Adam Siemieniuk
ec450b1900
[mlir][xegpu] Allow out-of-bounds writes ( #110811 )
...
Relaxes vector.transfer_write lowering to allow out-of-bound writes.
This aligns lowering with the current hardware specification which does
not update bytes in out-of-bound locations during block stores.
2024-10-09 18:59:14 +02:00
Adam Siemieniuk
6c25604df2
[mlir][xegpu] Convert Vector load and store to XeGPU ( #110826 )
...
Adds patterns to lower vector.load|store to XeGPU operations.
2024-10-03 08:59:39 +02:00
Kazu Hirata
b52885bc23
[mlir] Use std::optional::value_or (NFC) ( #109893 )
2024-09-26 09:53:43 -07:00
Chao Chen
8b5e841487
[MLIR][XeGPU] Updates XeGPU TensorDescAttr and Refine Gather/Scatter definition ( #109675 )
...
Bring back #109144 with fixes to VectorToXeGPU
2024-09-24 10:14:13 -05:00
Adam Siemieniuk
02d34d800b
[mlir][vector][xegpu] Vector to XeGPU conversion pass ( #107419 )
...
Add pass for Vector to XeGPU dialect conversion and initial conversion
patterns for vector.transfer_read|write operations.
2024-09-19 15:16:23 -05:00
