Dear Expert,
I have come across that I have freesurfer support cuda version 5, but I have already installed latest version. Can I use latest version for freeesurfer or should I install version 5? Since I am not expert in this field I need guidance to remove latest version of cuda and installing and enabling version 5.
Your help will be appreciated.
PC details: I am using Intel® Core™ i7-4790 CPU @ 3.60GHz × 8 processor with 15.6 GiB ram, GeForce GT 610/PCIe/SSE2 graphic card.
Thanks RB
On 16 May 2017 at 23:10, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
I have come across that I have freesurfer support cuda version 5, but I have already installed latest version. Can I use latest version for freeesurfer or should I install version 5? Since I am not expert in this field I need guidance to remove latest version of cuda and installing and enabling version 5.
Your help will be appreciated.
PC details: I am using Intel® Core™ i7-4790 CPU @ 3.60GHz × 8 processor with 15.6 GiB ram, GeForce GT 610/PCIe/SSE2 graphic card.
I have CUDA 8 on my machine, and it seems to run quite happily (although I do compile myself). I don't believe that NVIDIA have made any breaking changes in the CUDA SDK, although looking at things recently, some might be coming.
HTH,
Richard
Dear Richard, Thank you for your reply. I tried to compile by looking FAQ about CUDA in Ubuntu Ask, but it is get corrupted and I am forced to uninstall both NVIDIA driver and CUDA. I need help to compile it. Can I get some guidance or manual?
Thanks Ramesh
On Sun, May 21, 2017 at 2:18 AM, R Edgar freesurfer.rge@gmail.com wrote:
On 16 May 2017 at 23:10, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
I have come across that I have freesurfer support cuda version 5, but I
have
already installed latest version. Can I use latest version for freeesurfer or should I install version 5? Since I am not expert in this field I need guidance to remove latest
version
of cuda and installing and enabling version 5.
Your help will be appreciated.
PC details: I am using Intel® Core™ i7-4790 CPU @ 3.60GHz × 8 processor with 15.6 GiB ram, GeForce GT 610/PCIe/SSE2 graphic card.
I have CUDA 8 on my machine, and it seems to run quite happily (although I do compile myself). I don't believe that NVIDIA have made any breaking changes in the CUDA SDK, although looking at things recently, some might be coming.
HTH,
Richard
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
On 21 May 2017 at 23:40, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
Thank you for your reply. I tried to compile by looking FAQ about CUDA in Ubuntu Ask, but it is get corrupted and I am forced to uninstall both NVIDIA driver and CUDA. I need help to compile it. Can I get some guidance or manual?
How do you mean corrupted?
I have found it necessary to edit the configure script which is generated, and to manually pass it the path to nvcc, but I don't believe I've seen something get corrupted before.
Regards,
Richard
Dear Richard,
After compiling in the same way as given in the webpage, I couldn't reboot. It was showing a message that "running low graphic mode". It was not opening the desktop. Then through ctrl+alt+f1, I uninstalled nvida and cuda. Then I could boot the system.
Finally I have installed both cuda and nvidia after going through many web sources. I have installed NVIDIA drive through additional driver option in Ubuntu software & update centre. It installed recommended nvidia driver.
For cuda installation I followed these scripts, [chmod +x cuda_5.0.35_linux_64_ubuntu11.10-1.run sudo ./cuda_5.0.35_linux_64_ubuntu11.10-1.run --override sudo gedit ~/.bashrc sudo nano /usr/local/cuda/include/host_config.h
line: 115 comment out error //#error -- unsupported GNU version! gcc versions later than 4.9 are not supported!
rsync -av /usr/local/cuda/samples ]
I need your help to check whether cuda is connected with freesurfer. How to check this.
Thanks Ramesh
On Mon, May 22, 2017 at 2:51 PM, R Edgar freesurfer.rge@gmail.com wrote:
On 21 May 2017 at 23:40, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
Thank you for your reply. I tried to compile by looking FAQ about CUDA in Ubuntu Ask, but it is get corrupted and I am forced to uninstall both
NVIDIA
driver and CUDA. I need help to compile it. Can I get some guidance or manual?
