pytorch
0a4a7d4b - Use hipsolver for default svd case on ROCm (#103540)

Commit
1 year ago
Use hipsolver for default svd case on ROCm (#103540) Fixes #102678 Fixes #102629 Fixes #102558 HipSOLVER performance on ROCm5.4.2 and later no longer serves as massive bottleneck. Additionally, using magma on rocm in this case caused test_compare_cpu_lialg_pinv_singular_cuda_float32 to fail. Using hipSOLVER, the test now passes. Pull Request resolved: https://github.com/pytorch/pytorch/pull/103540 Approved by: https://github.com/lezcano
Author
Committer
Parents
Loading