied up a > bit though) and would like to contribute to the repository if it's not yet > available. > > It is probably not developed with the best computation efficiency, but > hopefully would still be useful to users who wish to test out the PICCS > method. > > > > If you think this could be useful to RTK, could you please let me know > what would be the best way to contribute my codes to RTK? > > > > Many thanks, > > Andy > > > > *DR CHUN-CHIEN (ANDY) SHIEH* | Research Associate > > Radiation Physics Laboratory | Sydney Medical School > > > > *THE UNIVERSITY OF SYDNEY* > Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 > * T* +61 2 9036 5012 > *E* *andy.shieh at sydney.edu.au * | *W* > *http://sydney.edu.au/medicine/radiation-physics* > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Wed Aug 24 17:58:28 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Wed, 24 Aug 2016 21:58:28 +0000 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: , Message-ID: <7587051e-cbb3-4865-99de-9b6ffe5bf201@sydney.edu.au> Thanks Simon. I will have a look and start working on it! Cheers, Andy Sent from Nine ________________________________ From: Simon Rit Sent: 24 Aug 2016 23:58 To: Chun-Chien (Andy) Shieh Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Contribute to RTK? Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh > wrote: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From theday79 at gmail.com Tue Aug 16 00:04:43 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 00:04:43 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: Message-ID: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in "cufft.h". It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang -------------- next part -------------- An HTML attachment was scrubbed... URL: From andreasg at phys.au.dk Tue Aug 16 04:21:04 2016 From: andreasg at phys.au.dk (Andreas Gravgaard Andersen) Date: Tue, 16 Aug 2016 10:21:04 +0200 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" wrote: > Hi all, > > > > I used to use RTK r1.03 without any problem. > > However, today I adapted rtkfdk to my own code and ran into the following > compile error messages.. > > > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan2d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan3d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecR2C referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecC2R referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftSetCompatibilityMode > referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const > &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@ > @YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftDestroy referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > > > I found that this error was caused by the following line in my code: > > > > #if CUDA_FOUND > > feldkamp = FDKCUDAType::New(); ? here! > > SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer > ())); > > #else > > .. > > I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, > cufftExecR2C, ..) are defined in ?cufft.h?. > > > > It is very weird to me because I did the exactly same thing in my previous > project and there was no problem at that time. > > Maybe I did something wrong but have no idea now. > > > > Any helps/suggestions will be highly appreciated! > > > > Regards, > > Yang > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cmake_mod.zip Type: application/zip Size: 7846 bytes Desc: not available URL: From theday79 at gmail.com Tue Aug 16 08:34:15 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 08:34:15 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: <001801d1f7ba$80901900$81b04b00$@gmail.com> Hi Andreas, Thanks a lot for your help! Now I can recall what happened in my previous project. I should have linked ?cufft.lib? manually in the CMakeLists file. Now it works fine. I believe this issue happens only in the old version of RTK and may have been fixed in the Nightly version. (I have been too lazy to move on to a newer version, though.) Thanks. Yang (To RTK staff: I?m sorry that I forgot to change the title even though this is a separate thread.) From: andreasga22 at gmail.com [mailto:andreasga22 at gmail.com] On Behalf Of Andreas Gravgaard Andersen Sent: Tuesday, August 16, 2016 4:21 AM To: Yang-Kyun Park Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Geometric phantom example on wiki Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" > wrote: Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in ?cufft.h?. It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Fri Aug 19 02:59:38 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Fri, 19 Aug 2016 06:59:38 +0000 Subject: [Rtk-users] Contribute to RTK? Message-ID: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics -------------- next part -------------- An HTML attachment was scrubbed... URL: From K.Niepel at physik.uni-muenchen.de Tue Aug 23 10:43:06 2016 From: K.Niepel at physik.uni-muenchen.de (K.Niepel) Date: Tue, 23 Aug 2016 16:43:06 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk Message-ID: <57BC60FA.40600@physik.uni-muenchen.de> Hello there, I am currently trying out the rtkfourdfdk application on a moving phantom data set with the number of phase bins varying from 15 