| # Nightly core benchmark subset. | |
| # One executable name per line. Lines starting with # are ignored. | |
| # These cover the most performance-critical dense linear algebra kernels. | |
| # BLAS-like operations | |
| bench_gemm | |
| bench_gemm_double | |
| bench_gemv | |
| bench_dot | |
| bench_vecadd | |
| bench_trsm | |
| bench_reductions | |
| # Decompositions | |
| bench_cholesky | |
| bench_cholesky_double | |
| bench_qr | |
| bench_svd | |
| bench_householder |