This is an archive of the discontinued LLVM Phabricator instance.

Differential D95129

Add log1p lowering from standard to ROCDL intrinsics
ClosedPublic

Authored by frgossen on Jan 21 2021, 4:55 AM.

Details

Reviewers
pifon2a
herhut
Commits
rG4ef38f9c1255: Add log1p lowering from standard to ROCDL intrinsics

Diff Detail

Event Timeline

frgossen created this revision.Jan 21 2021, 4:55 AM
Herald added subscribers: teijeong, rdzhabarov, tatianashp and 15 others. · View Herald TranscriptJan 21 2021, 4:55 AM
frgossen requested review of this revision.Jan 21 2021, 4:55 AM
Herald added a project: Restricted Project. · View Herald TranscriptJan 21 2021, 4:55 AM
Herald added subscribers: stephenneuendorffer, nicolasvasilache. · View Herald Transcript
herhut accepted this revision.Jan 21 2021, 4:58 AM

Thanks!

This revision is now accepted and ready to land.Jan 21 2021, 4:58 AM
This revision was landed with ongoing or failed builds.Jan 21 2021, 5:03 AM
Closed by commit rG4ef38f9c1255: Add log1p lowering from standard to ROCDL intrinsics (authored by frgossen). · Explain Why
This revision was automatically updated to reflect the committed changes.
frgossen added a commit: rG4ef38f9c1255: Add log1p lowering from standard to ROCDL intrinsics.
Harbormaster completed remote builds in B86069: Diff 318160.Jan 21 2021, 5:29 AM

Revision Contents

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

Diff 318164

mlir/lib/Conversion/GPUToROCDL/LowerGpuOpsToROCDLOps.cpp

Show First 20 Lines • Show All 115 Lines • ▼ Show 20 Linesvoid mlir::populateGpuToROCDLConversionPatterns(
patterns.insert<OpToFuncCallLowering<ExpOp>>(converter, "__ocml_exp_f32", patterns.insert<OpToFuncCallLowering<ExpOp>>(converter, "__ocml_exp_f32",
"__ocml_exp_f64"); "__ocml_exp_f64");
patterns.insert<OpToFuncCallLowering<FloorFOp>>(converter, "__ocml_floor_f32", patterns.insert<OpToFuncCallLowering<FloorFOp>>(converter, "__ocml_floor_f32",
"__ocml_floor_f64"); "__ocml_floor_f64");
patterns.insert<OpToFuncCallLowering<LogOp>>(converter, "__ocml_log_f32", patterns.insert<OpToFuncCallLowering<LogOp>>(converter, "__ocml_log_f32",
"__ocml_log_f64"); "__ocml_log_f64");
patterns.insert<OpToFuncCallLowering<Log10Op>>(converter, "__ocml_log10_f32", patterns.insert<OpToFuncCallLowering<Log10Op>>(converter, "__ocml_log10_f32",
"__ocml_log10_f64"); "__ocml_log10_f64");
patterns.insert<OpToFuncCallLowering<Log1pOp>>(converter, "__ocml_log1p_f32",
"__ocml_log1p_f64");
patterns.insert<OpToFuncCallLowering<Log2Op>>(converter, "__ocml_log2_f32", patterns.insert<OpToFuncCallLowering<Log2Op>>(converter, "__ocml_log2_f32",
"__ocml_log2_f64"); "__ocml_log2_f64");
patterns.insert<OpToFuncCallLowering<PowFOp>>(converter, "__ocml_pow_f32", patterns.insert<OpToFuncCallLowering<PowFOp>>(converter, "__ocml_pow_f32",
"__ocml_pow_f64"); "__ocml_pow_f64");
patterns.insert<OpToFuncCallLowering<RsqrtOp>>(converter, "__ocml_rsqrt_f32", patterns.insert<OpToFuncCallLowering<RsqrtOp>>(converter, "__ocml_rsqrt_f32",
"__ocml_rsqrt_f64"); "__ocml_rsqrt_f64");
patterns.insert<OpToFuncCallLowering<SinOp>>(converter, "__ocml_sin_f32", patterns.insert<OpToFuncCallLowering<SinOp>>(converter, "__ocml_sin_f32",
"__ocml_sin_f64"); "__ocml_sin_f64");
Show All 10 Lines

mlir/test/Conversion/GPUToROCDL/gpu-to-rocdl.mlir

Show First 20 Lines • Show All 193 Lines • ▼ Show 20 Lines func @gpu_log(%arg_f32 : f32, %arg_f64 : f64) -> (f32, f64) {
// CHECK: llvm.call @__ocml_log_f64(%{{.*}}) : (f64) -> f64 // CHECK: llvm.call @__ocml_log_f64(%{{.*}}) : (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 @__ocml_log1p_f32(f32) -> f32
// CHECK: llvm.func @__ocml_log1p_f64(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 @__ocml_log1p_f32(%{{.*}}) : (f32) -> f32
%result64 = std.log1p %arg_f64 : f64
// CHECK: llvm.call @__ocml_log1p_f64(%{{.*}}) : (f64) -> f64
std.return %result32, %result64 : f32, f64
}
}
// -----
gpu.module @test_module {
// CHECK: llvm.func @__ocml_log10_f32(f32) -> f32 // CHECK: llvm.func @__ocml_log10_f32(f32) -> f32
// CHECK: llvm.func @__ocml_log10_f64(f64) -> f64 // CHECK: llvm.func @__ocml_log10_f64(f64) -> f64
// CHECK-LABEL: func @gpu_log10 // CHECK-LABEL: func @gpu_log10
func @gpu_log10(%arg_f32 : f32, %arg_f64 : f64) -> (f32, f64) { func @gpu_log10(%arg_f32 : f32, %arg_f64 : f64) -> (f32, f64) {
%result32 = std.log10 %arg_f32 : f32 %result32 = std.log10 %arg_f32 : f32
// CHECK: llvm.call @__ocml_log10_f32(%{{.*}}) : (f32) -> f32 // CHECK: llvm.call @__ocml_log10_f32(%{{.*}}) : (f32) -> f32
%result64 = std.log10 %arg_f64 : f64 %result64 = std.log10 %arg_f64 : f64
// CHECK: llvm.call @__ocml_log10_f64(%{{.*}}) : (f64) -> f64 // CHECK: llvm.call @__ocml_log10_f64(%{{.*}}) : (f64) -> f64
▲ Show 20 Lines • Show All 118 Lines • Show Last 20 Lines