down to one. However, there is no noticeable blurring for lower fewer bins as with the 3D rtkfdk app - actually, the first frame seems to be exactly the same for any total number of frames. The only setting i changed is the --frames argument. Am I missing something? Thanks in advance, Kati // -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 05:04:13 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 11:04:13 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk In-Reply-To: <57BC60FA.40600@physik.uni-muenchen.de> References: <57BC60FA.40600@physik.uni-muenchen.de> Message-ID: Hi, Quickly checking the code, the reason is pretty simple: rtkfourdfdk uses rtkSelectOneProjectionPerCycleImageFilter.h, which selects the closest projection to a given phase in each respiratory cycle. So setting the number of frames will not increase the number of frames per phase (and, therefore, blur). I think it's quite a common way of sorting in the literature, see for example [Sonke et al, Med Phys, 2005]: "For each breathing cycle, the projection closest to each breathing phase is selected." The reason is explained in [Dietrich et al, PMB, 2006] or in a simulation work I did in the past. Simon On Tue, Aug 23, 2016 at 4:43 PM, K.Niepel wrote: > Hello there, > > I am currently trying out the rtkfourdfdk application on a moving phantom > data set with the number of phase bins varying from 15 down to one. > However, there is no noticeable blurring for lower fewer bins as with the > 3D rtkfdk app - actually, the first frame seems to be exactly the same for > any total number of frames. > > The only setting i changed is the --frames argument. Am I missing > something? > > Thanks in advance, > Kati > > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 09:58:11 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 15:58:11 +0200 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: Message-ID: Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page ) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh < andy.shieh at sydney.edu.au> wrote: > Hi Simon, > > > > First of all, thank you and all the contributors for such a great toolkit. > I've been a RTK user for many years and it has been extremely helpful to my > research. > > > > I've been developing my programs using RTK for quite a while but haven't > really been active in this community. > > I have implemented the PICCS method in RTK (still needs to be tidied up a > bit though) and would like to contribute to the repository if it's not yet > available. > > It is probably not developed with the best computation efficiency, but > hopefully would still be useful to users who wish to test out the PICCS > method. > > > > If you think this could be useful to RTK, could you please let me know > what would be the best way to contribute my codes to RTK? > > > > Many thanks, > > Andy > > > > *DR CHUN-CHIEN (ANDY) SHIEH* | Research Associate > > Radiation Physics Laboratory | Sydney Medical School > > > > *THE UNIVERSITY OF SYDNEY* > Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 > * T* +61 2 9036 5012 > *E* *andy.shieh at sydney.edu.au * | *W* > *http://sydney.edu.au/medicine/radiation-physics* > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Wed Aug 24 17:58:28 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Wed, 24 Aug 2016 21:58:28 +0000 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: , Message-ID: <7587051e-cbb3-4865-99de-9b6ffe5bf201@sydney.edu.au> Thanks Simon. I will have a look and start working on it! Cheers, Andy Sent from Nine ________________________________ From: Simon Rit Sent: 24 Aug 2016 23:58 To: Chun-Chien (Andy) Shieh Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Contribute to RTK? Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh > wrote: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From theday79 at gmail.com Tue Aug 16 00:04:43 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 00:04:43 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: Message-ID: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in "cufft.h". It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang -------------- next part -------------- An HTML attachment was scrubbed... URL: From andreasg at phys.au.dk Tue Aug 16 04:21:04 2016 From: andreasg at phys.au.dk (Andreas Gravgaard Andersen) Date: Tue, 16 Aug 2016 10:21:04 +0200 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" wrote: > Hi all, > > > > I used to use RTK r1.03 without any problem. > > However, today I adapted rtkfdk to my own code and ran into the following > compile error messages.. > > > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan2d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan3d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecR2C referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecC2R referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftSetCompatibilityMode > referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const > &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@ > @YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftDestroy referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > > > I found that this error was caused by the following line in my code: > > > > #if