h-murai
h-murai
We need the following three more patches to build NWChem 7.0.2 with Fujitsu compiler. ``` diff -u -r nwchem-7.0.2/src/nwdft/rt_tddft/rtutils/rt_tddft_spatial_potential.F nwchem-7.0.2-patched/src/nwdft/r t_tddft/rtutils/rt_tddft_spatial_potential.F --- nwchem-7.0.2/src/nwdft/rt_tddft/rtutils/rt_tddft_spatial_potential.F 2020-10-13 13:16:26.000000000 +0 900 +++ nwchem-7.0.2-patched/src/nwdft/rt_tddft/rtutils/rt_tddft_spatial_potential.F 2022-02-09...
I'm no sure, but there is no such version.
With the patch I posted above, we successfully compiled NWChem with Fujitsu compiler. But, as @william-dawson said, we actually need more debugging efforts to run it properly on A64FX.
The workaround @martin-frbg proposed doesn't work because the wrapper internally pass the `-mcpu` flag to gcc. When I commented out a line `preextend flags_list SPACK_TARGET_ARGS` from the wrapper (`lib/spack/env/gcc/gcc`), the...
Building tensorboard_data_server for aarch64 (with Spack) fails with the following error message. I think it's because `manylinux2010` doesn't support aarch64. Could you fix this in the near future? ``` ERROR:...