[llvm-branch-commits] [llvm] [mlir] [mlir][Parser] Add `nan` and `inf` keywords (PR #116176)

via llvm-branch-commits llvm-branch-commits at lists.llvm.org
Wed Nov 13 23:51:34 PST 2024


llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-llvm-support

@llvm/pr-subscribers-mlir

Author: Matthias Springer (matthias-springer)

<details>
<summary>Changes</summary>

Add two new keywords for parsing `nan` / `inf` floating-point literals. This is more convenient that writing the integer hexadecimal bit patterns by hand (which differ depending on the floating-point type).

Note: The printer still prints `nan` / `inf` literals as integer hexadecimals. That's because there can be multiple `nan` / `inf` bit patterns. When parsing a `nan` / `inf` keyword, the exact bit pattern is unspecified: we use whatever `APFloat::getInf`/`APFloat::getNaN` returns.

Depends on #<!-- -->116172.


---
Full diff: https://github.com/llvm/llvm-project/pull/116176.diff


8 Files Affected:

- (modified) llvm/include/llvm/ADT/APFloat.h (+2) 
- (modified) llvm/lib/Support/APFloat.cpp (+9) 
- (modified) mlir/lib/AsmParser/AttributeParser.cpp (+21-9) 
- (modified) mlir/lib/AsmParser/Parser.cpp (+27-2) 
- (modified) mlir/lib/AsmParser/TokenKinds.def (+2) 
- (modified) mlir/test/Dialect/Arith/canonicalize.mlir (+5-5) 
- (modified) mlir/test/IR/attribute.mlir (+54) 
- (modified) mlir/test/mlir-cpu-runner/math-polynomial-approx.mlir (+18-18) 


``````````diff
diff --git a/llvm/include/llvm/ADT/APFloat.h b/llvm/include/llvm/ADT/APFloat.h
index 97547fb577e0ec..40ad7ba92552ed 100644
--- a/llvm/include/llvm/ADT/APFloat.h
+++ b/llvm/include/llvm/ADT/APFloat.h
@@ -311,6 +311,8 @@ struct APFloatBase {
   static unsigned int semanticsIntSizeInBits(const fltSemantics&, bool);
   static bool semanticsHasZero(const fltSemantics &);
   static bool semanticsHasSignedRepr(const fltSemantics &);
+  static bool semanticsHasInf(const fltSemantics &);
+  static bool semanticsHasNan(const fltSemantics &);
 
   // Returns true if any number described by \p Src can be precisely represented
   // by a normal (not subnormal) value in \p Dst.
diff --git a/llvm/lib/Support/APFloat.cpp b/llvm/lib/Support/APFloat.cpp
index c566d489d11b03..8b9d9af2ca65b3 100644
--- a/llvm/lib/Support/APFloat.cpp
+++ b/llvm/lib/Support/APFloat.cpp
@@ -375,6 +375,15 @@ bool APFloatBase::semanticsHasSignedRepr(const fltSemantics &semantics) {
   return semantics.hasSignedRepr;
 }
 
+bool APFloatBase::semanticsHasInf(const fltSemantics &semantics) {
+  return semantics.nonFiniteBehavior != fltNonfiniteBehavior::NanOnly
+      && semantics.nonFiniteBehavior != fltNonfiniteBehavior::FiniteOnly;
+}
+
+bool APFloatBase::semanticsHasNan(const fltSemantics &semantics) {
+  return semantics.nonFiniteBehavior != fltNonfiniteBehavior::FiniteOnly;
+}
+
 bool APFloatBase::isRepresentableAsNormalIn(const fltSemantics &Src,
                                             const fltSemantics &Dst) {
   // Exponent range must be larger.
diff --git a/mlir/lib/AsmParser/AttributeParser.cpp b/mlir/lib/AsmParser/AttributeParser.cpp
index 9ebada076cd042..68da950f09e568 100644
--- a/mlir/lib/AsmParser/AttributeParser.cpp
+++ b/mlir/lib/AsmParser/AttributeParser.cpp
@@ -21,8 +21,10 @@
 #include "mlir/IR/DialectImplementation.h"
 #include "mlir/IR/DialectResourceBlobManager.h"
 #include "mlir/IR/IntegerSet.h"
+#include "llvm/ADT/APFloat.h"
 #include "llvm/ADT/StringExtras.h"
 #include "llvm/Support/Endian.h"
+#include <cmath>
 #include <optional>
 