CUDA_FOUND > > feldkamp = FDKCUDAType::New(); ? here! > > SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer > ())); > > #else > > .. > > I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, > cufftExecR2C, ..) are defined in ?cufft.h?. > > > > It is very weird to me because I did the exactly same thing in my previous > project and there was no problem at that time. > > Maybe I did something wrong but have no idea now. > > > > Any helps/suggestions will be highly appreciated! > > > > Regards, > > Yang > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cmake_mod.zip Type: application/zip Size: 7846 bytes Desc: not available URL: From theday79 at gmail.com Tue Aug 16 08:34:15 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 08:34:15 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: <001801d1f7ba$80901900$81b04b00$@gmail.com> Hi Andreas, Thanks a lot for your help! Now I can recall what happened in my previous project. I should have linked ?cufft.lib? manually in the CMakeLists file. Now it works fine. I believe this issue happens only in the old version of RTK and may have been fixed in the Nightly version. (I have been too lazy to move on to a newer version, though.) Thanks. Yang (To RTK staff: I?m sorry that I forgot to change the title even though this is a separate thread.) From: andreasga22 at gmail.com [mailto:andreasga22 at gmail.com] On Behalf Of Andreas Gravgaard Andersen Sent: Tuesday, August 16, 2016 4:21 AM To: Yang-Kyun Park Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Geometric phantom example on wiki Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" > wrote: Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in ?cufft.h?. It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Fri Aug 19 02:59:38 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Fri, 19 Aug 2016 06:59:38 +0000 Subject: [Rtk-users] Contribute to RTK? Message-ID: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics -------------- next part -------------- An HTML attachment was scrubbed... URL: From K.Niepel at physik.uni-muenchen.de Tue Aug 23 10:43:06 2016 From: K.Niepel at physik.uni-muenchen.de (K.Niepel) Date: Tue, 23 Aug 2016 16:43:06 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk Message-ID: <57BC60FA.40600@physik.uni-muenchen.de> Hello there, I am currently trying out the rtkfourdfdk application on a moving phantom data set with the number of phase bins varying from 15 down to one. However, there is no noticeable blurring for lower fewer bins as with the 3D rtkfdk app - actually, the first frame seems to be exactly the same for any total number of frames. The only setting i changed is the --frames argument. Am I missing something? Thanks in advance, Kati // -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 05:04:13 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 11:04:13 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk In-Reply-To: <57BC60FA.40600@physik.uni-muenchen.de> References: <57BC60FA.40600@physik.uni-muenchen.de> Message-ID: Hi, Quickly checking the code, the reason is pretty simple: rtkfourdfdk uses rtkSelectOneProjectionPerCycleImageFilter.h, which selects the closest projection to a given phase in each respiratory cycle. So setting the number of frames will not increase the number of frames per phase (and, therefore, blur). I think it's quite a common way of sorting in the literature, see for example [Sonke et al, Med Phys, 2005]: "For each breathing cycle, the projection closest to each breathing phase is selected." The reason is explained in [Dietrich et al, PMB, 2006] or in a simulation work I did in the past. Simon On Tue, Aug 23, 2016 at 4:43 PM, K.Niepel wrote: > Hello there, > > I am currently trying out the rtkfourdfdk application on a moving phantom > data set with the number of phase bins varying from 15 down to one. > However, there is no noticeable blurring for lower fewer bins as with the > 3D rtkfdk app - actually, the first frame seems to be exactly the same for > any total number of frames. > > The only setting i changed is the --frames argument. Am I missing > something? > > Thanks in advance, > Kati > > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 09:58:11 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 15:58:11 +0200 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: Message-ID: Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page ) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh < andy.shieh at sydney.edu.au> wrote: > Hi Simon, > > > > First of all, thank you and all the contributors for such a great toolkit. > I've been a RTK user for many years and it has been extremely helpful to my > research. > > > > I've been developing my programs using RTK for quite a while but haven't > really been active in this community. > > I have implemented the PICCS method in RTK (still needs to be tidied up a > bit though) and would like to contribute to the repository if it's not yet > available. > > It is probably not developed with the best computation efficiency, but > hopefully would still be useful to users who wish to test out the PICCS > method. > > > > If you think this could be useful to RTK, could you please let me know > what would be the best way to contribute my codes