[MLIR][Arith] FastMath extf conversion without NaN checks#180926
Merged
Conversation
This PR allows the expand op converter to consider the NoNaN fastmath attribute to disable the runtime checks for NaNs in E8M0 types. Default behaviour is still the same. The OCP document provides all-ones as NaN for E8M0, but for pre-MX I8 quantization, the checks for NaNs are prohibitively expensive if the hardware doesn't have native support for that type.
rengolin
requested review from
adam-smnk,
ftynse,
kuhar and
matthias-springer
Member
|
@llvm/pr-subscribers-mlir Author: Renato Golin (rengolin) ChangesThis PR allows the expand op converter to consider the NoNaN fastmath attribute to disable the runtime checks for NaNs in E8M0 types. Default behaviour is still the same. The OCP document provides all-ones as NaN for E8M0, but for pre-MX I8 quantization, the checks for NaNs are prohibitively expensive, especially if the hardware doesn't have native support for that type. Full diff: https://github.com/llvm/llvm-project/pull/180926.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Arith/Transforms/ExpandOps.cpp b/mlir/lib/Dialect/Arith/Transforms/ExpandOps.cpp
index c4e81e5dbed21..46f8c1037d47b 100644
--- a/mlir/lib/Dialect/Arith/Transforms/ExpandOps.cpp
+++ b/mlir/lib/Dialect/Arith/Transforms/ExpandOps.cpp
@@ -452,18 +452,25 @@ struct F8E8M0ExtFOpConverter : public OpRewritePattern<arith::ExtFOp> {
Type f32Ty = cloneToShapedType(operandTy, b.getF32Type());
Value bitcast = arith::BitcastOp::create(b, i8Ty, operand);
- // create constants for NaNs
- Value cF8NaN = createConst(op.getLoc(), i8Ty, 0xff, rewriter);
- Value cF32NaN = createConst(op.getLoc(), i32Ty, 0xffffffff, rewriter);
Value cF32MantissaWidth = createConst(op->getLoc(), i32Ty, 23, rewriter);
-
Value exti = arith::ExtUIOp::create(b, i32Ty, bitcast);
Value f32Bits = arith::ShLIOp::create(b, exti, cF32MantissaWidth);
- Value isNan =
- arith::CmpIOp::create(b, arith::CmpIPredicate::eq, bitcast, cF8NaN);
- // select for NaNs
- f32Bits = arith::SelectOp::create(b, isNan, cF32NaN, f32Bits);
+ // If FastMathFlag allows no NaN checks, skip it
+ auto fastMath = op.getFastmathAttr();
+ bool NoNaN = fastMath
+ ? (fastMath.getValue() & arith::FastMathFlags::nnan) ==
+ arith::FastMathFlags::nnan
+ : false;
+ if (!NoNaN) {
+ Value cF8NaN = createConst(op.getLoc(), i8Ty, 0xff, rewriter);
+ Value cF32NaN = createConst(op.getLoc(), i32Ty, 0xffffffff, rewriter);
+ Value isNan =
+ arith::CmpIOp::create(b, arith::CmpIPredicate::eq, bitcast, cF8NaN);
+ // select for NaNs
+ f32Bits = arith::SelectOp::create(b, isNan, cF32NaN, f32Bits);
+ }
+
Value result = arith::BitcastOp::create(b, f32Ty, f32Bits);
if (resultETy.getIntOrFloatBitWidth() < 32) {