 using namespace mlir;
@@ -121,6 +123,8 @@ Attribute Parser::parseAttribute(Type type) {
 
   // Parse floating point and integer attributes.
   case Token::floatliteral:
+  case Token::kw_inf:
+  case Token::kw_nan:
     return parseFloatAttr(type, /*isNegative=*/false);
   case Token::integer:
     return parseDecOrHexAttr(type, /*isNegative=*/false);
@@ -128,7 +132,8 @@ Attribute Parser::parseAttribute(Type type) {
     consumeToken(Token::minus);
     if (getToken().is(Token::integer))
       return parseDecOrHexAttr(type, /*isNegative=*/true);
-    if (getToken().is(Token::floatliteral))
+    if (getToken().is(Token::floatliteral) || getToken().is(Token::kw_inf) ||
+        getToken().is(Token::kw_nan))
       return parseFloatAttr(type, /*isNegative=*/true);
 
     return (emitWrongTokenError(
@@ -342,10 +347,8 @@ ParseResult Parser::parseAttributeDict(NamedAttrList &attributes) {
 
 /// Parse a float attribute.
 Attribute Parser::parseFloatAttr(Type type, bool isNegative) {
-  auto val = getToken().getFloatingPointValue();
-  if (!val)
-    return (emitError("floating point value too large for attribute"), nullptr);
-  consumeToken(Token::floatliteral);
+  const Token tok = getToken();
+  consumeToken();
   if (!type) {
     // Default to F64 when no type is specified.
     if (!consumeIf(Token::colon))
@@ -353,10 +356,16 @@ Attribute Parser::parseFloatAttr(Type type, bool isNegative) {
     else if (!(type = parseType()))
       return nullptr;
   }
-  if (!isa<FloatType>(type))
+  auto floatType = dyn_cast<FloatType>(type);
+  if (!floatType)
     return (emitError("floating point value not valid for specified type"),
             nullptr);
-  return FloatAttr::get(type, isNegative ? -*val : *val);
+  auto emitErrorAtTok = [&]() { return emitError(tok.getLoc()); };
+  FailureOr<APFloat> result = parseFloatFromLiteral(
+      emitErrorAtTok, tok, isNegative, floatType.getFloatSemantics());
+  if (failed(result))
+    return Attribute();
+  return FloatAttr::get(floatType, *result);
 }
 
 /// Construct an APint from a parsed value, a known attribute type and
@@ -623,7 +632,7 @@ TensorLiteralParser::getIntAttrElements(SMLoc loc, Type eltTy,
     }
 
     // Check to see if floating point values were parsed.
-    if (token.is(Token::floatliteral)) {
+    if (token.isAny(Token::floatliteral, Token::kw_inf, Token::kw_nan)) {
       return p.emitError(tokenLoc)
              << "expected integer elements, but parsed floating-point";
     }
@@ -731,6 +740,8 @@ ParseResult TensorLiteralParser::parseElement() {
   // Parse a boolean element.
   case Token::kw_true:
   case Token::kw_false:
+  case Token::kw_inf:
+  case Token::kw_nan:
   case Token::floatliteral:
   case Token::integer:
     storage.emplace_back(/*isNegative=*/false, p.getToken());
@@ -740,7 +751,8 @@ ParseResult TensorLiteralParser::parseElement() {
   // Parse a signed integer or a negative floating-point element.
   case Token::minus:
     p.consumeToken(Token::minus);
-    if (!p.getToken().isAny(Token::floatliteral, Token::integer))
+    if (!p.getToken().isAny(Token::floatliteral, Token::kw_inf, Token::kw_nan,
+                            Token::integer))
       return p.emitError("expected integer or floating point literal");
     storage.emplace_back(/*isNegative=*/true, p.getToken());
     p.consumeToken();
diff --git a/mlir/lib/AsmParser/Parser.cpp b/mlir/lib/AsmParser/Parser.cpp
index 15f3dd7a66c358..6ff43b1749b64e 100644
--- a/mlir/lib/AsmParser/Parser.cpp