to RTK? > > > > Many thanks, > > Andy > > > > *DR CHUN-CHIEN (ANDY) SHIEH* | Research Associate > > Radiation Physics Laboratory | Sydney Medical School > > > > *THE UNIVERSITY OF SYDNEY* > Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 > * T* +61 2 9036 5012 > *E* *andy.shieh at sydney.edu.au * | *W* > *http://sydney.edu.au/medicine/radiation-physics* > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Wed Aug 24 17:58:28 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Wed, 24 Aug 2016 21:58:28 +0000 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: , Message-ID: <7587051e-cbb3-4865-99de-9b6ffe5bf201@sydney.edu.au> Thanks Simon. I will have a look and start working on it! Cheers, Andy Sent from Nine ________________________________ From: Simon Rit Sent: 24 Aug 2016 23:58 To: Chun-Chien (Andy) Shieh Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Contribute to RTK? Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh > wrote: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From theday79 at gmail.com Tue Aug 16 00:04:43 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 00:04:43 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: Message-ID: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in "cufft.h". It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang -------------- next part -------------- An HTML attachment was scrubbed... URL: From andreasg at phys.au.dk Tue Aug 16 04:21:04 2016 From: andreasg at phys.au.dk (Andreas Gravgaard Andersen) Date: Tue, 16 Aug 2016 10:21:04 +0200 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" wrote: > Hi all, > > > > I used to use RTK r1.03 without any problem. > > However, today I adapted rtkfdk to my own code and ran into the following > compile error messages.. > > > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan2d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan3d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecR2C referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecC2R referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftSetCompatibilityMode > referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const > &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@ > @YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftDestroy referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > > > I found that this error was caused by the following line in my code: > > > > #if CUDA_FOUND > > feldkamp = FDKCUDAType::New(); ? here! > > SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer > ())); > > #else > > .. > > I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, > cufftExecR2C, ..) are defined in ?cufft.h?. > > > > It is very weird to me because I did the exactly same thing in my previous > project and there was no problem at that time. > > Maybe I did something wrong but have no idea now. > > > > Any helps/suggestions will be highly appreciated! > > > > Regards, > > Yang > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cmake_mod.zip Type: application/zip Size: 7846 bytes Desc: not available URL: From theday79 at gmail.com Tue Aug 16 08:34:15 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 08:34:15 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: <001801d1f7ba$80901900$81b04b00$@gmail.com> Hi Andreas, Thanks a lot for your help! Now I can recall what happened in my previous project. I should have linked ?cufft.lib? manually in the CMakeLists file. Now it works fine. I believe this issue happens only in the old version of RTK and may have been fixed in the Nightly version. (I have been too lazy to move on to a newer version, though.) Thanks. Yang (To RTK staff: I?m sorry that I forgot to change the title even though this is a separate thread.) From: andreasga22 at gmail.com [mailto:andreasga22 at gmail.com] On Behalf Of Andreas Gravgaard Andersen Sent: Tuesday, August 16, 2016 4:21 AM To: Yang-Kyun Park Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Geometric phantom example on wiki Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" > wrote: Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in ?cufft.h?. It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Fri Aug 19 02:59:38 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Fri, 19 Aug 2016 06:59:38 +0000 Subject: [Rtk-users] Contribute to RTK? Message-ID: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics -------------- next part -------------- An HTML attachment was scrubbed... URL: From K.Niepel at physik.uni-muenchen.de Tue Aug 23 10:43:06 2016 From: K.Niepel at physik.uni-muenchen.de (K.Niepel) Date: Tue, 23 Aug 2016 16:43:06 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk Message-ID: <57BC60FA.40600@physik.uni-muenchen.de> Hello there, I am currently trying out the rtkfourdfdk application on a moving phantom data set with the number of phase bins varying from 15 down to one. However, there is no noticeable blurring for lower fewer bins as with the 3D rtkfdk app - actually, the first frame seems to be exactly the same for any total number of frames. The only setting i changed is the --frames argument. Am I missing something? Thanks in advance, Kati // -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 05:04:13 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 