[Gate-users] Gate-users Digest, Vol 127, Issue 15

[Edwin Leung]

Hi Reza,

Please take a look at the following link as the proposed solution in the
link may solve your problem. It solved mine.

http://lists.opengatecollaboration.org/mailman/private/gate-users/
2016-June/009003.html

I solved the problem by modifying and recompiling the GATE source code. In
the source code folder, you will see a file named CMakeLists.txt and inside
you will find a line with CUDA_NVCC_FLAGS and two "-gencode
arch=compute_XX,code=sm_XX;" where XX = 20 or 30 (which corresponds to
Compute Capability 2.0 or 3.0).

Compute Capability 2.0 and 3.0 are for GPUs that are ~5+ years old so your
GPU most likely will have a Compute Capability that is higher than those
two. To identify the Compute Capability of your GPU, please refer to the
Compute Capability you see when running CUDA example 5 or the link below:

https://developer.nvidia.com/cuda-gpus

I imagine that this is patched in the next GATE release since my pull
request has been accepted on the GitHub develop branch.

>From what I understand, CUDA doesn't come with SDK anymore. While the link
to the proposed solution (that I provided earlier) recommended we use the
same filepath for both CUDA_TOOLKIT_ROOT_DIR and CUDA_SDK_ROOT_DIR, the
example script worked with a blank CUDA_SDK_ROOT_DIR for me.

Hope that helps!

Sincerely,

Ed

On Wed, Dec 14, 2016 at 3:00 AM, <
gate-users-request at lists.opengatecollaboration.org> wrote:

> Send Gate-users mailing list submissions to
>         gate-users at lists.opengatecollaboration.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         http://lists.opengatecollaboration.org/mailman/listinfo/gate-users
> or, via email, send a message with subject or body 'help' to
>         gate-users-request at lists.opengatecollaboration.org
>
> You can reach the person managing the list at
>         gate-users-owner at lists.opengatecollaboration.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Gate-users digest..."
>
>
> Today's Topics:
>
>    1. GATE v7.2 GPU Compilation (M.R Teimoori)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Tue, 13 Dec 2016 10:13:40 -0600
> From: "M.R Teimoori" <reza.teimoori.s at gmail.com>
> To: gate-users <gate-users at lists.opengatecollaboration.org>
> Subject: [Gate-users] GATE v7.2 GPU Compilation
> Message-ID:
>         <CACEZ9i8X-Y5D4Jn4ihVkpBs-VXXTaWWdDqwP7gUuS2ixoM8iTA@
> mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Hi All,
>
> I'm trying to compile GATE v7.2 with the GPU flag on. My GPU is GeForce GTX
> Titan Black and I have CUDA 8.0 installed on Ubuntu 16.04. The default
> variables don't lead to a successful compilation. If anyone has done the
> compilation successfully, can you please let me know what exactly I need to
> set the following variables:
>
> CUDA_HOST_COMPILER = default set to '/usr/bin/cc'
> CUDA_SDK_ROOT_DIR = no default folder has been chosen - It says it couldn't
> be found
> CUDA_TOOLKIT_ROOT_DIR = default set to '/usr/local/cuda-8.0'
> CUDA_USE_STATIC_CUDA_RUNTIME = default is 'on'
> CUDA_rt_LIBRARY = default set to '/usr/lib/x86_64-linux-gnu/librt.so'
>
> Thank you,
>
> Cheers,
> Reza
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://lists.opengatecollaboration.org/mailman/private/gate-users/
> attachments/20161213/1f4633bb/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> Gate-users mailing list
> Gate-users at lists.opengatecollaboration.org
> http://lists.opengatecollaboration.org/mailman/listinfo/gate-users
>
> ------------------------------
>
> End of Gate-users Digest, Vol 127, Issue 15
> *******************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opengatecollaboration.org/mailman/private/gate-users/attachments/20161214/6b98e98b/attachment.html>