How do you mean corrupted?
I have found it necessary to edit the configure script which is generated, and to manually pass it the path to nvcc, but I don't believe I've seen something get corrupted before.
Regards,
Richard _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
On 22 May 2017 at 08:44, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
After compiling in the same way as given in the webpage, I couldn't reboot. It was showing a message that "running low graphic mode". It was not opening the desktop. Then through ctrl+alt+f1, I uninstalled nvida and cuda. Then I could boot the system.
Finally I have installed both cuda and nvidia after going through many web sources. I have installed NVIDIA drive through additional driver option in Ubuntu software & update centre. It installed recommended nvidia driver.
For cuda installation I followed these scripts, [chmod +x cuda_5.0.35_linux_64_ubuntu11.10-1.run sudo ./cuda_5.0.35_linux_64_ubuntu11.10-1.run --override sudo gedit ~/.bashrc sudo nano /usr/local/cuda/include/host_config.h
line: 115 comment out error //#error -- unsupported GNU version! gcc versions later than 4.9 are not supported!
rsync -av /usr/local/cuda/samples ]
I need your help to check whether cuda is connected with freesurfer. How to check this.
Compiling freesurfer should not result in a non-booting computer.
I'm afraid I wasn't using CUDA around v5. On my CentOS box, I installed the SDK using yum (i.e. rpm), and it upgraded itself to CUDA 8.0 when that came out.
Is there some reason you can't go to CUDA 8.0?
Richard
Dear Richard,
Ok. I will try cuda 8.0 and let you know the result. Should I install cuda 5 before installing cuda 8? or Can I run both?
Thanks Ramesh
On Tue, May 23, 2017 at 1:54 AM, R Edgar freesurfer.rge@gmail.com wrote:
On 22 May 2017 at 08:44, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
After compiling in the same way as given in the webpage, I couldn't
reboot.
It was showing a message that "running low graphic mode". It was not
opening
the desktop. Then through ctrl+alt+f1, I uninstalled nvida and cuda.
Then I
could boot the system.
Finally I have installed both cuda and nvidia after going through many
web
sources. I have installed NVIDIA drive through additional driver option in Ubuntu software & update centre. It installed recommended nvidia driver.
For cuda installation I followed these scripts, [chmod +x cuda_5.0.35_linux_64_ubuntu11.10-1.run sudo ./cuda_5.0.35_linux_64_ubuntu11.10-1.run --override sudo gedit ~/.bashrc sudo nano /usr/local/cuda/include/host_config.h
line: 115 comment out error //#error -- unsupported GNU version! gcc versions later than 4.9 are not supported!
rsync -av /usr/local/cuda/samples ]
I need your help to check whether cuda is connected with freesurfer. How
to
check this.
Compiling freesurfer should not result in a non-booting computer.
I'm afraid I wasn't using CUDA around v5. On my CentOS box, I installed the SDK using yum (i.e. rpm), and it upgraded itself to CUDA 8.0 when that came out.
Is there some reason you can't go to CUDA 8.0?
Richard _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
On 23 May 2017 at 01:13, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
Ok. I will try cuda 8.0 and let you know the result. Should I install cuda 5 before installing cuda 8? or Can I run both?
You'll only be able to have one version of the CUDA driver, but you should be able to have multiple CUDA SDKs.
