<div dir="ltr">Hi Dave,<div><br></div><div>Without the lowmem flag all projection images may be loaded into graphics memory at once causing an out-of-memory issues.</div><div>You can also check this thread:</div><div><a href="http://public.kitware.com/pipermail/rtk-users/2015-June/000510.html">http://public.kitware.com/pipermail/rtk-users/2015-June/000510.html</a><br></div><div><br></div><div>Regards,</div><div>Chao</div></div><div class="gmail_extra"><br><div class="gmail_quote">2015-10-14 22:27 GMT+02:00 sco user <span dir="ltr"><<a href="mailto:scouser27@gmail.com" target="_blank">scouser27@gmail.com</a>></span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div><div><div><div><div><div><div><div>Hi<br><br></div>I'm a new user to RTK. <br></div>I've been able to perform reconstructions of the Catphan and my own copies of Varian CBCT images with some success. I enabled the CUDA flags when making RTK so as to take advantage of the capabilities of my GPU.<br><br></div>When I use --hardware="cuda" with the rtkfdk application, I have found that I must specify the lowmem option in order to get the reconstruction to run successfully. If I don't use lowmem, I get the following error:<br><br>####<br>Reconstructing and writing... /home/dave/RTK/utilities/ITKCudaCommon/include/itkCudaDataManager.h:55 @ unknown : Cuda Error : out of memory<br><br>ExceptionObject caught with writer->Update() in file /home/dave/RTK/applications/rtkfdk/rtkfdk.cxx line 238<br><br>itk::ExceptionObject (0x18e3460)<br>Location: "unknown" <br>File: /home/dave/RTK/utilities/ITKCudaCommon/include/itkCudaDataManager.h<br>Line: 55<br>Description: Cuda Error : out of memory<br>####<br><br></div>I take from this that I'm out of memory on the GPU. Is that correct? The GPU has 2Gb of memory. Can anyone tell me what the minimum memory should be in order to get a typical Varian reconstruction to run without the lowmem option.<br><br></div>On a side note, I know that the cuda option is working because it's a lot quicker than hardware="cpu", despite only having 96 cuda cores.<br><br><br></div>Another question I'd like to ask is if anyone can shed some light on useful values for I-zero, wpc, scatter etc. <br><br></div>With thanks in advance,<br></div>Dave.<br></div>
<br>
<br>_______________________________________________<br>
Rtk-users mailing list<br>
<a href="mailto:Rtk-users@public.kitware.com">Rtk-users@public.kitware.com</a><br>
<a href="http://public.kitware.com/mailman/listinfo/rtk-users" rel="noreferrer" target="_blank">http://public.kitware.com/mailman/listinfo/rtk-users</a><br>
<br></blockquote></div><br></div>