mirror-llvm-project/mlir/test/Dialect/X86
Arun Thangamani b32d4f0c45
[mlir][x86] Decouple Accumulator from ADD-Based Vector Contractions (over loops) (#204327)
This path transforms `vector.contract(A, B, Acc)` into
`vector.contract(A, B, 0) + Acc` to decouple the contraction computation
from the accumulator update for contractions over the loop.
2026-08-02 07:33:00 +05:30
..
AMX [mlir][x86] Adds extra pattern for AMX lowering (#213228) 2026-07-31 20:25:06 +05:30
cvt-packed-f32-to-bf16.mlir
dot-bf16.mlir
legalize-for-llvm.mlir
move-acc-for-contract-loop.mlir [mlir][x86] Decouple Accumulator from ADD-Based Vector Contractions (over loops) (#204327) 2026-08-02 07:33:00 +05:30
roundtrip.mlir
shuffle-vector-fmas.mlir
sink-vector-producer-ops.mlir
vector-contract-bf16-to-fma.mlir [mlir][x86] Decouple Accumulator from ADD-Based Vector Contractions (over loops) (#204327) 2026-08-02 07:33:00 +05:30
vector-contract-to-fma.mlir
vector-contract-to-packed-type-dotproduct.mlir [mlir][x86] Decouple Accumulator from ADD-Based Vector Contractions (over loops) (#204327) 2026-08-02 07:33:00 +05:30