[llvm-branch-commits] [llvm] [BOLT] Fix counts aggregation in merge-fdata (PR #119652)
https://github.com/aaupov edited https://github.com/llvm/llvm-project/pull/119652 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Fix counts aggregation in merge-fdata (PR #119652)
https://github.com/aaupov updated
https://github.com/llvm/llvm-project/pull/119652
>From cc7cea0276fef36e896ecef149ef680b66bb9c1f Mon Sep 17 00:00:00 2001
From: Amir Ayupov
Date: Wed, 11 Dec 2024 19:11:07 -0800
Subject: [PATCH 1/2] =?UTF-8?q?[=F0=9D=98=80=F0=9D=97=BD=F0=9D=97=BF]=20in?=
=?UTF-8?q?itial=20version?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Created using spr 1.3.4
---
bolt/test/merge-fdata.test | 8 ++
bolt/tools/merge-fdata/merge-fdata.cpp | 39 ++
2 files changed, 36 insertions(+), 11 deletions(-)
create mode 100644 bolt/test/merge-fdata.test
diff --git a/bolt/test/merge-fdata.test b/bolt/test/merge-fdata.test
new file mode 100644
index 00..0ea0f2cb4398b6
--- /dev/null
+++ b/bolt/test/merge-fdata.test
@@ -0,0 +1,8 @@
+## Reproduces an issue where counts were not accumulated by merge-fdata
+
+# RUN: split-file %s %t
+# RUN: merge-fdata %t/fdata | FileCheck %s
+# CHECK: 1 main 10 1 main 1a 1 20
+#--- fdata
+1 main 10 1 main 1a 0 10
+1 main 10 1 main 1a 1 10
diff --git a/bolt/tools/merge-fdata/merge-fdata.cpp
b/bolt/tools/merge-fdata/merge-fdata.cpp
index 89ca46c1c0a8fa..a74e9099df9059 100644
--- a/bolt/tools/merge-fdata/merge-fdata.cpp
+++ b/bolt/tools/merge-fdata/merge-fdata.cpp
@@ -19,6 +19,7 @@
#include "llvm/Support/FileSystem.h"
#include "llvm/Support/ManagedStatic.h"
#include "llvm/Support/PrettyStackTrace.h"
+#include "llvm/Support/Regex.h"
#include "llvm/Support/Signals.h"
#include "llvm/Support/ThreadPool.h"
#include
@@ -266,7 +267,19 @@ void mergeLegacyProfiles(const
SmallVectorImpl &Filenames) {
errs() << "Using legacy profile format.\n";
std::optional BoltedCollection;
std::mutex BoltedCollectionMutex;
- typedef StringMap ProfileTy;
+ constexpr static const char *const FdataCountersPattern =
+ "(.*) ([0-9]+) ([0-9]+)";
+ Regex FdataRegex(FdataCountersPattern);
+ struct CounterTy {
+uint64_t Count;
+uint64_t MispredCount;
+CounterTy &operator+(const CounterTy &O) {
+ Count += O.Count;
+ MispredCount += O.MispredCount;
+ return *this;
+}
+ };
+ typedef StringMap ProfileTy;
auto ParseProfile = [&](const std::string &Filename, auto &Profiles) {
const llvm::thread::id tid = llvm::this_thread::get_id();
@@ -304,15 +317,19 @@ void mergeLegacyProfiles(const
SmallVectorImpl &Filenames) {
SmallVector Lines;
SplitString(Buf, Lines, "\n");
for (StringRef Line : Lines) {
- size_t Pos = Line.rfind(" ");
- if (Pos == StringRef::npos)
+ CounterTy CurrCount;
+ SmallVector Fields;
+ if (!FdataRegex.match(Line, &Fields))
report_error(Filename, "Malformed / corrupted profile");
- StringRef Signature = Line.substr(0, Pos);
- uint64_t Count;
- if (Line.substr(Pos + 1, Line.size() - Pos).getAsInteger(10, Count))
-report_error(Filename, "Malformed / corrupted profile counter");
- Count += Profile->lookup(Signature);
- Profile->insert_or_assign(Signature, Count);
+ StringRef Signature = Fields[1];
+ if (Fields[2].getAsInteger(10, CurrCount.MispredCount))
+report_error(Filename, "Malformed / corrupted execution count");
