llvm-project

18ecdbfe - [mlir][vector-to-gpu]: Lower transposed strided transfer_read

[mlir][vector-to-gpu]: Lower transposed strided transfer_read Add support for lowering vector.transfer_read to gpu.subgroup_mma_load_matrix with transpose permutation_map with non-minor dimensions e.g. (d0, d1, d2) -> (d2, d0)