CMake config for IntelLLVM (ifx) and NVIDIA HPC SDK (nvfortran)

[ACSimon33]

Description

Updated the CMake config to set the correct compiler flags for IntelLLVM (ifx) and NVIDIA HPC SDK (nvfortran). I tested ifx on Linux (oneAPI 2023.2.0) and Windows (oneAPI 2023.1.0). There are still some errors but way less than before (see below).

nvfortran (NVIDIA HPC SDK 23.7) is only available on Linux and it works alright. I didn't fix / silence the compiler warnings for now. There are also some numerical errors (see below).

ifx test summary:

                        -->   LAPACK TESTING SUMMARY  <--
                Processing LAPACK Testing output found in the TESTING directory
SUMMARY                 nb test run     numerical error         other error  
================        ===========     =================       ================  
REAL                    1328283         0       (0.000%)        0       (0.000%)
DOUBLE PRECISION        1329105         0       (0.000%)        0       (0.000%)
COMPLEX                 608384          8012    (1.317%)        8880    (1.460%)
COMPLEX16               608707          8081    (1.328%)        8880    (1.459%)

--> ALL PRECISIONS      3874479         16093   (0.415%)        17760   (0.458%)

nvfortran test summary:

                        -->   LAPACK TESTING SUMMARY  <--
                Processing LAPACK Testing output found in the TESTING directory
SUMMARY                 nb test run     numerical error         other error  
================        ===========     =================       ================  
REAL                    1100967         317     (0.029%)        0       (0.000%)
DOUBLE PRECISION        1255005         280     (0.022%)        0       (0.000%)
COMPLEX                 652783          132     (0.020%)        0       (0.000%)
COMPLEX16               694202          287     (0.041%)        0       (0.000%)

--> ALL PRECISIONS      3702957         1016    (0.027%)        0       (0.000%)

Some of the errors are most likely due to non-conforming handling of INF and NAN as both compilers fail some of the Fortran compiler tests (-DTEST_FORTRAN_COMPILER=ON). We are already working with Intel to resolve the errors in ifx.

Simon, NAG Ltd.

[codecov]

Codecov Report

Patch and project coverage have no change.

Comparison is base (2983298) 0.00% compared to head (055a0e1) 0.00%.
Report is 2 commits behind head on master.

❗ Current head 055a0e1 differs from pull request most recent head 64d95ed. Consider uploading reports for the commit 64d95ed to get more accurate results

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #898   +/-   ##
=======================================
  Coverage    0.00%    0.00%           
=======================================
  Files        1918     1918           
  Lines      188614   188614           
=======================================
  Misses     188614   188614           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[weslleyspereira]

Thanks @ACSimon33 !