result = arith::TruncFOp::create(b, resultTy, result, nullptr,
diff --git a/mlir/test/Dialect/Arith/expand-ops.mlir b/mlir/test/Dialect/Arith/expand-ops.mlir
index 61e22af31f030..75c4de2168761 100644
--- a/mlir/test/Dialect/Arith/expand-ops.mlir
+++ b/mlir/test/Dialect/Arith/expand-ops.mlir
@@ -383,11 +383,11 @@ func.func @extf_f8E8M0FNU_to_f32(%arg0 : f8E8M0FNU) -> f32 {
// CHECK-LABEL: @extf_f8E8M0FNU_to_f32
// CHECK: %[[BITCAST:.+]] = arith.bitcast %arg0 : f8E8M0FNU to i8
-// CHECK-DAG: %[[CF8NAN:.+]] = arith.constant -1 : i8
-// CHECK-DAG: %[[CF32NAN:.+]] = arith.constant -1 : i32
-// CHECK-DAG: %[[C23_i32:.+]] = arith.constant 23 : i32
+// CHECK: %[[C23_i32:.+]] = arith.constant 23 : i32
// CHECK: %[[EXTUI:.+]] = arith.extui %[[BITCAST]] : i8 to i32
// CHECK: %[[SHLI:.+]] = arith.shli %[[EXTUI]], %[[C23_i32]] : i32
+// CHECK-DAG: %[[CF8NAN:.+]] = arith.constant -1 : i8
+// CHECK-DAG: %[[CF32NAN:.+]] = arith.constant -1 : i32
// CHECK: %[[CMP_NAN:.+]] = arith.cmpi eq, %[[BITCAST]], %[[CF8NAN]] : i8
// CHECK: %[[SELECT_NAN:.+]] = arith.select %[[CMP_NAN]], %[[CF32NAN]], %[[SHLI]] : i32
// CHECK: %[[RESULT:.+]] = arith.bitcast %[[SELECT_NAN]] : i32 to f32
@@ -395,6 +395,21 @@ func.func @extf_f8E8M0FNU_to_f32(%arg0 : f8E8M0FNU) -> f32 {
// -----
+func.func @extf_f8E8M0FNU_to_f32_no_nan(%arg0 : f8E8M0FNU) -> f32 {
+ %0 = arith.extf %arg0 fastmath<nnan> : f8E8M0FNU to f32
+ return %0 : f32
+}
+
+// CHECK-LABEL: @extf_f8E8M0FNU_to_f32_no_nan
+// CHECK: %[[BITCAST:.+]] = arith.bitcast %arg0 : f8E8M0FNU to i8
+// CHECK: %[[C23_i32:.+]] = arith.constant 23 : i32
+// CHECK: %[[EXTUI:.+]] = arith.extui %[[BITCAST]] : i8 to i32
+// CHECK: %[[SHLI:.+]] = arith.shli %[[EXTUI]], %[[C23_i32]] : i32
+// CHECK: %[[RESULT:.+]] = arith.bitcast %[[SHLI]] : i32 to f32
+// CHECK: return %[[RESULT]]
+
+// -----
+
func.func @extf_f8E8M0FNU_to_f16(%arg0 : f8E8M0FNU) -> f16 {
%0 = arith.extf %arg0 : f8E8M0FNU to f16
return %0 : f16
@@ -402,11 +417,11 @@ func.func @extf_f8E8M0FNU_to_f16(%arg0 : f8E8M0FNU) -> f16 {
// CHECK-LABEL: @extf_f8E8M0FNU_to_f16
// CHECK: %[[BITCAST:.+]] = arith.bitcast %arg0 : f8E8M0FNU to i8
-// CHECK-DAG: %[[CF8NAN:.+]] = arith.constant -1 : i8
-// CHECK-DAG: %[[CF32NAN:.+]] = arith.constant -1 : i32
// CHECK-DAG: %[[C23_i32:.+]] = arith.constant 23 : i32
// CHECK: %[[EXTUI:.+]] = arith.extui %[[BITCAST]] : i8 to i32
// CHECK: %[[SHLI:.+]] = arith.shli %[[EXTUI]], %[[C23_i32]] : i32
+// CHECK-DAG: %[[CF8NAN:.+]] = arith.constant -1 : i8
+// CHECK-DAG: %[[CF32NAN:.+]] = arith.constant -1 : i32
// CHECK: %[[CMP_NAN:.+]] = arith.cmpi eq, %[[BITCAST]], %[[CF8NAN]] : i8
// CHECK: %[[SELECT_NAN:.+]] = arith.select %[[CMP_NAN]], %[[CF32NAN]], %[[SHLI]] : i32
// CHECK: %[[F32_RESULT:.+]] = arith.bitcast %[[SELECT_NAN]] : i32 to f32
