I am compiling a NAS Parallel Benchmark with OpenMP target directives. I am getting this warning from clang 9.0.0:
txinvr.c:50:3: warning: loop not vectorized: the optimizer was unable to perform the requested transformation; the transformation might be disabled or specified as part of an unsupported transformation ordering [-Wpass-failed=transform-warning]
#pragma omp teams distribute parallel for simd collapse(3) private(t1, t2, t3, ac, ru1, uu, vv, ww, r1, r2, r3, r4, r5, ac2inv)
Whenever the “simd” clause is being used. Is simd not supported for offload in clang?
Thanks and greetings,