:cmake .. -DGMX_BUILD_OWN_FFTW=ON -DREGRESSIONTEST_DOWNLOAD=ON -DGMX_GPU=on -DGPU_DEPLOYMENT_KIT_ROOT_DIR=/usr/local/cuda-7.5 -DCMAKE_C_FLAGS="-D_FORCE_INLINES" and then make.None worked :/ Maybe I did something wrong? Regards,Diogo ----------------- From: Mark A
Hi Mark, I know, but at that moment I wasn't working in the same room as the CPU and I wanted to be sure on what I had to try next. I tried the cmake command and got the same message. I used grep -R -I "CUDA_NVCC_FLAGS" and grep -R -I "CMAKE_CXX_FLAGS" to find what files I could try to add "-D_FORC
to find the correct file to modify.
>
> Best regards,
> Diogo
>
>
> From: Mark Abraham <http://gmane.org/get-address.php?address=mark.j.abraham%2dRe5JQEeQqe8AvxtiuMwx3w%40public.gmane.org>
> >
>
> Subject: Re: [gmx-users] Error while re-compiling Gromacs 5.1.2 wit
o modify a CMakefile.txt or Makefile? If so, which? I've been looking around but can't seem to find the correct file to modify. Best regards,Diogo From: Mark Abraham <mark.j.abraham@...> Subject: Re: [gmx-users] Error while re-compiling Gromacs 5.1.2 with GP
Hi,
That's the standard content of CUDA - that deliberate compiler error has to
be suppressed (per the link in the original post) if you want to use a
different compiler.
The issue here is that the version of the C standard library that ships on
the operating system has a new feature that interac
Em 31-05-2016 17:25, Szilárd Páll escreveu:
Hi,
Just because gcc 5.3 and CUDA/nvcc 7.5 are in some Ubuntu repos (partner
AFAIR), it does not mean they're automatically compatible. The NVIDIA
documentation clearly indicates they're not.
Surely the ubuntu cuda maintainers patched 'include/host
On Tue, May 31, 2016 at 10:25 PM, Szilárd Páll wrote:
> Hi,
>
> Just because gcc 5.3 and CUDA/nvcc 7.5 are in some Ubuntu repos (partner
> AFAIR), it does not mean they're automatically compatible. The NVIDIA
> documentation clearly indicates they're not.
>
> However, this typically means that min
mpiler> (ICC)
> or Microsoft Visual C <https://en.wikipedia.org/wiki/Microsoft_Visual_C>
> Compiler, and sends the device code (the part which will run on the GPU) to
> the GPU. The device code is further compiled by NVCC.
>
> Regards
> Diogo
>
>
> From: Florent Hé
Hi,
Just because gcc 5.3 and CUDA/nvcc 7.5 are in some Ubuntu repos (partner
AFAIR), it does not mean they're automatically compatible. The NVIDIA
documentation clearly indicates they're not.
However, this typically means that minor issue can prevent building. In
this case, it looks like a glibc
un on the GPU) to the GPU. The device code is further compiled by NVCC. RegardsDiogo From: Florent Hédin <work@...> Subject: Re: [gmx-users] Error while re-compiling Gromacs 5.1.2 with GPU and gcc 5.3.1 Newsgroups: gmane.science.biology.gromacs.user Date: 2016-05-31 19:09:01 GMT (44 minutes ago)
Hi,
As far as I know cuda 7.5 only supports branch 4.X of gcc, not 5.X ;
what about forcing the use of gcc-4.9:
CC=gcc-4.9 CXX=g++-4.9 cmake [OTHER FLAGS] ..
Regards,
Florent Hédin
On 31/05/16 19:42, Diogo Martins de Sá wrote:
Hi Nikhil,
You can get CUDA toolkit 7.5 using apt-get install nv
Hi Nikhil, You can get CUDA toolkit 7.5 using apt-get install nvidia-cuda-dev nvidia-cuda-toolkit. Also, if you open the link I've sent, you`ll see the step on how to install directly from provider. NAMD ran with CUDA this way. So it is supported. Installing directly with apt-get gives the same res
Hi
as per my understanding, CUDA is not supported Ubuntu 16.04 from the
provider
--
Gromacs Users mailing list
* Please search the archive at
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
* For (un)
Hi guys, I am running kubuntu 16.04 and installed CUDA toolkit 7.5 like was done here: https://www.pugetsystems.com/labs/articles/NVIDIA-CUDA-with-Ubuntu-16-04-beta-on-a-laptop-if-you-just-cannot-wait-775/ I used cmake with the following command line: cmake .. -DGMX_BUILD_OWN_FFTW=ON -DREGRESSIONTE
14 matches
Mail list logo
