just yesterday i re-tried cuda toolkit 3.2RC,
cufftExec returned CUFFT_SUCCESS but issues were the same both on pc and notebook.
pc configuration :
ubuntu 10.10 64bit, gcc 4.5, devdriver 260.24, eclipse with cuda plugin, AthlonX2 7850, 4gb ram, geforce gts250 1gb vram
notebook configuration:
ubuntu 10.10 64bit, gcc 4.5, devdriver 260.24, eclipse with cuda plugin, core i3 330M, 4 gb ram, geforce gt320M 1 gb vram.
For me it’s not a problem because i can work, and i actually work very well both on pc and notebook, with cuda 3.1.1 but if you tell me you need more informations about these problems with
3.2RC i can send you more images, code,etc etc…
just yesterday i re-tried cuda toolkit 3.2RC,
cufftExec returned CUFFT_SUCCESS but issues were the same both on pc and notebook.
pc configuration :
ubuntu 10.10 64bit, gcc 4.5, devdriver 260.24, eclipse with cuda plugin, AthlonX2 7850, 4gb ram, geforce gts250 1gb vram
notebook configuration:
ubuntu 10.10 64bit, gcc 4.5, devdriver 260.24, eclipse with cuda plugin, core i3 330M, 4 gb ram, geforce gt320M 1 gb vram.
For me it’s not a problem because i can work, and i actually work very well both on pc and notebook, with cuda 3.1.1 but if you tell me you need more informations about these problems with
3.2RC i can send you more images, code,etc etc…
For me it’s not a problem because i can work, and i actually work very well both on pc and notebook, with cuda 3.1.1 but if you tell me you need more informations about these problems with
3.2RC i can send you more images, code,etc etc…
Code that reproduces the problem would definitely be the most helpful thing. Can you send it to me via private message?
Thanks!
Cliff
For me it’s not a problem because i can work, and i actually work very well both on pc and notebook, with cuda 3.1.1 but if you tell me you need more informations about these problems with
3.2RC i can send you more images, code,etc etc…
Code that reproduces the problem would definitely be the most helpful thing. Can you send it to me via private message?
Thanks!
Cliff
i send to you all stuff you requested via private message,
please reply to me if you find something interesting.
thanks,
Andrea Dossi
i send to you all stuff you requested via private message,
please reply to me if you find something interesting.
thanks,
Andrea Dossi
i send to you all stuff you requested via private message,
please reply to me if you find something interesting.
thanks,
Andrea Dossi
sorry i was unable to send you readme.txt with instructions so i post it here…
Instructions:
+) extract all files of matlab_code archive and launch CreateProjectionphantom3d.m script
after have set up directory where save the images. (this will take lot of time…have a coffee break :) )
+) compile C code
+) transfer executable in projection directory and launch it
+) load volume using script LoadVol.m (A = LoadVol(‘FourierSliceCUFFT3d’); )
+) visualize reconstruction (figure, imagesc(A(:,:,192 **)), colormap(‘gray’) )
**: deep you want to see… around DIM/2 is most interesting
NB: parameters in C code and matlab code must be the same, so if you want to try
with different volume size (256, 300, 350, 380 ecc ecc) you have to recompile project and to recreate
projections in matlab with different size, default size is 384.
i’ve noticed problems using rc3.2 with volumes bigger than 300x300x300 as posted in the image post #8 …
for any question email me!
andreadossiAThotmail.com
ramarromarrone1985ATgmail.com
andrea.dossiATstudenti.unimi.it
i send to you all stuff you requested via private message,
please reply to me if you find something interesting.