+ if (Fields[3].getAsInteger(10, CurrCount.Count))
+report_error(Filename, "Malformed / corrupted misprediction count");
+
+ CounterTy Counter = Profile->lookup(Signature);
+ Counter = Counter + CurrCount;
+ Profile->insert_or_assign(Signature, Counter);
}
};
@@ -330,14 +347,14 @@ void mergeLegacyProfiles(const
SmallVectorImpl &Filenames) {
ProfileTy MergedProfile;
for (const auto &[Thread, Profile] : ParsedProfiles)
for (const auto &[Key, Value] : Profile) {
- uint64_t Count = MergedProfile.lookup(Key) + Value;
+ auto Count = MergedProfile.lookup(Key) + Value;
MergedProfile.insert_or_assign(Key, Count);
}
if (BoltedCollection.value_or(false))
output() << "boltedcollection\n";
for (const auto &[Key, Value] : MergedProfile)
-output() << Key << " " << Value << "\n";
+output() << Key << " " << Value.MispredCount << " " << Value.Count << "\n";
errs() << "Profile from " << Filenames.size() << " files merged.\n";
}
>From 040deec056483f8268d134ce826a20f1e511dcd4 Mon Sep 17 00:00:00 2001
From: Amir Ayupov
Date: Wed, 11 Dec 2024 19:22:23 -0800
Subject: [PATCH 2/2] simplify
Created using spr 1.3.4
---
bolt/tools/merge-fdata/merge-fdata.cpp | 24
1 file changed, 12 insertions(+), 12 deletions(-)
diff --git a/bolt/tools/merge-fdata/merge-fdata.cpp
b/bolt/tools/merge-fdata/merge-fdata.cpp
index a74e9099df9059..83616e1de99254 100644
--- a/bolt/tools/merge-fdata/merge-fdata.cpp
+++ b/bolt/tools/merge-fdata/merge-fdata.cpp
@@ -271,13 +271,14 @@ void mergeLegacyProfiles(const
SmallVectorImpl &Filenames) {
"(.*) ([0-9]+) ([0-9]+)";
Regex FdataRegex(FdataCountersPattern);
struct CounterTy {
[llvm-branch-commits] [llvm] [BOLT] Fix counts aggregation in merge-fdata (PR #119652)
https://github.com/aaupov updated
https://github.com/llvm/llvm-project/pull/119652
>From cc7cea0276fef36e896ecef149ef680b66bb9c1f Mon Sep 17 00:00:00 2001
From: Amir Ayupov
Date: Wed, 11 Dec 2024 19:11:07 -0800
Subject: [PATCH 1/2] =?UTF-8?q?[=F0=9D=98=80=F0=9D=97=BD=F0=9D=97=BF]=20in?=
=?UTF-8?q?itial=20version?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Created using spr 1.3.4
---
bolt/test/merge-fdata.test | 8 ++
bolt/tools/merge-fdata/merge-fdata.cpp | 39 ++
2 files changed, 36 insertions(+), 11 deletions(-)
create mode 100644 bolt/test/merge-fdata.test
diff --git a/bolt/test/merge-fdata.test b/bolt/test/merge-fdata.test
new file mode 100644
index 00..0ea0f2cb4398b6
--- /dev/null
+++ b/bolt/test/merge-fdata.test
@@ -0,0 +1,8 @@
+## Reproduces an issue where counts were not accumulated by merge-fdata
+
+# RUN: split-file %s %t
+# RUN: merge-fdata %t/fdata | FileCheck %s
+# CHECK: 1 main 10 1 main 1a 1 20
+#--- fdata
+1 main 10 1 main 1a 0 10
+1 main 10 1 main 1a 1 10
diff --git a/bolt/tools/merge-fdata/merge-fdata.cpp
b/bolt/tools/merge-fdata/merge-fdata.cpp
index 89ca46c1c0a8fa..a74e9099df9059 100644
--- a/bolt/tools/merge-fdata/merge-fdata.cpp
+++ b/bolt/tools/merge-fdata/merge-fdata.cpp
@@ -19,6 +19,7 @@
#include "llvm/Support/FileSystem.h"
#include "llvm/Support/ManagedStatic.h"
#include "llvm/Support/PrettyStackTrace.h"
+#include "llvm/Support/Regex.h"
#include "llvm/Support/Signals.h"
#include "llvm/Support/ThreadPool.h"
#include
@@ -266,7 +267,19 @@ void mergeLegacyProfiles(const
