[XeGPU] SLM support for scattered tensors.
Open
- Dominant language
- MLIR
- Stars
- 156
- Forks
- 45
- Avg merge
- 4h 14m
- Merged PRs (30d)
- 22
Description
Currently, scattered tensors in slm are not supported for load/store ops:
```#xegpu.tdesc_attr>```
They can only load/store to global memory (ugm), as indicated by the hardcoded `sfid` argument of the intrinsic call which is set to 15. Changing it to 14 should allow loads/stores to be performed upon shared local memory (slm).
Contributor guide
Assessment
This issue has not been assessed yet.