mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-07-17 17:19:12 +00:00
Fix ThreadwiseTensorSliceTransfer_v4::Run (Fuse scale)
This commit is contained in:
@@ -1496,7 +1496,7 @@ struct ThreadwiseTensorSliceTransfer_v4
|
||||
}
|
||||
});
|
||||
}
|
||||
#if 0
|
||||
|
||||
// Fuse scale
|
||||
template <typename SrcRefToOriginDisplacement,
|
||||
typename DstOriginIdx,
|
||||
@@ -1696,7 +1696,7 @@ struct ThreadwiseTensorSliceTransfer_v4
|
||||
}
|
||||
});
|
||||
}
|
||||
#endif
|
||||
|
||||
template <typename SrcSliceMoveStepIdx>
|
||||
__device__ void MoveSrcSliceWindow(const SrcDesc&,
|
||||
const SrcSliceMoveStepIdx& src_slice_move_step_idx)
|
||||
|
||||
Reference in New Issue
Block a user