|
Member
|
@llvm/pr-subscribers-mlir-arith Author: Renato Golin (rengolin) ChangesThis PR allows the expand op converter to consider the NoNaN fastmath attribute to disable the runtime checks for NaNs in E8M0 types. Default behaviour is still the same. The OCP document provides all-ones as NaN for E8M0, but for pre-MX I8 quantization, the checks for NaNs are prohibitively expensive, especially if the hardware doesn't have native support for that type. Full diff: https://github.com/llvm/llvm-project/pull/180926.diff 2 Files Affected:
diff --git a/mlir/lib/Dialect/Arith/Transforms/ExpandOps.cpp b/mlir/lib/Dialect/Arith/Transforms/ExpandOps.cpp
index c4e81e5dbed21..46f8c1037d47b 100644
--- a/mlir/lib/Dialect/Arith/Transforms/ExpandOps.cpp
+++ b/mlir/lib/Dialect/Arith/Transforms/ExpandOps.cpp
@@ -452,18 +452,25 @@ struct F8E8M0ExtFOpConverter : public OpRewritePattern<arith::ExtFOp> {
Type f32Ty = cloneToShapedType(operandTy, b.getF32Type());
Value bitcast = arith::BitcastOp::create(b, i8Ty, operand);
- // create constants for NaNs
- Value cF8NaN = createConst(op.getLoc(), i8Ty, 0xff, rewriter);
- Value cF32NaN = createConst(op.getLoc(), i32Ty, 0xffffffff, rewriter);
Value cF32MantissaWidth = createConst(op->getLoc(), i32Ty, 23, rewriter);
-
Value exti = arith::ExtUIOp::create(b, i32Ty, bitcast);
Value f32Bits = arith::ShLIOp::create(b, exti, cF32MantissaWidth);
- Value isNan =
- arith::CmpIOp::create(b, arith::CmpIPredicate::eq, bitcast, cF8NaN);
- // select for NaNs
- f32Bits = arith::SelectOp::create(b, isNan, cF32NaN, f32Bits);
+ // If FastMathFlag allows no NaN checks, skip it
+ auto fastMath = op.getFastmathAttr();
+ bool NoNaN = fastMath
+ ? (fastMath.getValue() & arith::FastMathFlags::nnan) ==
+ arith::FastMathFlags::nnan
+ : false;
+ if (!NoNaN) {
+ Value cF8NaN = createConst(op.getLoc(), i8Ty, 0xff, rewriter);
+ Value cF32NaN = createConst(op.getLoc(), i32Ty, 0xffffffff, rewriter);
+ Value isNan =
+ arith::CmpIOp::create(b, arith::CmpIPredicate::eq, bitcast, cF8NaN);
+ // select for NaNs
+ f32Bits = arith::SelectOp::create(b, isNan, cF32NaN, f32Bits);
+ }
+
Value result = arith::BitcastOp::create(b, f32Ty, f32Bits);
if (resultETy.getIntOrFloatBitWidth() < 32) {
result = arith::TruncFOp::create(b, resultTy, result, nullptr,
diff --git a/mlir/test/Dialect/Arith/expand-ops.mlir b/mlir/test/Dialect/Arith/expand-ops.mlir
index 61e22af31f030..75c4de2168761 100644
--- a/mlir/test/Dialect/Arith/expand-ops.mlir
+++ b/mlir/test/Dialect/Arith/expand-ops.mlir
@@ -383,11 +383,11 @@ func.func @extf_f8E8M0FNU_to_f32(%arg0 : f8E8M0FNU) -> f32 {
// CHECK-LABEL: @extf_f8E8M0FNU_to_f32
// CHECK: %[[BITCAST:.+]] = arith.bitcast %arg0 : f8E8M0FNU to i8
-// CHECK-DAG: %[[CF8NAN:.+]] = arith.constant -1 : i8
