Commit 294e2544 authored by Frederik Gossen's avatar Frederik Gossen
Browse files

Add log1p lowering from standard to NVVM intrinsics 

Differential Revision: https://reviews.llvm.org/D95130
parent 0ca81b90

mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp

+2 −0
Original line number Diff line number Diff line
@@ -187,6 +187,8 @@ void mlir::populateGpuToNVVMConversionPatterns( 
                                                  "__nv_floor");

  patterns.insert<OpToFuncCallLowering<LogOp>>(converter, "__nv_logf",

                                               "__nv_log");

  patterns.insert<OpToFuncCallLowering<Log1pOp>>(converter, "__nv_log1pf",

                                                 "__nv_log1p");

  patterns.insert<OpToFuncCallLowering<Log10Op>>(converter, "__nv_log10f",

                                                 "__nv_log10");

  patterns.insert<OpToFuncCallLowering<Log2Op>>(converter, "__nv_log2f",

mlir/test/Conversion/GPUToNVVM/gpu-to-nvvm.mlir

+15 −0
Original line number Diff line number Diff line
@@ -246,6 +246,21 @@ gpu.module @test_module { 


// -----



gpu.module @test_module {

  // CHECK: llvm.func @__nv_log1pf(f32) -> f32

  // CHECK: llvm.func @__nv_log1p(f64) -> f64

  // CHECK-LABEL: func @gpu_log1p

  func @gpu_log1p(%arg_f32 : f32, %arg_f64 : f64) -> (f32, f64) {

    %result32 = std.log1p %arg_f32 : f32

    // CHECK: llvm.call @__nv_log1pf(%{{.*}}) : (f32) -> f32

    %result64 = std.log1p %arg_f64 : f64

    // CHECK: llvm.call @__nv_log1p(%{{.*}}) : (f64) -> f64

    std.return %result32, %result64 : f32, f64

  }

}



// -----



gpu.module @test_module {

  // CHECK: llvm.func @__nv_log2f(f32) -> f32

  // CHECK: llvm.func @__nv_log2(f64) -> f64