Richard
Dear Richard, Now I am getting confusion, really which version of cuda I have?. Because when I give these command "cudadetect" nvcc -V, and cat /usr/local/cuda/version.txt I am getting different answers.
rb@rameshbabu:~$ cudadetect Detecting CUDA... There is 1 device supporting CUDA:
Device 0: "GeForce GT 610" CUDA Driver Version: 8.0 CUDA Runtime Version: 5.0 CUDA Capability Major revision number: 2 CUDA Capability Minor revision number: 1 Total amount of global memory: 2080440320 bytes Number of multiprocessors: 1 Number of cores: 32 Total amount of constant memory: 65536 bytes Total amount of shared memory per block: 49152 bytes Total number of registers available per block: 32768 Warp size: 32 Maximum number of threads per block: 1024 Maximum sizes of each dimension of a block: 1024 x 1024 x 64 Maximum sizes of each dimension of a grid: 65535 x 65535 x 65535 Maximum memory pitch: 2147483647 bytes Texture alignment: 512 bytes Clock rate: 1.62 GHz Concurrent copy and execution: Yes Run time limit on kernels: Yes Integrated: No Support host page-locked memory mapping: Yes Compute mode: Default (multiple host threads can use this device simultaneously)
rb@rameshbabu:~$ cat /usr/local/cuda/version.txt cat: /usr/local/cuda/version.txt: No such file or directory
rb@rameshbabu:~$ nvcc -V The program 'nvcc' is currently not installed. You can install it by typing: sudo apt install nvidia-cuda-toolkit
The first command reports that I have CUDA Driver Version: 8.0 and CUDA Runtime Version: 5.0, but why I am not getting any report for other two commands.
Please provide a solution.
Thanks Ramesh
On Tue, May 23, 2017 at 3:38 PM, R Edgar freesurfer.rge@gmail.com wrote:
On 23 May 2017 at 01:13, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
Ok. I will try cuda 8.0 and let you know the result. Should I install
cuda 5
before installing cuda 8? or Can I run both?
You'll only be able to have one version of the CUDA driver, but you should be able to have multiple CUDA SDKs.
Richard _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
On 24 May 2017 at 00:36, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
Now I am getting confusion, really which version of cuda I have?. Because when I give these command "cudadetect" nvcc -V, and cat /usr/local/cuda/version.txt I am getting different answers.
rb@rameshbabu:~$ cudadetect Detecting CUDA... There is 1 device supporting CUDA:
Device 0: "GeForce GT 610" CUDA Driver Version: 8.0 CUDA Runtime Version: 5.0 CUDA Capability Major revision number: 2 CUDA Capability Minor revision number: 1 Total amount of global memory: 2080440320 bytes Number of multiprocessors: 1 Number of cores: 32 Total amount of constant memory: 65536 bytes Total amount of shared memory per block: 49152 bytes Total number of registers available per block: 32768 Warp size: 32 Maximum number of threads per block: 1024 Maximum sizes of each dimension of a block: 1024 x 1024 x 64 Maximum sizes of each dimension of a grid: 65535 x 65535 x 65535 Maximum memory pitch: 2147483647 bytes Texture alignment: 512 bytes Clock rate: 1.62 GHz Concurrent copy and execution: Yes Run time limit on kernels: Yes Integrated: No Support host page-locked memory mapping: Yes Compute mode: Default (multiple host threads can use this device simultaneously)
cudadetect is installed by freesurfer. For me, the output starts: [rge21@cudastation gpu_utils]$ ./cudadetect Detecting CUDA... There is 1 device supporting CUDA:
Device 0: "Quadro K1200" CUDA Driver Version: 8.0 CUDA Runtime Version: 8.0 CUDA Capability Major revision number: 5 CUDA Capability Minor revision number: 0
rb@rameshbabu:~$ cat /usr/local/cuda/version.txt cat: /usr/local/cuda/version.txt: No such file or directory
rb@rameshbabu:~$ nvcc -V The program 'nvcc' is currently not installed. You can install it by typing: sudo apt install nvidia-cuda-toolkit
Did you install the CUDA SDK? The response from your OS implies you didn't (likewise the absence of /usr/local/cuda). However, if you did, then check your $PATH.
From my machine:
[rge21@cudastation gpu_utils]$ which nvcc /usr/local/cuda-8.0/bin/nvcc [rge21@cudastation gpu_utils]$ nvcc --version nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2016 NVIDIA Corporation Built on Tue_Jan_10_13:22:03_CST_2017 Cuda compilation tools, release 8.0, V8.0.61
The first command reports that I have CUDA Driver Version: 8.0 and CUDA Runtime Version: 5.0, but why I am not getting any report for other two commands.
