hardmax moved to todo

mlir-assigner/tests/Ops/NeedsCompilerWork/Hardmax/HardMaxAxis.mlir

@@ -0,0 +1,50 @@
+module attributes {llvm.data_layout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128", llvm.target_triple = "x86_64-pc-linux-gnu", "onnx-mlir.symbol-postfix" = "hardmaxaxis.mlir"} {
+  func.func @main_graph(%arg0: memref<1x10x20x30xf32>) -> memref<1x10x20x30xf32> attributes {input_names = ["in_a"], llvm.emit_c_interface, output_names = ["out_a"]} {
+    %cst = arith.constant 0.000000e+00 : f32
+    %cst_0 = arith.constant 1.000000e+00 : f32
+    %c0 = arith.constant 0 : index
+    %alloc = memref.alloc() {alignment = 16 : i64} : memref<1x10x20x30xf32>
+    %alloc_1 = memref.alloc() {alignment = 16 : i64} : memref<1x10x20x30xindex>
+    affine.for %arg1 = 0 to 1 {
+      affine.for %arg2 = 0 to 10 {
+        affine.for %arg3 = 0 to 20 {
+          affine.for %arg4 = 0 to 30 {
+            affine.store %c0, %alloc_1[%arg1, %arg2, %arg3, %arg4] : memref<1x10x20x30xindex>
+          }
+        }
+      }
+    }
+    affine.for %arg1 = 0 to 1 {
+      affine.for %arg2 = 0 to 10 {
+        affine.for %arg3 = 0 to 20 {
+          affine.for %arg4 = 0 to 30 {
+            %0 = affine.load %alloc_1[%c0, %arg2, %arg3, %arg4] : memref<1x10x20x30xindex>
+            %1 = memref.load %arg0[%0, %arg2, %arg3, %arg4] : memref<1x10x20x30xf32>
+            %2 = affine.load %arg0[%arg1, %arg2, %arg3, %arg4] : memref<1x10x20x30xf32>
+            %3 = arith.cmpf ogt, %2, %1 : f32
+            scf.if %3 {
+              affine.store %arg1, %alloc_1[%c0, %arg2, %arg3, %arg4] : memref<1x10x20x30xindex>
+            }
+          }
+        }
+      }
+    }
+    affine.for %arg1 = 0 to 1 {
+      affine.for %arg2 = 0 to 10 {
+        affine.for %arg3 = 0 to 20 {
+          affine.for %arg4 = 0 to 30 {
+            %0 = affine.load %alloc_1[%c0, %arg2, %arg3, %arg4] : memref<1x10x20x30xindex>
+            %1 = arith.cmpi eq, %0, %arg1 : index
+            scf.if %1 {
+              affine.store %cst_0, %alloc[%arg1, %arg2, %arg3, %arg4] : memref<1x10x20x30xf32>
+            } else {
+              affine.store %cst, %alloc[%arg1, %arg2, %arg3, %arg4] : memref<1x10x20x30xf32>
+            }
+          }
+        }
+      }
+    }
+    return %alloc : memref<1x10x20x30xf32>
+  }
+  "krnl.entry_point"() {func = @main_graph, numInputs = 1 : i32, numOutputs = 1 : i32, signature = "[    { \22type\22 : \22f32\22 , \22dims\22 : [1 , 10 , 20 , 30] , \22name\22 : \22in_a\22 }\0A\0A]\00@[   { \22type\22 : \22f32\22 , \22dims\22 : [1 , 10 , 20 , 30] , \22name\22 : \22out_a\22 }\0A\0A]\00"} : () -> ()
+}

mlir-assigner/tests/Ops/NeedsCompilerWork/Hardmax/HardMaxSimple.mlir

@@ -0,0 +1,50 @@
+module attributes {llvm.data_layout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128", llvm.target_triple = "x86_64-pc-linux-gnu", "onnx-mlir.symbol-postfix" = "hardmaxsimple.mlir"} {
+  func.func @main_graph(%arg0: memref<1x10x20x30xf32>) -> memref<1x10x20x30xf32> attributes {input_names = ["in_a"], llvm.emit_c_interface, output_names = ["out_a"]} {
+    %cst = arith.constant 0.000000e+00 : f32
+    %cst_0 = arith.constant 1.000000e+00 : f32
+    %c0 = arith.constant 0 : index
+    %alloc = memref.alloc() {alignment = 16 : i64} : memref<1x10x20x30xf32>
+    %alloc_1 = memref.alloc() {alignment = 16 : i64} : memref<1x10x20x1xindex>
+    affine.for %arg1 = 0 to 1 {
+      affine.for %arg2 = 0 to 10 {
+        affine.for %arg3 = 0 to 20 {
+          affine.for %arg4 = 0 to 1 {
+            affine.store %c0, %alloc_1[%arg1, %arg2, %arg3, %arg4] : memref<1x10x20x1xindex>
+          }
+        }
+      }
+    }
+    affine.for %arg1 = 0 to 1 {
+      affine.for %arg2 = 0 to 10 {
+        affine.for %arg3 = 0 to 20 {
+          affine.for %arg4 = 0 to 30 {
+            %0 = affine.load %alloc_1[%arg1, %arg2, %arg3, %c0] : memref<1x10x20x1xindex>
+            %1 = memref.load %arg0[%arg1, %arg2, %arg3, %0] : memref<1x10x20x30xf32>
+            %2 = affine.load %arg0[%arg1, %arg2, %arg3, %arg4] : memref<1x10x20x30xf32>
+            %3 = arith.cmpf ogt, %2, %1 : f32
+            scf.if %3 {
+              affine.store %arg4, %alloc_1[%arg1, %arg2, %arg3, %c0] : memref<1x10x20x1xindex>
+            }
+          }
+        }
+      }
+    }
+    affine.for %arg1 = 0 to 1 {
+      affine.for %arg2 = 0 to 10 {
+        affine.for %arg3 = 0 to 20 {
+          affine.for %arg4 = 0 to 30 {
+            %0 = affine.load %alloc_1[%arg1, %arg2, %arg3, %c0] : memref<1x10x20x1xindex>
+            %1 = arith.cmpi eq, %0, %arg4 : index
+            scf.if %1 {
+              affine.store %cst_0, %alloc[%arg1, %arg2, %arg3, %arg4] : memref<1x10x20x30xf32>
+            } else {
+              affine.store %cst, %alloc[%arg1, %arg2, %arg3, %arg4] : memref<1x10x20x30xf32>
+            }
+          }
+        }
+      }
+    }
+    return %alloc : memref<1x10x20x30xf32>
+  }
+  "krnl.entry_point"() {func = @main_graph, numInputs = 1 : i32, numOutputs = 1 : i32, signature = "[    { \22type\22 : \22f32\22 , \22dims\22 : [1 , 10 , 20 , 30] , \22name\22 : \22in_a\22 }\0A\0A]\00@[   { \22type\22 : \22f32\22 , \22dims\22 : [1 , 10 , 20 , 30] , \22name\22 : \22out_a\22 }\0A\0A]\00"} : () -> ()
+}

mlir-assigner/tests/Ops/NeedsCompilerWork/Hardmax/HardMaxSimple.onnx

@@ -0,0 +1,17 @@
+	:h
+
+in_aout_a"HardmaxHardMaxSimpleZ
+in_a
+

+

+
+
+
+b
+out_a
+

+

+
+
+
+B