+++ b/mlir/lib/AsmParser/Parser.cpp
@@ -103,11 +103,36 @@ FailureOr<APFloat>
 detail::parseFloatFromLiteral(function_ref<InFlightDiagnostic()> emitError,
                               const Token &tok, bool isNegative,
                               const llvm::fltSemantics &semantics) {
+  // Check for inf keyword.
+  if (tok.is(Token::kw_inf)) {
+    if (!APFloat::semanticsHasInf(semantics)) {
+      emitError() << "floating point type does not support infinity";
+      return failure();
+    }
+    return APFloat::getInf(semantics, isNegative);
+  }
+
+  // Check for NaN keyword.
+  if (tok.is(Token::kw_nan)) {
+    if (!APFloat::semanticsHasNan(semantics)) {
+      emitError() << "floating point type does not support NaN";
+      return failure();
+    }
+    return APFloat::getNaN(semantics, isNegative);
+  }
+
   // Check for a floating point value.
   if (tok.is(Token::floatliteral)) {
     auto val = tok.getFloatingPointValue();
-    if (!val)
-      return emitError() << "floating point value too large";
+    if (!val) {
+      emitError() << "floating point value too large";
+      return failure();
+    }
+    if (std::fpclassify(*val) == FP_ZERO &&
+        !APFloat::semanticsHasZero(semantics)) {
+      emitError() << "floating point type does not support zero";
+      return failure();
+    }
 
     APFloat result(isNegative ? -*val : *val);
     bool unused;
diff --git a/mlir/lib/AsmParser/TokenKinds.def b/mlir/lib/AsmParser/TokenKinds.def
index 49da8c3dea5fa5..9208c8adddcfce 100644
--- a/mlir/lib/AsmParser/TokenKinds.def
+++ b/mlir/lib/AsmParser/TokenKinds.def
@@ -111,11 +111,13 @@ TOK_KEYWORD(floordiv)
 TOK_KEYWORD(for)
 TOK_KEYWORD(func)
 TOK_KEYWORD(index)
+TOK_KEYWORD(inf)
 TOK_KEYWORD(loc)
 TOK_KEYWORD(max)
 TOK_KEYWORD(memref)
 TOK_KEYWORD(min)
 TOK_KEYWORD(mod)
+TOK_KEYWORD(nan)
 TOK_KEYWORD(none)
 TOK_KEYWORD(offset)
 TOK_KEYWORD(size)
diff --git a/mlir/test/Dialect/Arith/canonicalize.mlir b/mlir/test/Dialect/Arith/canonicalize.mlir
index a386a178b78995..c86b2b5f63f016 100644
--- a/mlir/test/Dialect/Arith/canonicalize.mlir
+++ b/mlir/test/Dialect/Arith/canonicalize.mlir
@@ -1880,7 +1880,7 @@ func.func @test_minimumf(%arg0 : f32) -> (f32, f32, f32) {
   // CHECK-NEXT:  %[[X:.+]] = arith.minimumf %arg0, %[[C0]]
   // CHECK-NEXT:  return %[[X]], %arg0, %arg0
   %c0 = arith.constant 0.0 : f32
-  %inf = arith.constant 0x7F800000 : f32
+  %inf = arith.constant inf : f32
   %0 = arith.minimumf %c0, %arg0 : f32
   %1 = arith.minimumf %arg0, %arg0 : f32
   %2 = arith.minimumf %inf, %arg0 : f32
@@ -1895,7 +1895,7 @@ func.func @test_maximumf(%arg0 : f32) -> (f32, f32, f32) {
   // CHECK-NEXT:  %[[X:.+]] = arith.maximumf %arg0, %[[C0]]
   // CHECK-NEXT:   return %[[X]], %arg0, %arg0
   %c0 = arith.constant 0.0 : f32
-  %-inf = arith.constant 0xFF800000 : f32
+  %-inf = arith.constant -inf : f32
   %0 = arith.maximumf %c0, %arg0 : f32
   %1 = arith.maximumf %arg0, %arg0 : f32
   %2 = arith.maximumf %-inf, %arg0 : f32
@@ -1910,7 +1910,7 @@ func.func @test_minnumf(%arg0 : f32) -> (f32, f32, f32) {
   // CHECK-NEXT:  %[[X:.+]] = arith.minnumf %arg0, %[[C0]]
   // CHECK-NEXT:  return %[[X]], %arg0, %arg0
   %c0 = arith.constant 0.0 : f32
-  %inf = arith.constant 0x7F800000 : f32
+  %inf = arith.constant inf : f32
   %0 = arith.minnumf %c0, %arg0 : f32
   %1 = arith.minnumf %arg0, %arg0 : f32
   %2 = arith.minnumf %inf, %arg0 : f32
@@ -1925,7 +1925,7 @@ func.func @test_maxnumf(%arg0 : f32) -> (f32, f32, f32) {
