Sami Remes b619c374eb [rocm-libraries] ROCm/rocm-libraries#5438 (commit 7000562) ...

[CK_TILE] Normalize gpu_target before LDS_SIZE_MAP lookup
 (#5438)

GPU targets passed with feature suffixes (e.g. `gfx950:xnack+`) were
falling through to `DEFAULT_LDS_SIZE` instead of matching their entry in
`LDS_SIZE_MAP`, potentially causing incorrect tile acceptance/rejection.

## Changes

- **`gemm_validation_utils.py`**: Strip everything after `:` from
`gpu_target` before the `LDS_SIZE_MAP` lookup; use the normalized base
arch name in the error message as well.

```python
# Before
hw_lds_size = LDS_SIZE_MAP.get(gpu_target, DEFAULT_LDS_SIZE)

# After
base_gpu_target = gpu_target.split(":")[0] if gpu_target else gpu_target
hw_lds_size = LDS_SIZE_MAP.get(base_gpu_target, DEFAULT_LDS_SIZE)
```

2026-05-29 16:33:15 +00:00

..

include

feat(precommit-hooks): add check for correct copyright header (#3302)

2025-12-10 22:50:43 -08:00

ops

[rocm-libraries] ROCm/rocm-libraries#5438 (commit 7000562)

2026-05-29 16:33:15 +00:00

sampling

[rocm-libraries] ROCm/rocm-libraries#7311 (commit 79d8cae)

2026-05-21 02:17:42 -05:00

CMakeLists.txt

[rocm-libraries] ROCm/rocm-libraries#7311 (commit 79d8cae)

2026-05-21 02:17:42 -05:00

operation_support_matrix.md

[rocm-libraries] ROCm/rocm-libraries#5260 (commit a1834d2)

2026-05-17 00:29:40 -07:00