Okay, I managed to get an older version of gcc installed using anaconda
(gcc7.3.0), however, the same error is still occurring.
As shown in the last command below, I can verify using 'which gcc' that the
version being used is older than version 8.
make[3]: Leaving directory `/home/wesley/amber18/AmberTools/src/arpack'
> /cm/shared/apps/cuda10.2/toolkit/10.2.89/bin/nvcc -gencode
> arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode
> arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode
> arch=compute_52,code=sm_52 -gencode arch=compute_53,code=sm_53 -gencode
> arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode
> arch=compute_60,code=sm_70, -gencode arch=compute_61,code=sm_70
> -Wno-deprecated-declarations -use_fast_math -O3 -ccbin g++ -o
> cuda_mg_wrapper.o -c cuda_mg_wrapper.cu
> In file included from
> /cm/shared/apps/cuda10.2/toolkit/10.2.89/bin/../targets/x86_64-linux/include/cuda_runtime.h:83,
> from <command-line>:
> /cm/shared/apps/cuda10.2/toolkit/10.2.89/bin/../targets/x86_64-linux/include/crt/host_config.h:138:2:
> error: #error -- unsupported GNU version! gcc versions later than 8 are not
> supported!
> 138 | #error -- unsupported GNU version! gcc versions later than 8 are
> not supported!
> | ^~~~~
> make[2]: *** [cuda_mg_wrapper.o] Error 1
> make[2]: Leaving directory `/home/wesley/amber18/AmberTools/src/pbsa'
> make[1]: *** [cuda_serial] Error 2
> make[1]: Leaving directory `/home/wesley/amber18/AmberTools/src'
> make: *** [install] Error 2
> (base) [wesley.bright90 amber18]$ which gcc
> ~/bin/anaconda3/libexec/gcc/x86_64-conda_cos6-linux-gnu/7.3.0/gcc
On Thu, Oct 22, 2020 at 1:02 PM David A Case <david.case.rutgers.edu> wrote:
> On Thu, Oct 22, 2020, Wes Smith wrote:
>
> >> error: #error -- unsupported GNU version! gcc versions later than 8 are
> not
> >> supported!
>
> This message comes from NVIDIA, not from any Amber code. You can either
> install gcc8 and use that, or move to CUDA 11. See:
>
> https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html
>
> Note: please don't try gcc10 (yet): there are a few pieces of AmberTools
> that don't work with gcc10.
>
> Second note: upgrading to CUDA 11 won't help if you need amber18 GPU
> code; you'll have to downgrade gcc. Amber20 is fine with CUDA 11.
>
> >
> >Anyone know how to fix this error? I would perfer not having to install a
> >separate, older, version of GCC just to run amber18 if possible as this
> >could lead to a lot of complications since all our system library files
> >seem to be compiled using the newer gcc...
>
> You don't need to install gcc8 on the system, or even have root
> privileges: just put gcc8 somewhere (say in a folder under your home
> directory). Then create a ~/gcc8bin folder that links gfortran, gcc and
> g++ to the gcc8 binaries. Finally, put ~/gcc8bin at the start of your
> PATH. Your amber build will then use gcc8. Remove ~/gcc8bin from your
> PATH once you are done, and you'll be back to gcc9.
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Oct 23 2020 - 21:00:02 PDT