Did you install CUDA yourself? What is the output from "locate libcudart" ?
Regards,
Richard
Dear Richard,
There is some response for cudadetect command which I have pated in earlier mail. But no ./cudadetect file.
rb@rameshbabu:~$ ./cudadetect bash: ./cudadetect: No such file or directory
I have nvcc file in /usr/local/cuda/bin but still not located.
rb@rameshbabu:~$ which nvcc (no response) rb@rameshbabu:~$ ll /usr/local/cuda/bin/ total 31836 drwxr-xr-x 3 root root 4096 May 19 19:56 ./ drwxr-xr-x 16 root root 4096 May 19 19:56 ../ -rwxr-xr-x 1 root root 63288 May 19 19:56 bin2c* lrwxrwxrwx 1 root root 4 May 19 19:56 computeprof -> nvvp* drwxr-xr-x 2 root root 4096 May 19 19:56 crt/ -rwxr-xr-x 1 root root 3240648 May 19 19:56 cudafe* -rwxr-xr-x 1 root root 2904184 May 19 19:56 cudafe++* -rwxr-xr-x 1 root root 5698534 May 19 19:56 cuda-gdb* -rwxr-xr-x 1 root root 170664 May 19 19:56 cuda-memcheck* -rwxr-xr-x 1 root root 6460832 May 19 19:56 cuobjdump* -rwxr-xr-x 1 root root 164992 May 19 19:56 fatbin* -rwxr-xr-x 1 root root 125584 May 19 19:56 fatbinary* -rwxr-xr-x 1 root root 58680 May 19 19:56 filehash* -rwxr-xr-x 1 root root 210 May 19 19:56 nsight* -rwxr-xr-x 1 root root 160576 May 19 19:56 nvcc* -rw-r--r-- 1 root root 306 May 19 19:56 nvcc.profile -rwxr-xr-x 1 root root 5729960 May 19 19:56 nvlink* -rwxr-xr-x 1 root root 2166448 May 19 19:56 nvprof* -rwxr-xr-x 1 root root 206 May 19 19:56 nvvp* -rwxr-xr-x 1 root root 5600776 May 19 19:56 ptxas*
rb@rameshbabu:~$ nvcc --version The program 'nvcc' is currently not installed. You can install it by typing: sudo apt install nvidia-cuda-toolkit
rb@rameshbabu:~$ locate libcudart /usr/local/MATLAB/R2017a/bin/glnxa64/libcudart.so.8.0 /usr/local/MATLAB/R2017a/bin/glnxa64/libcudart.so.8.0.44 /usr/local/cuda-5.0/lib/libcudart.so /usr/local/cuda-5.0/lib/libcudart.so.5.0 /usr/local/cuda-5.0/lib/libcudart.so.5.0.35 /usr/local/cuda-5.0/lib64/libcudart.so /usr/local/cuda-5.0/lib64/libcudart.so.5.0 /usr/local/cuda-5.0/lib64/libcudart.so.5.0.35 /usr/local/freesurfer/MCRv80/bin/glnxa64/libcudart.so.4 /usr/local/freesurfer/MCRv80/bin/glnxa64/libcudart.so.4.2.9
As far as I know, freesurfer supports cuda-5 and it requires libcudart.so-5.0. That's why I have installed cuda-5. How did you enable freesurfer with cuda-8?