   // CHECK-NEXT:  %[[X:.+]] = arith.maxnumf %arg0, %[[C0]]
   // CHECK-NEXT:   return %[[X]], %arg0, %arg0
   %c0 = arith.constant 0.0 : f32
-  %-inf = arith.constant 0xFF800000 : f32
+  %-inf = arith.constant -inf : f32
   %0 = arith.maxnumf %c0, %arg0 : f32
   %1 = arith.maxnumf %arg0, %arg0 : f32
   %2 = arith.maxnumf %-inf, %arg0 : f32
@@ -2024,7 +2024,7 @@ func.func @test_cmpf(%arg0 : f32) -> (i1, i1, i1, i1) {
 //   CHECK-DAG:   %[[T:.*]] = arith.constant true
 //   CHECK-DAG:   %[[F:.*]] = arith.constant false
 //       CHECK:   return %[[F]], %[[F]], %[[T]], %[[T]]
-  %nan = arith.constant 0x7fffffff : f32
+  %nan = arith.constant nan : f32
   %0 = arith.cmpf olt, %nan, %arg0 : f32
   %1 = arith.cmpf olt, %arg0, %nan : f32
   %2 = arith.cmpf ugt, %nan, %arg0 : f32
diff --git a/mlir/test/IR/attribute.mlir b/mlir/test/IR/attribute.mlir
index a62de3f5004d73..1fbb4986ab2ea7 100644
--- a/mlir/test/IR/attribute.mlir
+++ b/mlir/test/IR/attribute.mlir
@@ -108,9 +108,63 @@ func.func @float_attrs_pass() {
     // CHECK: float_attr = 2.000000e+00 : f128
     float_attr = 2. : f128
   } : () -> ()
+  "test.float_attrs"() {
+    // Note: nan/inf are printed in binary format because there may be multiple
+    // nan/inf representations.
+    // CHECK: float_attr = 0x7FC00000 : f32
+    float_attr = nan : f32
+  } : () -> ()
+  "test.float_attrs"() {
+    // CHECK: float_attr = 0x7C : f8E4M3
+    float_attr = nan : f8E4M3
+  } : () -> ()
+  "test.float_attrs"() {
+    // CHECK: float_attr = 0xFFC00000 : f32
+    float_attr = -nan : f32
+  } : () -> ()
+  "test.float_attrs"() {
+    // CHECK: float_attr = 0xFC : f8E4M3
+    float_attr = -nan : f8E4M3
+  } : () -> ()
+  "test.float_attrs"() {
+    // CHECK: float_attr = 0x7F800000 : f32
+    float_attr = inf : f32
+  } : () -> ()
+  "test.float_attrs"() {
+    // CHECK: float_attr = 0x78 : f8E4M3
+    float_attr = inf : f8E4M3
+  } : () -> ()
+  "test.float_attrs"() {
+    // CHECK: float_attr = 0xFF800000 : f32
+    float_attr = -inf : f32
+  } : () -> ()
+  "test.float_attrs"() {
+    // CHECK: float_attr = 0xF8 : f8E4M3
+    float_attr = -inf : f8E4M3
+  } : () -> ()
   return
 }
 
+// -----
+
+func.func @float_nan_unsupported() {
+  "test.float_attrs"() {
+    // expected-error @below{{floating point type does not support NaN}}
+    float_attr = nan : f4E2M1FN
+  } : () -> ()
+}
+
+// -----
+
+func.func @float_inf_unsupported() {
+  "test.float_attrs"() {
+    // expected-error @below{{floating point type does not support infinity}}
+    float_attr = inf : f4E2M1FN
+  } : () -> ()
+}
+
+// -----
+
 //===----------------------------------------------------------------------===//
 // Test integer attributes
 //===----------------------------------------------------------------------===//
diff --git a/mlir/test/mlir-cpu-runner/math-polynomial-approx.mlir b/mlir/test/mlir-cpu-runner/math-polynomial-approx.mlir
index b8861198d596b0..28b656b0da5f1a 100644
--- a/mlir/test/mlir-cpu-runner/math-polynomial-approx.mlir
+++ b/mlir/test/mlir-cpu-runner/math-polynomial-approx.mlir
@@ -41,7 +41,7 @@ func.func @tanh() {
   call @tanh_8xf32(%v2) : (vector<8xf32>) -> ()
 
   // CHECK: nan
-  %nan = arith.constant 0x7fc00000 : f32
+  %nan = arith.constant nan : f32
   call @tanh_f32(%nan) : (f32) -> ()
 
  return
@@ -87,15 +87,15 @@ func.func @log() {
   call @log_f32(%zero) : (f32) -> ()
 
   // CHECK: nan
-  %nan = arith.constant 0x7fc00000 : f32
+  %nan = arith.constant nan : f32
   call @log_f32(%nan) : (f32) -> ()
 
