MLIR: add SinOp Lowering to __nv_sinf and __nv_sin

Added lowering rule from `SinOp` to `__nv_sinf` and `__nv_sin`

Reviewed By: ftynse

Differential Revision: https://reviews.llvm.org/D90147

diff --git a/mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp b/mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp
index 0dd1f98..61f97d4 100644 (file)
--- a/mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp
+++ b/mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp
@@ -138,7 +138,7 @@ struct LowerGpuOpsToNVVMOpsPass
     target.addIllegalDialect<gpu::GPUDialect>();
     target.addIllegalOp<LLVM::CosOp, LLVM::ExpOp, LLVM::FAbsOp, LLVM::FCeilOp,
                         LLVM::FFloorOp, LLVM::LogOp, LLVM::Log10Op,
-                        LLVM::Log2Op>();
+                        LLVM::Log2Op, LLVM::SinOp>();
     target.addIllegalOp<FuncOp>();
     target.addLegalDialect<NVVM::NVVMDialect>();
     // TODO: Remove once we support replacing non-root ops.
@@ -183,6 +183,8 @@ void mlir::populateGpuToNVVMConversionPatterns(
                                                  "__nv_log10");
   patterns.insert<OpToFuncCallLowering<Log2Op>>(converter, "__nv_log2f",
                                                 "__nv_log2");
+  patterns.insert<OpToFuncCallLowering<SinOp>>(converter, "__nv_sinf",
+                                               "__nv_sin");
   patterns.insert<OpToFuncCallLowering<TanhOp>>(converter, "__nv_tanhf",
                                                 "__nv_tanh");
 }

diff --git a/mlir/test/Conversion/GPUToNVVM/gpu-to-nvvm.mlir b/mlir/test/Conversion/GPUToNVVM/gpu-to-nvvm.mlir
index 6b071a0..8e4e451 100644 (file)
--- a/mlir/test/Conversion/GPUToNVVM/gpu-to-nvvm.mlir
+++ b/mlir/test/Conversion/GPUToNVVM/gpu-to-nvvm.mlir
@@ -262,6 +262,21 @@ gpu.module @test_module {
 // -----
 
 gpu.module @test_module {
+  // CHECK: llvm.func @__nv_sinf(!llvm.float) -> !llvm.float
+  // CHECK: llvm.func @__nv_sin(!llvm.double) -> !llvm.double
+  // CHECK-LABEL: func @gpu_sin
+  func @gpu_sin(%arg_f32 : f32, %arg_f64 : f64) -> (f32, f64) {
+    %result32 = std.sin %arg_f32 : f32
+    // CHECK: llvm.call @__nv_sinf(%{{.*}}) : (!llvm.float) -> !llvm.float
+    %result64 = std.sin %arg_f64 : f64
+    // CHECK: llvm.call @__nv_sin(%{{.*}}) : (!llvm.double) -> !llvm.double
+    std.return %result32, %result64 : f32, f64
+  }
+}
+
+// -----
+
+gpu.module @test_module {
   // CHECK: llvm.func @__nv_tanhf(!llvm.float) -> !llvm.float
   // CHECK: llvm.func @__nv_tanh(!llvm.double) -> !llvm.double
   // CHECK-LABEL: func @gpu_tanh