-// CHECK-DAG: %[[CF32NAN:.+]] = arith.constant -1 : i32
-// CHECK-DAG: %[[C23_i32:.+]] = arith.constant 23 : i32
+// CHECK: %[[C23_i32:.+]] = arith.constant 23 : i32
// CHECK: %[[EXTUI:.+]] = arith.extui %[[BITCAST]] : i8 to i32
// CHECK: %[[SHLI:.+]] = arith.shli %[[EXTUI]], %[[C23_i32]] : i32
+// CHECK-DAG: %[[CF8NAN:.+]] = arith.constant -1 : i8
+// CHECK-DAG: %[[CF32NAN:.+]] = arith.constant -1 : i32
// CHECK: %[[CMP_NAN:.+]] = arith.cmpi eq, %[[BITCAST]], %[[CF8NAN]] : i8
// CHECK: %[[SELECT_NAN:.+]] = arith.select %[[CMP_NAN]], %[[CF32NAN]], %[[SHLI]] : i32
// CHECK: %[[RESULT:.+]] = arith.bitcast %[[SELECT_NAN]] : i32 to f32
@@ -395,6 +395,21 @@ func.func @extf_f8E8M0FNU_to_f32(%arg0 : f8E8M0FNU) -> f32 {
// -----
+func.func @extf_f8E8M0FNU_to_f32_no_nan(%arg0 : f8E8M0FNU) -> f32 {
+ %0 = arith.extf %arg0 fastmath<nnan> : f8E8M0FNU to f32
+ return %0 : f32
+}
+
+// CHECK-LABEL: @extf_f8E8M0FNU_to_f32_no_nan
+// CHECK: %[[BITCAST:.+]] = arith.bitcast %arg0 : f8E8M0FNU to i8
+// CHECK: %[[C23_i32:.+]] = arith.constant 23 : i32
+// CHECK: %[[EXTUI:.+]] = arith.extui %[[BITCAST]] : i8 to i32
+// CHECK: %[[SHLI:.+]] = arith.shli %[[EXTUI]], %[[C23_i32]] : i32
+// CHECK: %[[RESULT:.+]] = arith.bitcast %[[SHLI]] : i32 to f32
+// CHECK: return %[[RESULT]]
+
+// -----
+
func.func @extf_f8E8M0FNU_to_f16(%arg0 : f8E8M0FNU) -> f16 {
%0 = arith.extf %arg0 : f8E8M0FNU to f16
return %0 : f16
@@ -402,11 +417,11 @@ func.func @extf_f8E8M0FNU_to_f16(%arg0 : f8E8M0FNU) -> f16 {
// CHECK-LABEL: @extf_f8E8M0FNU_to_f16
// CHECK: %[[BITCAST:.+]] = arith.bitcast %arg0 : f8E8M0FNU to i8
-// CHECK-DAG: %[[CF8NAN:.+]] = arith.constant -1 : i8
-// CHECK-DAG: %[[CF32NAN:.+]] = arith.constant -1 : i32
// CHECK-DAG: %[[C23_i32:.+]] = arith.constant 23 : i32
// CHECK: %[[EXTUI:.+]] = arith.extui %[[BITCAST]] : i8 to i32
// CHECK: %[[SHLI:.+]] = arith.shli %[[EXTUI]], %[[C23_i32]] : i32
+// CHECK-DAG: %[[CF8NAN:.+]] = arith.constant -1 : i8
+// CHECK-DAG: %[[CF32NAN:.+]] = arith.constant -1 : i32
// CHECK: %[[CMP_NAN:.+]] = arith.cmpi eq, %[[BITCAST]], %[[CF8NAN]] : i8
// CHECK: %[[SELECT_NAN:.+]] = arith.select %[[CMP_NAN]], %[[CF32NAN]], %[[SHLI]] : i32
// CHECK: %[[F32_RESULT:.+]] = arith.bitcast %[[SELECT_NAN]] : i32 to f32
|
matthias-springer
approved these changes
Feb 11, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR allows the expand op converter to consider the NoNaN fastmath attribute to disable the runtime checks for NaNs in E8M0 types. Default behaviour is still the same.
The OCP document provides all-ones as NaN for E8M0, but for pre-MX I8 quantization, the checks for NaNs are prohibitively expensive, especially if the hardware doesn't have native support for that type.