thanks,
Andrea Dossi
sorry i was unable to send you readme.txt with instructions so i post it here…
Instructions:
+) extract all files of matlab_code archive and launch CreateProjectionphantom3d.m script
after have set up directory where save the images. (this will take lot of time…have a coffee break :) )
+) compile C code
+) transfer executable in projection directory and launch it
+) load volume using script LoadVol.m (A = LoadVol(‘FourierSliceCUFFT3d’); )
+) visualize reconstruction (figure, imagesc(A(:,:,192 **)), colormap(‘gray’) )
**: deep you want to see… around DIM/2 is most interesting
NB: parameters in C code and matlab code must be the same, so if you want to try
with different volume size (256, 300, 350, 380 ecc ecc) you have to recompile project and to recreate
projections in matlab with different size, default size is 384.
i’ve noticed problems using rc3.2 with volumes bigger than 300x300x300 as posted in the image post #8 …
for any question email me!
andreadossiAThotmail.com
ramarromarrone1985ATgmail.com
andrea.dossiATstudenti.unimi.it
NB: parameters in C code and matlab code must be the same, so if you want to try
with different volume size (256, 300, 350, 380 ecc ecc) you have to recompile project and to recreate
projections in matlab with different size, default size is 384.
i’ve noticed problems using rc3.2 with volumes bigger than 300x300x300 as posted in the image post #8 …
Hi Andrea,
I tried your code last week but wasn’t able to reproduce the problem you were seeing on my C2050. Perhaps I need to try on a GPU with less vidmem? I’ll give it another go that way this week sometime. Just wanted to give you an update.
–Cliff
NB: parameters in C code and matlab code must be the same, so if you want to try
with different volume size (256, 300, 350, 380 ecc ecc) you have to recompile project and to recreate
projections in matlab with different size, default size is 384.
i’ve noticed problems using rc3.2 with volumes bigger than 300x300x300 as posted in the image post #8 …
Hi Andrea,
I tried your code last week but wasn’t able to reproduce the problem you were seeing on my C2050. Perhaps I need to try on a GPU with less vidmem? I’ll give it another go that way this week sometime. Just wanted to give you an update.
–Cliff
Hi Cliff!
thank you so much for the update and for the time you spent running my code,
remember that i work with a GeForce 250GTS 1GB (cuda capability 1.1) and a GeForce320M 1GB (cuda capability 1.2) so i think you should try (if you have the opportunity) the code on a geforce 9600,9800,240,250 or something similar.
Andrea
Hi Cliff!
thank you so much for the update and for the time you spent running my code,
remember that i work with a GeForce 250GTS 1GB (cuda capability 1.1) and a GeForce320M 1GB (cuda capability 1.2) so i think you should try (if you have the opportunity) the code on a geforce 9600,9800,240,250 or something similar.
Andrea
Good day,
i didn’t receive any more news so i have waited for final version of cuda toolkit.
now cuda toolkit 3.2 has been released and i decide to try it.
the issues are the same noted in the beta and release-candidate versions.
i’ll continue to use cuda 3.1 but i hope that these problems will be resolved as soon as possible.
Best Regards
Andrea Dossi
Good day,
i didn’t receive any more news so i have waited for final version of cuda toolkit.
now cuda toolkit 3.2 has been released and i decide to try it.
the issues are the same noted in the beta and release-candidate versions.
i’ll continue to use cuda 3.1 but i hope that these problems will be resolved as soon as possible.
Best Regards
Andrea Dossi
i didn’t receive any more news so i have waited for final version of cuda toolkit.
now cuda toolkit 3.2 has been released and i decide to try it.
the issues are the same noted in the beta and release-candidate versions.
Hi Andrea,
Sorry that this is still being problematic for you. I was never able to reproduce the problem myself. If you could do me a favor and file a bug on our registered developer site, that would help to ensure that this issue gets attention from our QA team for the next time around.
Thanks,
Cliff
i didn’t receive any more news so i have waited for final version of cuda toolkit.
now cuda toolkit 3.2 has been released and i decide to try it.
the issues are the same noted in the beta and release-candidate versions.
Hi Andrea,
Sorry that this is still being problematic for you. I was never able to reproduce the problem myself. If you could do me a favor and file a bug on our registered developer site, that would help to ensure that this issue gets attention from our QA team for the next time around.
Thanks,
Cliff