SmallVectorImpl &Filenames) {
errs() << "Using legacy profile format.\n";
std::optional BoltedCollection;
std::mutex BoltedCollectionMutex;
- typedef StringMap ProfileTy;
+ constexpr static const char *const FdataCountersPattern =
+ "(.*) ([0-9]+) ([0-9]+)";
+ Regex FdataRegex(FdataCountersPattern);
+ struct CounterTy {
+uint64_t Count;
+uint64_t MispredCount;
+CounterTy &operator+(const CounterTy &O) {
+ Count += O.Count;
+ MispredCount += O.MispredCount;
+ return *this;
+}
+ };
+ typedef StringMap ProfileTy;
auto ParseProfile = [&](const std::string &Filename, auto &Profiles) {
const llvm::thread::id tid = llvm::this_thread::get_id();
@@ -304,15 +317,19 @@ void mergeLegacyProfiles(const
SmallVectorImpl &Filenames) {
SmallVector Lines;
SplitString(Buf, Lines, "\n");
for (StringRef Line : Lines) {
- size_t Pos = Line.rfind(" ");
- if (Pos == StringRef::npos)
+ CounterTy CurrCount;
+ SmallVector Fields;
+ if (!FdataRegex.match(Line, &Fields))
report_error(Filename, "Malformed / corrupted profile");
- StringRef Signature = Line.substr(0, Pos);
- uint64_t Count;
- if (Line.substr(Pos + 1, Line.size() - Pos).getAsInteger(10, Count))
-report_error(Filename, "Malformed / corrupted profile counter");
- Count += Profile->lookup(Signature);
- Profile->insert_or_assign(Signature, Count);
+ StringRef Signature = Fields[1];
+ if (Fields[2].getAsInteger(10, CurrCount.MispredCount))
+report_error(Filename, "Malformed / corrupted execution count");
+ if (Fields[3].getAsInteger(10, CurrCount.Count))
+report_error(Filename, "Malformed / corrupted misprediction count");
+
+ CounterTy Counter = Profile->lookup(Signature);
+ Counter = Counter + CurrCount;
+ Profile->insert_or_assign(Signature, Counter);
}
};
@@ -330,14 +347,14 @@ void mergeLegacyProfiles(const
SmallVectorImpl &Filenames) {
ProfileTy MergedProfile;
for (const auto &[Thread, Profile] : ParsedProfiles)
for (const auto &[Key, Value] : Profile) {
- uint64_t Count = MergedProfile.lookup(Key) + Value;
+ auto Count = MergedProfile.lookup(Key) + Value;
MergedProfile.insert_or_assign(Key, Count);
}
if (BoltedCollection.value_or(false))
output() << "boltedcollection\n";
for (const auto &[Key, Value] : MergedProfile)
-output() << Key << " " << Value << "\n";
+output() << Key << " " << Value.MispredCount << " " << Value.Count << "\n";
errs() << "Profile from " << Filenames.size() << " files merged.\n";
}
>From 040deec056483f8268d134ce826a20f1e511dcd4 Mon Sep 17 00:00:00 2001
From: Amir Ayupov
Date: Wed, 11 Dec 2024 19:22:23 -0800
Subject: [PATCH 2/2] simplify
Created using spr 1.3.4
---
bolt/tools/merge-fdata/merge-fdata.cpp | 24
1 file changed, 12 insertions(+), 12 deletions(-)
diff --git a/bolt/tools/merge-fdata/merge-fdata.cpp
b/bolt/tools/merge-fdata/merge-fdata.cpp
index a74e9099df9059..83616e1de99254 100644
--- a/bolt/tools/merge-fdata/merge-fdata.cpp
+++ b/bolt/tools/merge-fdata/merge-fdata.cpp
@@ -271,13 +271,14 @@ void mergeLegacyProfiles(const
SmallVectorImpl &Filenames) {
"(.*) ([0-9]+) ([0-9]+)";
Regex FdataRegex(FdataCountersPattern);
struct CounterTy {
[llvm-branch-commits] [mlir] [mlir][Transforms] Support 1:N mappings in `ConversionValueMapping` (PR #116524)