11:04:13 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk In-Reply-To: <57BC60FA.40600@physik.uni-muenchen.de> References: <57BC60FA.40600@physik.uni-muenchen.de> Message-ID: Hi, Quickly checking the code, the reason is pretty simple: rtkfourdfdk uses rtkSelectOneProjectionPerCycleImageFilter.h, which selects the closest projection to a given phase in each respiratory cycle. So setting the number of frames will not increase the number of frames per phase (and, therefore, blur). I think it's quite a common way of sorting in the literature, see for example [Sonke et al, Med Phys, 2005]: "For each breathing cycle, the projection closest to each breathing phase is selected." The reason is explained in [Dietrich et al, PMB, 2006] or in a simulation work I did in the past. Simon On Tue, Aug 23, 2016 at 4:43 PM, K.Niepel wrote: > Hello there, > > I am currently trying out the rtkfourdfdk application on a moving phantom > data set with the number of phase bins varying from 15 down to one. > However, there is no noticeable blurring for lower fewer bins as with the > 3D rtkfdk app - actually, the first frame seems to be exactly the same for > any total number of frames. > > The only setting i changed is the --frames argument. Am I missing > something? > > Thanks in advance, > Kati > > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 09:58:11 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 15:58:11 +0200 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: Message-ID: Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page ) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh < andy.shieh at sydney.edu.au> wrote: > Hi Simon, > > > > First of all, thank you and all the contributors for such a great toolkit. > I've been a RTK user for many years and it has been extremely helpful to my > research. > > > > I've been developing my programs using RTK for quite a while but haven't > really been active in this community. > > I have implemented the PICCS method in RTK (still needs to be tidied up a > bit though) and would like to contribute to the repository if it's not yet > available. > > It is probably not developed with the best computation efficiency, but > hopefully would still be useful to users who wish to test out the PICCS > method. > > > > If you think this could be useful to RTK, could you please let me know > what would be the best way to contribute my codes to RTK? > > > > Many thanks, > > Andy > > > > *DR CHUN-CHIEN (ANDY) SHIEH* | Research Associate > > Radiation Physics Laboratory | Sydney Medical School > > > > *THE UNIVERSITY OF SYDNEY* > Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 > * T* +61 2 9036 5012 > *E* *andy.shieh at sydney.edu.au * | *W* > *http://sydney.edu.au/medicine/radiation-physics* > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Wed Aug 24 17:58:28 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Wed, 24 Aug 2016 21:58:28 +0000 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: , Message-ID: <7587051e-cbb3-4865-99de-9b6ffe5bf201@sydney.edu.au> Thanks Simon. I will have a look and start working on it! Cheers, Andy Sent from Nine ________________________________ From: Simon Rit Sent: 24 Aug 2016 23:58 To: Chun-Chien (Andy) Shieh Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Contribute to RTK? Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh > wrote: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From theday79 at gmail.com Tue Aug 16 00:04:43 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 00:04:43 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: Message-ID: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in "cufft.h". It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang -------------- next part -------------- An HTML attachment was scrubbed... URL: From andreasg at phys.au.dk Tue Aug 16 04:21:04 2016 From: andreasg at phys.au.dk (Andreas Gravgaard Andersen) Date: Tue, 16 Aug 2016 10:21:04 +0200 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" wrote: > Hi all, > > > > I used to use RTK r1.03 without any problem. > > However, today I adapted rtkfdk to my own code and ran into the following > compile error messages.. > > > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan2d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftPlan3d referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecR2C referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftExecC2R referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftSetCompatibilityMode > referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const > &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@ > @YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) > > 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : > error LNK2019: unresolved external symbol cufftDestroy referenced in > function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 > const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@ > @AEBUint2@@PEAMPEAUfloat2@@@Z) > > > > I found that this error was caused by the following line in my code: > > > > #if CUDA_FOUND > > feldkamp = FDKCUDAType::New(); ? here! > > SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer > ())); > > #else > > .. > > I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, > cufftExecR2C, ..) are defined in ?cufft.h?. > > > > It is very weird to me because I did the exactly same thing in my previous > project and there was no problem at that time. > > Maybe I did something wrong but have no idea now. > > > > Any helps/suggestions will be highly appreciated! > > > > Regards, > > Yang > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cmake_mod.zip Type: application/zip Size: 7846 bytes Desc: not available URL: From theday79 at gmail.com Tue Aug 16 08:34:15 2016 From: theday79 at gmail.com (Yang K Park) Date: Tue, 16 Aug 2016 08:34:15 -0400 Subject: [Rtk-users] Geometric phantom example on wiki In-Reply-To: References: <005c01d1f773$5228f290$f67ad7b0$@gmail.com> Message-ID: <001801d1f7ba$80901900$81b04b00$@gmail.com> Hi Andreas, Thanks a lot for your help! Now I can recall what happened in my previous project. I should have linked ?cufft.lib? manually in the CMakeLists file. Now it works fine. I believe this issue happens only in the old version of RTK and may have been fixed in the Nightly version. (I have been too lazy to move on to a newer version, though.) Thanks. Yang (To RTK staff: I?m sorry that I forgot to change the title even though this is a separate thread.) From: andreasga22 at gmail.com [mailto:andreasga22 at gmail.com] On Behalf Of Andreas Gravgaard Andersen Sent: Tuesday, August 16, 2016 4:21 AM To: Yang-Kyun Park Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Geometric phantom example on wiki Hi Yang, I had a similar problem when I was compiling your CBCT reconstruction software - I thought it was due me having a different version of CUDA. I solved it by changing the cmakelists.txt to use a FindCuda_wrap.cmake and associated files (copied and slightly modified from the rtk one.) to let those files do the linking. I also changed the CMakelists.txt to make it compile on non-cuda hardware, so I'm not completely sure whether this might have been the solution to that. Attached in zip archive. :: My current CBCTrecon -> https://github.com/agravgaard/cbctrecon.git I have since changed the source to use the newest git version of RTK (keeping the OpenCL code in my own fork ), but I don't think that made any difference in the linking. Best regards Andreas On 16 Aug 2016 6:04 am, "Yang K Park" > wrote: Hi all, I used to use RTK r1.03 without any problem. However, today I adapted rtkfdk to my own code and ran into the following compile error messages.. 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan2d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftPlan3d referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecR2C referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftExecC2R referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftSetCompatibilityMode referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) 1>rtkcuda.lib(cuda_compile_generated_rtkCudaFFTRampImageFilter.cu.obj) : error LNK2019: unresolved external symbol cufftDestroy referenced in function "void __cdecl CUDA_fft_convolution(struct int3 const &,struct int2 const &,float *,struct float2 *)" (?CUDA_fft_convolution@@YAXAEBUint3@@AEBUint2@@PEAMPEAUfloat2@@@Z) I found that this error was caused by the following line in my code: #if CUDA_FOUND feldkamp = FDKCUDAType::New(); <-- here! SET_FELDKAMP_OPTIONS(static_cast(feldkamp.GetPointer())); #else .. I also found that those problematic symbols (cufftPlan2d, cufftPlan3d, cufftExecR2C, ..) are defined in ?cufft.h?. It is very weird to me because I did the exactly same thing in my previous project and there was no problem at that time. Maybe I did something wrong but have no idea now. Any helps/suggestions will be highly appreciated! Regards, Yang _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Fri Aug 19 02:59:38 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Fri, 19 Aug 2016 06:59:38 +0000 Subject: [Rtk-users] Contribute to RTK? Message-ID: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics -------------- next part -------------- An HTML attachment was scrubbed... URL: From K.Niepel at physik.uni-muenchen.de Tue Aug 23 10:43:06 2016 From: K.Niepel at physik.uni-muenchen.de (K.Niepel) Date: Tue, 23 Aug 2016 16:43:06 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk Message-ID: <57BC60FA.40600@physik.uni-muenchen.de> Hello there, I am currently trying out the rtkfourdfdk application on a moving phantom data set with the number of phase bins varying from 15 down to one. However, there is no noticeable blurring for lower fewer bins as with the 3D rtkfdk app - actually, the first frame seems to be exactly the same for any total number of frames. The only setting i changed is the --frames argument. Am I missing something? Thanks in advance, Kati // -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 05:04:13 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 11:04:13 +0200 Subject: [Rtk-users] Frames in rtkfourdfdk In-Reply-To: <57BC60FA.40600@physik.uni-muenchen.de> References: <57BC60FA.40600@physik.uni-muenchen.de> Message-ID: Hi, Quickly checking the code, the reason is pretty simple: rtkfourdfdk uses rtkSelectOneProjectionPerCycleImageFilter.h, which selects the closest projection to a given phase in each respiratory