[Mlir-commits] [mlir] [MLIR][Linalg] Remove unary named ops (PR #220905)

[Renato Golin]

================
@@ -190,61 +160,12 @@ func.func @unary_ops(%A: tensor<?x?x?xf32>, %Out: tensor<?x?x?xf32>) -> tensor<?
 // ALL-LABEL: unary_ops
 // ALL-SAME: %[[A:.+]]: tensor<?x?x?xf32>, %[[OUT:.+]]: tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
 
-// NAMED-NOT: linalg.generic
-// NAMED: %[[RES0:.+]] = linalg.exp
-// NAMED-SAME: ins(%[[A]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES1:.+]] = linalg.log
-// NAMED-SAME: ins(%[[RES0]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES2:.+]] = linalg.abs
-// NAMED-SAME: ins(%[[RES1]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES3:.+]] = linalg.ceil
-// NAMED-SAME: ins(%[[RES2]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES4:.+]] = linalg.floor
-// NAMED-SAME: ins(%[[RES3]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES5:.+]] = linalg.negf
-// NAMED-SAME: ins(%[[RES4]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES6:.+]] = linalg.reciprocal
-// NAMED-SAME: ins(%[[RES5]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES7:.+]] = linalg.round
-// NAMED-SAME: ins(%[[RES6]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES8:.+]] = linalg.sqrt
-// NAMED-SAME: ins(%[[RES7]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES9:.+]] = linalg.rsqrt
-// NAMED-SAME: ins(%[[RES8]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES10:.+]] = linalg.square
-// NAMED-SAME: ins(%[[RES9]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES11:.+]] = linalg.tanh
-// NAMED-SAME: ins(%[[RES10]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// NAMED: %[[RES12:.+]] = linalg.erf
-// NAMED-SAME: ins(%[[RES11]] : tensor<?x?x?xf32>)
-// NAMED-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-
-// CATEGORY: %[[RES0:.+]] = linalg.elementwise kind=#linalg.elementwise_kind<exp>
-// CATEGORY-SAME: ins(%[[A]] : tensor<?x?x?xf32>)
-// CATEGORY-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// CATEGORY: %[[RES1:.+]] = linalg.elementwise kind=#linalg.elementwise_kind<log>
-// CATEGORY-SAME: ins(%[[RES0]] : tensor<?x?x?xf32>)
-// CATEGORY-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// CATEGORY: %[[RES2:.+]] = linalg.elementwise kind=#linalg.elementwise_kind<abs>
-// CATEGORY-SAME: ins(%[[RES1]] : tensor<?x?x?xf32>)
-// CATEGORY-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
-// CATEGORY: %[[RES3:.+]] = linalg.elementwise kind=#linalg.elementwise_kind<ceil>
-// CATEGORY-SAME: ins(%[[RES2]] : tensor<?x?x?xf32>)
-// CATEGORY-SAME: outs(%[[OUT]] : tensor<?x?x?xf32>) -> tensor<?x?x?xf32>
+// No unary linalg named ops remain, so generic-to-named leaves them as generics.
+// NAMED-NOT: linalg.elementwise
+// NAMED: linalg.generic
----------------
rengolin wrote:

Yes, but this is a request to get named ops, not category, this is why it remains generic. There's a discussion if we want to do that or not, but this is not for this PR. #221266

https://github.com/llvm/llvm-project/pull/220905