https://github.com/matthias-springer edited https://github.com/llvm/llvm-project/pull/116524 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [mlir] [mlir][Transforms] Support 1:N mappings in `ConversionValueMapping` (PR #116524)
https://github.com/matthias-springer edited https://github.com/llvm/llvm-project/pull/116524 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [mlir] [mlir][Vector] Clean up `populateVectorToLLVMConversionPatterns` (PR #119975)
https://github.com/matthias-springer created
https://github.com/llvm/llvm-project/pull/119975
Clean up `populateVectorToLLVMConversionPatterns` so that it populates only
conversion patterns. All rewrite patterns that do not lower to LLVM should be
populated into a separate greedy pattern rewrite.
The current combination of rewrite patterns and conversion patterns triggered
an edge case when merging the 1:1 and 1:N dialect conversions.
Depends on #119973.
>From e5926b63835eec731c513d91ce9c451c429ca572 Mon Sep 17 00:00:00 2001
From: Matthias Springer
Date: Sat, 14 Dec 2024 17:19:18 +0100
Subject: [PATCH] [mlir][Vector] Clean up
`populateVectorToLLVMConversionPatterns`
---
.../Vector/Transforms/LoweringPatterns.h | 4 +++
.../GPUCommon/GPUToLLVMConversion.cpp | 12 +
.../VectorToLLVM/ConvertVectorToLLVM.cpp | 27 ++-
.../VectorToLLVM/ConvertVectorToLLVMPass.cpp | 6 -
.../Conversion/GPUCommon/lower-vector.mlir| 4 +--
.../VectorToLLVM/vector-to-llvm.mlir | 5
6 files changed, 37 insertions(+), 21 deletions(-)
diff --git a/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
b/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
index 3d643c96b45008..c507b23c6d4de6 100644
--- a/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
+++ b/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
@@ -292,6 +292,10 @@ void
populateVectorBitCastLoweringPatterns(RewritePatternSet &patterns,
int64_t targetRank = 1,
PatternBenefit benefit = 1);
+/// Populates a pattern that rank-reduces n-D FMAs into (n-1)-D FMAs where
+/// n > 1.
+void populateVectorRankReducingFMAPattern(RewritePatternSet &patterns);
+
} // namespace vector
} // namespace mlir
#endif // MLIR_DIALECT_VECTOR_TRANSFORMS_LOWERINGPATTERNS_H
diff --git a/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
b/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
index 1497d662dcdbdd..2fe3b1302e5e5b 100644
--- a/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
+++ b/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
@@ -32,10 +32,12 @@
#include "mlir/Dialect/GPU/Transforms/Passes.h"
#include "mlir/Dialect/LLVMIR/LLVMDialect.h"
#include "mlir/Dialect/MemRef/IR/MemRef.h"
+#include "mlir/Dialect/Vector/Transforms/LoweringPatterns.h"
#include "mlir/IR/Attributes.h"
#include "mlir/IR/Builders.h"
#include "mlir/IR/BuiltinOps.h"
#include "mlir/IR/BuiltinTypes.h"
+#include "mlir/Transforms/GreedyPatternRewriteDriver.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/Support/Error.h"
@@ -522,6 +524,16 @@
DECLARE_CONVERT_OP_TO_GPU_RUNTIME_CALL_PATTERN(SetCsrPointersOp)
void GpuToLLVMConversionPass::runOnOperation() {
MLIRContext *context = &getContext();
+
+ // Perform progressive lowering of vector transfer operations.
+ {
+RewritePatternSet patterns(&getContext());
+// Vector transfer ops with rank > 1 should be lowered with VectorToSCF.
