use named op for matmul

opencompl · Nov 4, 2023 · 605b9cd · 605b9cd
1 parent 2f43517
commit 605b9cd
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 7 deletions.
diff --git a/kernels/matmul/16x16xf64/linalg.csv b/kernels/matmul/16x16xf64/linalg.csv
@@ -0,0 +1 @@
+49320
diff --git a/kernels/matmul/16x16xf64/linalg.mlir b/kernels/matmul/16x16xf64/linalg.mlir
@@ -1,12 +1,6 @@
 func.func public @matmul(%X: memref<16x16xf64>,
                          %Y: memref<16x16xf64>,
                          %Z: memref<16x16xf64>) {
-  "linalg.generic"(%X, %Y, %Z) ({
-  ^bb0(%x: f64, %y: f64, %z: f64):
-    %r0 = arith.mulf %x, %y : f64
-    %r1 = arith.addf %z, %r0 : f64
-    "linalg.yield"(%r1) : (f64) -> ()
-  }) {indexing_maps = [affine_map<(m, n, k) -> (m, k)>, affine_map<(m, n, k) -> (k, n)>, affine_map<(m, n, k) -> (m,
-  n)>], iterator_types = [#linalg.iterator_type<parallel>, #linalg.iterator_type<parallel>, #linalg.iterator_type<reduction>], operand_segment_sizes = array<i32: 2, 1>} : (memref<16x16xf64>, memref<16x16xf64>, memref<16x16xf64>) -> ()
+  linalg.matmul ins(%X, %Y : memref<16x16xf64>, memref<16x16xf64>) outs(%Z : memref<16x16xf64>) -> ()
   func.return
 }