[ITK-users] GPU support situation for ITK

[Xiaopeng Yang]

Hi Joel,

 

Actually I have been thinking for a long time whether I should implement GPU
or not. I am using DemonsRegistrationFilter to align multi-phase abdominal
CT datasets. I have three different phases, so each time I need to perform
two registrations which take more than 12 min (each 6 min). My customers
really hate this long time computation. That is why I am trying to solve the
problem as much as I can, but I need a right direction to take my effort on
since I do not have much time. Thank you.

 

Best regards,

Xiaopeng

 

From: Insight-users [mailto:insight-users-bounces at itk.org] On Behalf Of Jo?l
Schaerer
Sent: Friday, April 18, 2014 4:34 AM
To: insight-users at itk.org
Subject: Re: [ITK-users] GPU support situation for ITK

 

Hi Xiaopeng,

while it would probably be a fun project to work on, I would recommend
reflecting a bit more on your problem before rushing into this.

What is your application? What registration algorithm are you using? Is
every computation performed during these 6 minutes absolutely needed for
your final application?

Sometimes thinking a bit about algorithms and the application at hand is
even more beneficial than rolling out a super optimized implementation.

just my 2 cents,

joel

On 04/17/2014 04:46 PM, Xiaopeng Yang wrote:

Then do you think it is worthwhile to give a shot? Right now the algorithm I
used for non-rigid registration takes about 6 min on a desktop with Intel
Core i7 3.2Ghz and 32G Ram. I would like to get the results in almost real
time, so I am thinking about implementing the GPU.

 

Best regards,

Xiaopeng

From: Jim Miller [mailto:millerjv at gmail.com] 
Sent: Thursday, April 17, 2014 11:18 PM
To: Xiaopeng Yang
Cc:  <mailto:insight-users at itk.org> <insight-users at itk.org>
Subject: Re: [ITK-users] GPU support situation for ITK

 

Implementing algorithms on the GPU is a skill. Also, it is tricky to get a
GPU implementation to work across a large variety of GPUs. 

Jim


On Apr 17, 2014, at 10:05 AM, "Xiaopeng Yang" <yxp233 at postech.ac.kr> wrote:

Thank you Jim. I am still using ITK 3. I will try the latest version. By the
way, is it difficult to implement the GPU to other algorithms? Can the GPU
be implemented to all other algorithms?

 

Thanks,

Xiaopeng

 

From: Jim Miller [mailto:millerjv at gmail.com] 
Sent: Thursday, April 17, 2014 10:46 PM
To: Xiaopeng Yang
Cc: <insight-users at itk.org>
Subject: Re: [ITK-users] GPU support situation for ITK

 

A framework for GPU filters is in place in ITK. A limited number of
algorithms have been implemented using the GPU to serve as models as to how
other algorithms can be implemented. 

Jim


On Apr 17, 2014, at 9:25 AM, "Xiaopeng Yang" <yxp233 at postech.ac.kr> wrote:

Dear Users,

 

I am wondering what is the current situation of GPU implementation in ITK.
For example, for segmentation algorithms like region growing and level set
methods, has GPU been implemented? How about the registration algorithms?
Thank you very much.

 

Best regards,

Xiaopeng

_____________________________________
Powered by www.kitware.com

Visit other Kitware open-source projects at
http://www.kitware.com/opensource/opensource.html

Kitware offers ITK Training Courses, for more information visit:
http://www.kitware.com/products/protraining.php

Please keep messages on-topic and check the ITK FAQ at:
http://www.itk.org/Wiki/ITK_FAQ

Follow this link to subscribe/unsubscribe:
http://www.itk.org/mailman/listinfo/insight-users






_____________________________________
Powered by www.kitware.com
 
Visit other Kitware open-source projects at
http://www.kitware.com/opensource/opensource.html
 
Kitware offers ITK Training Courses, for more information visit:
http://www.kitware.com/products/protraining.php
 
Please keep messages on-topic and check the ITK FAQ at:
http://www.itk.org/Wiki/ITK_FAQ
 
Follow this link to subscribe/unsubscribe:
http://www.itk.org/mailman/listinfo/insight-users

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20140418/32fb0956/attachment.html>