Thanks Ramesh
On Thu, May 25, 2017 at 5:42 AM, R Edgar freesurfer.rge@gmail.com wrote:
On 24 May 2017 at 00:36, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
Now I am getting confusion, really which version of cuda I have?. Because when I give these command "cudadetect" nvcc -V, and cat /usr/local/cuda/version.txt I am getting different answers.
rb@rameshbabu:~$ cudadetect Detecting CUDA... There is 1 device supporting CUDA:
Device 0: "GeForce GT 610" CUDA Driver Version: 8.0 CUDA Runtime Version: 5.0 CUDA Capability Major revision number: 2 CUDA Capability Minor revision number: 1 Total amount of global memory: 2080440320 bytes Number of multiprocessors: 1 Number of cores: 32 Total amount of constant memory: 65536 bytes Total amount of shared memory per block: 49152 bytes Total number of registers available per block: 32768 Warp size: 32 Maximum number of threads per block: 1024 Maximum sizes of each dimension of a block: 1024 x 1024 x 64 Maximum sizes of each dimension of a grid: 65535 x 65535 x 65535 Maximum memory pitch: 2147483647 bytes Texture alignment: 512 bytes Clock rate: 1.62 GHz Concurrent copy and execution: Yes Run time limit on kernels: Yes Integrated: No Support host page-locked memory mapping: Yes Compute mode: Default (multiple host threads can use this device simultaneously)
cudadetect is installed by freesurfer. For me, the output starts: [rge21@cudastation gpu_utils]$ ./cudadetect Detecting CUDA... There is 1 device supporting CUDA:
Device 0: "Quadro K1200" CUDA Driver Version: 8.0 CUDA Runtime Version: 8.0 CUDA Capability Major revision number: 5 CUDA Capability Minor revision number: 0
rb@rameshbabu:~$ cat /usr/local/cuda/version.txt cat: /usr/local/cuda/version.txt: No such file or directory
rb@rameshbabu:~$ nvcc -V The program 'nvcc' is currently not installed. You can install it by
typing:
sudo apt install nvidia-cuda-toolkit
Did you install the CUDA SDK? The response from your OS implies you didn't (likewise the absence of /usr/local/cuda). However, if you did, then check your $PATH.
From my machine:
[rge21@cudastation gpu_utils]$ which nvcc /usr/local/cuda-8.0/bin/nvcc [rge21@cudastation gpu_utils]$ nvcc --version nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2016 NVIDIA Corporation Built on Tue_Jan_10_13:22:03_CST_2017 Cuda compilation tools, release 8.0, V8.0.61
The first command reports that I have CUDA Driver Version: 8.0 and CUDA Runtime Version: 5.0, but why I am not getting any report for other two commands.
Did you install CUDA yourself? What is the output from "locate libcudart" ?
Regards,
Richard _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
On 25 May 2017 at 01:20, Ramesh Babu mgrameshbabu2013@gmail.com wrote:
There is some response for cudadetect command which I have pated in earlier mail. But no ./cudadetect file.
rb@rameshbabu:~$ ./cudadetect bash: ./cudadetect: No such file or directory
That's... not altogether surprising.
rb@rameshbabu:~$ locate libcudart /usr/local/MATLAB/R2017a/bin/glnxa64/libcudart.so.8.0 /usr/local/MATLAB/R2017a/bin/glnxa64/libcudart.so.8.0.44 /usr/local/cuda-5.0/lib/libcudart.so /usr/local/cuda-5.0/lib/libcudart.so.5.0 /usr/local/cuda-5.0/lib/libcudart.so.5.0.35 /usr/local/cuda-5.0/lib64/libcudart.so /usr/local/cuda-5.0/lib64/libcudart.so.5.0 /usr/local/cuda-5.0/lib64/libcudart.so.5.0.35 /usr/local/freesurfer/MCRv80/bin/glnxa64/libcudart.so.4 /usr/local/freesurfer/MCRv80/bin/glnxa64/libcudart.so.4.2.9
Which one of those is LD_LIBRARY_PATH going to find?
As far as I know, freesurfer supports cuda-5 and it requires libcudart.so-5.0. That's why I have installed cuda-5. How did you enable freesurfer with cuda-8?
I compiled it myself after following NVIDIA's instructions for install the latest driver and SDK. Your machine was even telling you what to do in order to get nvcc (which would imply the presence of the SDK).
HTH,
Richard
freesurfer@nmr.mgh.harvard.edu
-
R Edgar -
Ramesh Babu