https://bugs.llvm.org/show_bug.cgi?id=49752

            Bug ID: 49752
           Summary: Offload kernel performance regression
           Product: OpenMP
           Version: unspecified
          Hardware: PC
                OS: Linux
            Status: CONFIRMED
          Severity: enhancement
          Priority: P
         Component: Clang Compiler Support
          Assignee: [email protected]
          Reporter: [email protected]
                CC: [email protected]

both 12.x and main branch are affected by this bug.
offending commit
bd756286d2e774716a12f55db27d070595058d53

before
nvlink info    : Function properties for
'__omp_offloading_10304_1c00a0d__ZN11qmcplusplus17einspline_spo_ompIfE18multi_evaluate_vghERKSt6vectorIPNS_6SPOSetESaIS4_EERKS2_IPNS_11ParticleSetESaISA_EEi_l412':
nvlink info    : used 244 registers, 80 stack, 1134 bytes smem, 416 bytes
cmem[0], 24 bytes cmem[2], 0 bytes lmem
after
nvlink info    : Function properties for
'__omp_offloading_10304_1c00a0d__ZN11qmcplusplus17einspline_spo_ompIfE18multi_evaluate_vghERKSt6vectorIPNS_6SPOSetESaIS4_EERKS2_IPNS_11ParticleSetESaISA_EEi_l412':
nvlink info    : used 244 registers, 360 stack, 1134 bytes smem, 416 bytes
cmem[0], 24 bytes cmem[2], 0 bytes lmem

huge increase in stack.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs

Reply via email to