   // CHECK: inf
-  %inf = arith.constant 0x7f800000 : f32
+  %inf = arith.constant inf : f32
   call @log_f32(%inf) : (f32) -> ()
 
   // CHECK: -inf, nan, inf, 0.693147
-  %special_vec = arith.constant dense<[0.0, -1.0, 0x7f800000, 2.0]> : vector<4xf32>
+  %special_vec = arith.constant dense<[0.0, -1.0, inf, 2.0]> : vector<4xf32>
   call @log_4xf32(%special_vec) : (vector<4xf32>) -> ()
 
   return
@@ -141,11 +141,11 @@ func.func @log2() {
   call @log2_f32(%neg_one) : (f32) -> ()
 
   // CHECK: inf
-  %inf = arith.constant 0x7f800000 : f32
+  %inf = arith.constant inf : f32
   call @log2_f32(%inf) : (f32) -> ()
 
   // CHECK: -inf, nan, inf, 1.58496
-  %special_vec = arith.constant dense<[0.0, -1.0, 0x7f800000, 3.0]> : vector<4xf32>
+  %special_vec = arith.constant dense<[0.0, -1.0, inf, 3.0]> : vector<4xf32>
   call @log2_4xf32(%special_vec) : (vector<4xf32>) -> ()
 
   return
@@ -192,11 +192,11 @@ func.func @log1p() {
   call @log1p_f32(%neg_two) : (f32) -> ()
 
   // CHECK: inf
-  %inf = arith.constant 0x7f800000 : f32
+  %inf = arith.constant inf : f32
   call @log1p_f32(%inf) : (f32) -> ()
 
   // CHECK: -inf, nan, inf, 9.99995e-06
-  %special_vec = arith.constant dense<[-1.0, -1.1, 0x7f800000, 0.00001]> : vector<4xf32>
+  %special_vec = arith.constant dense<[-1.0, -1.1, inf, 0.00001]> : vector<4xf32>
   call @log1p_4xf32(%special_vec) : (vector<4xf32>) -> ()
 
   return
@@ -247,7 +247,7 @@ func.func @erf() {
   call @erf_f32(%val7) : (f32) -> ()
 
   // CHECK: -1
-  %negativeInf = arith.constant 0xff800000 : f32
+  %negativeInf = arith.constant -inf : f32
   call @erf_f32(%negativeInf) : (f32) -> ()
 
   // CHECK: -1, -1, -0.913759, -0.731446
@@ -263,11 +263,11 @@ func.func @erf() {
   call @erf_4xf32(%vecVals3) : (vector<4xf32>) -> ()
 
   // CHECK: 1
-  %inf = arith.constant 0x7f800000 : f32
+  %inf = arith.constant inf : f32
   call @erf_f32(%inf) : (f32) -> ()
 
   // CHECK: nan
-  %nan = arith.constant 0x7fc00000 : f32
+  %nan = arith.constant nan : f32
   call @erf_f32(%nan) : (f32) -> ()
 
   return
@@ -306,15 +306,15 @@ func.func @exp() {
   call @exp_4xf32(%special_vec) : (vector<4xf32>) -> ()
 
   // CHECK: inf
-  %inf = arith.constant 0x7f800000 : f32
+  %inf = arith.constant inf : f32
   call @exp_f32(%inf) : (f32) -> ()
 
   // CHECK: 0
-  %negative_inf = arith.constant 0xff800000 : f32
+  %negative_inf = arith.constant -inf : f32
   call @exp_f32(%negative_inf) : (f32) -> ()
 
   // CHECK: nan
-  %nan = arith.constant 0x7fc00000 : f32
+  %nan = arith.constant nan : f32
   call @exp_f32(%nan) : (f32) -> ()
 
   return
@@ -358,19 +358,19 @@ func.func @expm1() {
   call @expm1_8xf32(%v2) : (vector<8xf32>) -> ()
 
   // CHECK: -1
-  %neg_inf = arith.constant 0xff800000 : f32
+  %neg_inf = arith.constant -inf : f32
   call @expm1_f32(%neg_inf) : (f32) -> ()
 
   // CHECK: inf
-  %inf = arith.constant 0x7f800000 : f32
+  %inf = arith.constant inf : f32
   call @expm1_f32(%inf) : (f32) -> ()
 
   // CHECK: -1, inf, 1e-10
-  %special_vec = arith.constant dense<[0xff800000, 0x7f800000, 1.0e-10]> : vector<3xf32>
+  %special_vec = arith.constant dense<[-inf, inf, 1.0e-10]> : vector<3xf32>
   call @expm1_3xf32(%special_vec) : (vector<3xf32>) -> ()
 
   // CHECK: nan
-  %nan = arith.constant 0x7fc00000 : f32
+  %nan = arith.constant nan : f32
   call @expm1_f32(%nan) : (f32) -> ()
 
   return

``````````

</details>


https://github.com/llvm/llvm-project/pull/116176


More information about the llvm-branch-commits mailing list