cycle. So setting the number of frames will not increase the number of frames per phase (and, therefore, blur). I think it's quite a common way of sorting in the literature, see for example [Sonke et al, Med Phys, 2005]: "For each breathing cycle, the projection closest to each breathing phase is selected." The reason is explained in [Dietrich et al, PMB, 2006] or in a simulation work I did in the past. Simon On Tue, Aug 23, 2016 at 4:43 PM, K.Niepel wrote: > Hello there, > > I am currently trying out the rtkfourdfdk application on a moving phantom > data set with the number of phase bins varying from 15 down to one. > However, there is no noticeable blurring for lower fewer bins as with the > 3D rtkfdk app - actually, the first frame seems to be exactly the same for > any total number of frames. > > The only setting i changed is the --frames argument. Am I missing > something? > > Thanks in advance, > Kati > > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From simon.rit at creatis.insa-lyon.fr Wed Aug 24 09:58:11 2016 From: simon.rit at creatis.insa-lyon.fr (Simon Rit) Date: Wed, 24 Aug 2016 15:58:11 +0200 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: Message-ID: Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page ) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh < andy.shieh at sydney.edu.au> wrote: > Hi Simon, > > > > First of all, thank you and all the contributors for such a great toolkit. > I've been a RTK user for many years and it has been extremely helpful to my > research. > > > > I've been developing my programs using RTK for quite a while but haven't > really been active in this community. > > I have implemented the PICCS method in RTK (still needs to be tidied up a > bit though) and would like to contribute to the repository if it's not yet > available. > > It is probably not developed with the best computation efficiency, but > hopefully would still be useful to users who wish to test out the PICCS > method. > > > > If you think this could be useful to RTK, could you please let me know > what would be the best way to contribute my codes to RTK? > > > > Many thanks, > > Andy > > > > *DR CHUN-CHIEN (ANDY) SHIEH* | Research Associate > > Radiation Physics Laboratory | Sydney Medical School > > > > *THE UNIVERSITY OF SYDNEY* > Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 > * T* +61 2 9036 5012 > *E* *andy.shieh at sydney.edu.au * | *W* > *http://sydney.edu.au/medicine/radiation-physics* > > > _______________________________________________ > Rtk-users mailing list > Rtk-users at public.kitware.com > http://public.kitware.com/mailman/listinfo/rtk-users > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From andy.shieh at sydney.edu.au Wed Aug 24 17:58:28 2016 From: andy.shieh at sydney.edu.au (Chun-Chien (Andy) Shieh) Date: Wed, 24 Aug 2016 21:58:28 +0000 Subject: [Rtk-users] Contribute to RTK? In-Reply-To: References: , Message-ID: <7587051e-cbb3-4865-99de-9b6ffe5bf201@sydney.edu.au> Thanks Simon. I will have a look and start working on it! Cheers, Andy Sent from Nine ________________________________ From: Simon Rit Sent: 24 Aug 2016 23:58 To: Chun-Chien (Andy) Shieh Cc: rtk-users at public.kitware.com Subject: Re: [Rtk-users] Contribute to RTK? Hi Andy, Thanks, always good to hear that this is useful to some. Of course your dev are interesting to other, particularly a new recon algorithm. The best way to contribute is to do a fork on github (like other people, see this page) and to prepare your contribution on a branch in your own github. When it's ready, send a pull-request for review and we'll integrate it if it's ready or comment on the pull-request. For a new recon algorithm, we normally do a corresponding functional test. See the testing subfolder for examples. It's usually a simple Shepp Logan test. That's even better if you can also dev this test. I'm looking forward to your contributions! Simon On Fri, Aug 19, 2016 at 8:59 AM, Chun-Chien (Andy) Shieh > wrote: Hi Simon, First of all, thank you and all the contributors for such a great toolkit. I've been a RTK user for many years and it has been extremely helpful to my research. I've been developing my programs using RTK for quite a while but haven't really been active in this community. I have implemented the PICCS method in RTK (still needs to be tidied up a bit though) and would like to contribute to the repository if it's not yet available. It is probably not developed with the best computation efficiency, but hopefully would still be useful to users who wish to test out the PICCS method. If you think this could be useful to RTK, could you please let me know what would be the best way to contribute my codes to RTK? Many thanks, Andy DR CHUN-CHIEN (ANDY) SHIEH | Research Associate Radiation Physics Laboratory | Sydney Medical School THE UNIVERSITY OF SYDNEY Room 479, Blackburn Building D06 | The University of Sydney | NSW | 2006 T +61 2 9036 5012 E andy.shieh at sydney.edu.au | W http://sydney.edu.au/medicine/radiation-physics _______________________________________________ Rtk-users mailing list Rtk-users at public.kitware.com http://public.kitware.com/mailman/listinfo/rtk-users -------------- next part -------------- An HTML attachment was scrubbed... URL: