Fallout from latest staging merge (gcc-15)

[markuskowa]

The update to gcc-15 in the latest nixpkgs staging update has led to broken packages:
(listing all packages, even if the problem is upstream)

• amd-scalapack
• apbs (apbs: fix build NixOS/nixpkgs#483075)
• crest
• gabedit
• libxsmm
• molcas
• molden
• mrchem (mark broken)
• nwchem
• pcmsolver
• python3.qmcpack (mark broken)
• python311.vmd-python
• openbabel (openbabel: disable failing tests NixOS/nixpkgs#482731)
• tblite
• molcasWfa (is strange case where cflags for moclas and wfa colide).
• wannier90

CC @sheepforce

[sheepforce]

• molden: molden: ensure gcc15 compatibility NixOS/nixpkgs#479318
• tblite: Compiles fine on nixpkgs/master
• molcas: openmolcas: ensure gcc15 compatibility NixOS/nixpkgs#479325
• nwchem: nwchem: ensure gcc15 compatibility NixOS/nixpkgs#479352
• wannier90: wannier90: relax numerical thresholds NixOS/nixpkgs#481917
• libxsmm: libxsmm: ensure gcc15 compatibility NixOS/nixpkgs#479357
• pcmsolver: pcmsolver: ensure gcc15 compatibility NixOS/nixpkgs#479365

[sheepforce]

python311.vmd-python

3.12 is also broken. One issue is python3.pkgs.jaraco-test upstream, which only works with python 3.13+. I consider marking vmd-python as broken until they fix their python version compatibility upstream.

[markuskowa]

wannier90 might be one of those dynamic CPU detection issues or an optimization issue. It fails on my AMD EPYC:

tests/testw90_example11_2 - silicon.win: **FAILED**.
final_spreads
    ERROR: absolute error 3.59e-06 greater than 3.00e-06. (Test: 7.20268747.  Benchmark: 7.20269106.)

[sheepforce]

Hm, that is funny, I'm running this on a ThreadRipper 7945WX. I assume they have a quite similar instruction set? Mine reports

$ grep "flags" /proc/cpuinfo | tail -n 1
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good amd_lbr_v2 nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba perfmon_v2 ibrs ibpb stibp ibrs_enhanced vmmcall fsgsbase bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local user_shstk avx512_bf16 clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin cppc arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic vgif x2avic v_spec_ctrl vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq la57 rdpid overflow_recov succor smca fsrm flush_l1d debug_swap

I guess there are two options: disable the tests altogether or I'm going through the code to look for that threshold and increase it slightly.

[sheepforce]

Wannier90 has quite a lot of thresholds for the tests. I'm afraid if I change one of them, where 3.0e-6 is most likely that one here https://github.com/wannier-developers/wannier90/blob/develop/test-suite/tests/userconfig#L12, another one may fail on another CPU 🙄 With some crazy patching magic we could probably all multiply by a factor of 2 or something, but that's probably not the best way forward.

[sheepforce]

Wannie90 with relaxed numerical thresholds: NixOS/nixpkgs#481917

[markuskowa]

Ok, I switched wannier90 to vanialla stdenv.