- #Ansys 15 installation guide fortran compiler software#
- #Ansys 15 installation guide fortran compiler code#
The following modules are installed via Spack environments, areĬonsidered experimental, and are subject to change. More information on the XL compiler options is here: Level, add debug symbols, and block semantic changing optimizations: ForĮxample, the following flags will respectively reduce the optimization
Specifying flags on the command line will override these defaults. Please refer to documentation on the STRICT/NOSTRICT option for more information. If the compiler applies such an optimization the following warningġ500-036: (I) The NOSTRICT option (default at OPT(3)) has the potential to alter the semantics of a program. The default MPICH compiler wrapper flags -O3 -qipa -qhot will performĪggressive optimizations that could alter the semantics of your program.
#Ansys 15 installation guide fortran compiler code#
Nvcc is called with -ccbin $(CXX) otherwise linking will fail.ĬUDA code should be compiled with -arch=sm_70 for the Volta V100 GPUs. When mixing CUDA and MPI, please make sure an xl module is loaded and
Please pay carefulĪttention to which modules you have loaded.Ĭurrently the following are available as modules: Those packages can be overridden by modules. Some tools are available without loading any modules and a subset of
#Ansys 15 installation guide fortran compiler software#
Many packages for building software are available as modules. Note, to measure performance of your application it is strongly suggested that you disable node sharing with other jobs by passing -gres=gpu:6 (or -gres=gpu:16g:4 on the nodes with four GPUs) to Slurm. The flops/cycle/core is from Dirk Pleiter's SC18 tutorial "IBM POWER9 Processor,NVIDIA V100 GPU andIBM AC922 Node Hardware Architecture": These results are from Nvidia Tesla V100 DataĬombined, the two Power9 CPUs on an AiMOS node can perform 1 TeraFLOPs in double precision.Ģ sockets * 20 cores/socket * 8 flops/cycle/core * 3.15 giga-cycles/second ~= 1 TeraFLOPs double precision