This is an archive of the discontinued LLVM Phabricator instance.

Differential D95130

Add log1p lowering from standard to NVVM intrinsics
ClosedPublic

Authored by frgossen on Jan 21 2021, 4:56 AM.

Details

Reviewers
pifon2a
herhut
Commits
rG294e2544c992: Add log1p lowering from standard to NVVM intrinsics

Diff Detail

Event Timeline

frgossen created this revision.Jan 21 2021, 4:56 AM
Herald added subscribers: teijeong, rdzhabarov, tatianashp and 16 others. · View Herald TranscriptJan 21 2021, 4:56 AM
frgossen requested review of this revision.Jan 21 2021, 4:56 AM
Herald added a project: Restricted Project. · View Herald TranscriptJan 21 2021, 4:56 AM
Herald added subscribers: stephenneuendorffer, nicolasvasilache. · View Herald Transcript
herhut accepted this revision.Jan 21 2021, 4:59 AM
This revision is now accepted and ready to land.Jan 21 2021, 4:59 AM
This revision was landed with ongoing or failed builds.Jan 21 2021, 5:01 AM
Closed by commit rG294e2544c992: Add log1p lowering from standard to NVVM intrinsics (authored by frgossen). · Explain Why
This revision was automatically updated to reflect the committed changes.
frgossen added a commit: rG294e2544c992: Add log1p lowering from standard to NVVM intrinsics.
Harbormaster completed remote builds in B86070: Diff 318161.Jan 21 2021, 5:47 AM

Revision Contents

PathSize
mlir/
lib/
Conversion/
GPUToNVVM/
2 lines
test/
Conversion/
GPUToNVVM/
15 lines

Diff 318162

mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp

Show First 20 Lines • Show All 181 Lines • ▼ Show 20 Linesvoid mlir::populateGpuToNVVMConversionPatterns(
patterns.insert<OpToFuncCallLowering<CosOp>>(converter, "__nv_cosf", patterns.insert<OpToFuncCallLowering<CosOp>>(converter, "__nv_cosf",
"__nv_cos"); "__nv_cos");
patterns.insert<OpToFuncCallLowering<ExpOp>>(converter, "__nv_expf", patterns.insert<OpToFuncCallLowering<ExpOp>>(converter, "__nv_expf",
"__nv_exp"); "__nv_exp");
patterns.insert<OpToFuncCallLowering<FloorFOp>>(converter, "__nv_floorf", patterns.insert<OpToFuncCallLowering<FloorFOp>>(converter, "__nv_floorf",
"__nv_floor"); "__nv_floor");
patterns.insert<OpToFuncCallLowering<LogOp>>(converter, "__nv_logf", patterns.insert<OpToFuncCallLowering<LogOp>>(converter, "__nv_logf",
"__nv_log"); "__nv_log");
patterns.insert<OpToFuncCallLowering<Log1pOp>>(converter, "__nv_log1pf",
"__nv_log1p");
patterns.insert<OpToFuncCallLowering<Log10Op>>(converter, "__nv_log10f", patterns.insert<OpToFuncCallLowering<Log10Op>>(converter, "__nv_log10f",
"__nv_log10"); "__nv_log10");
patterns.insert<OpToFuncCallLowering<Log2Op>>(converter, "__nv_log2f", patterns.insert<OpToFuncCallLowering<Log2Op>>(converter, "__nv_log2f",
"__nv_log2"); "__nv_log2");
patterns.insert<OpToFuncCallLowering<PowFOp>>(converter, "__nv_powf", patterns.insert<OpToFuncCallLowering<PowFOp>>(converter, "__nv_powf",
"__nv_pow"); "__nv_pow");
patterns.insert<OpToFuncCallLowering<RsqrtOp>>(converter, "__nv_rsqrtf", patterns.insert<OpToFuncCallLowering<RsqrtOp>>(converter, "__nv_rsqrtf",
"__nv_rsqrt"); "__nv_rsqrt");
Show All 12 Lines

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

Show First 20 Lines • Show All 241 Lines • ▼ Show 20 Lines func @gpu_log10(%arg_f32 : f32, %arg_f64 : f64) -> (f32, f64) {
// CHECK: llvm.call @__nv_log10(%{{.*}}) : (f64) -> f64 // CHECK: llvm.call @__nv_log10(%{{.*}}) : (f64) -> f64
std.return %result32, %result64 : f32, f64 std.return %result32, %result64 : f32, f64
} }
} }
// ----- // -----
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_log2f(f32) -> f32
// CHECK: llvm.func @__nv_log2(f64) -> f64 // CHECK: llvm.func @__nv_log2(f64) -> f64
// CHECK-LABEL: func @gpu_log2 // CHECK-LABEL: func @gpu_log2
func @gpu_log2(%arg_f32 : f32, %arg_f64 : f64) -> (f32, f64) { func @gpu_log2(%arg_f32 : f32, %arg_f64 : f64) -> (f32, f64) {
%result32 = std.log2 %arg_f32 : f32 %result32 = std.log2 %arg_f32 : f32
// CHECK: llvm.call @__nv_log2f(%{{.*}}) : (f32) -> f32 // CHECK: llvm.call @__nv_log2f(%{{.*}}) : (f32) -> f32
%result64 = std.log2 %arg_f64 : f64 %result64 = std.log2 %arg_f64 : f64
// CHECK: llvm.call @__nv_log2(%{{.*}}) : (f64) -> f64 // CHECK: llvm.call @__nv_log2(%{{.*}}) : (f64) -> f64
▲ Show 20 Lines • Show All 153 Lines • Show Last 20 Lines