+vector::populateVectorTransferLoweringPatterns(patterns,
+ /*maxTransferRank=*/1);
+(void)applyPatternsAndFoldGreedily(getOperation(), std::move(patterns));
+ }
+
LowerToLLVMOptions options(context);
options.useBarePtrCallConv = hostBarePtrCallConv;
RewritePatternSet patterns(context);
diff --git a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
b/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
index a9a07c323c7358..577b74bb7e0c26 100644
--- a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
+++ b/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
@@ -1475,16 +1475,16 @@ class VectorTypeCastOpConversion
/// Conversion pattern for a `vector.create_mask` (1-D scalable vectors only).
/// Non-scalable versions of this operation are handled in Vector Transforms.
-class VectorCreateMaskOpRewritePattern
-: public OpRewritePattern {
+class VectorCreateMaskOpConversion
+: public OpConversionPattern {
public:
- explicit VectorCreateMaskOpRewritePattern(MLIRContext *context,
+ explicit VectorCreateMaskOpConversion(MLIRContext *context,
bool enableIndexOpt)
- : OpRewritePattern(context),
+ : OpConversionPattern(context),
force32BitVectorIndices(enableIndexOpt) {}
- LogicalResult matchAndRewrite(vector::CreateMaskOp op,
-PatternRewriter &rewriter) const override {
+ LogicalResult matchAndRewrite(vector::CreateMaskOp op, OpAdaptor adaptor,
+ConversionPatternRewriter &rewriter) const
override {
auto dstType = op.getType();
if (dstType.getRank() != 1 || !cast(dstType).isScalable())
return failure();
@@ -1495,7 +1495,7 @@ class VectorCreateMaskOpRewritePattern
loc, LLVM::getVectorType(idxType, dstType.getShape()[0],
[llvm-branch-commits] [mlir] [mlir][Vector] Clean up `populateVectorToLLVMConversionPatterns` (PR #119975)
llvmbot wrote:
@llvm/pr-subscribers-mlir
Author: Matthias Springer (matthias-springer)
Changes
Clean up `populateVectorToLLVMConversionPatterns` so that it populates only
conversion patterns. All rewrite patterns that do not lower to LLVM should be
populated into a separate greedy pattern rewrite.
The current combination of rewrite patterns and conversion patterns triggered
an edge case when merging the 1:1 and 1:N dialect conversions.
Depends on #119973.
---
Full diff: https://github.com/llvm/llvm-project/pull/119975.diff
6 Files Affected:
- (modified) mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
(+4)
- (modified) mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp (+12)
- (modified) mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp (+14-13)
- (modified) mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
(+5-1)
- (modified) mlir/test/Conversion/GPUCommon/lower-vector.mlir (+2-2)
- (modified) mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir (-5)
``diff
diff --git a/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
b/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
index 3d643c96b45008..c507b23c6d4de6 100644
--- a/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
+++ b/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
@@ -292,6 +292,10 @@ void
populateVectorBitCastLoweringPatterns(RewritePatternSet &patterns,
int64_t targetRank = 1,
PatternBenefit benefit = 1);
+/// Populates a pattern that rank-reduces n-D FMAs into (n-1)-D FMAs where
+/// n > 1.
+void populateVectorRankReducingFMAPattern(RewritePatternSet &patterns);
+
} // namespace vector
} // namespace mlir
#endif // MLIR_DIALECT_VECTOR_TRANSFORMS_LOWERINGPATTERNS_H
diff --git a/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
b/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
index 1497d662dcdbdd..2fe3b1302e5e5b 100644
--- a/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
+++ b/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
@@ -32,10 +32,12 @@
#include "mlir/Dialect/GPU/Transforms/Passes.h"
#include "mlir/Dialect/LLVMIR/LLVMDialect.h"
#include "mlir/Dialect/MemRef/IR/MemRef.h"
+#include "mlir/Dialect/Vector/Transforms/LoweringPatterns.h"
#include "mlir/IR/Attributes.h"
#include "mlir/IR/Builders.h"
#include "mlir/IR/BuiltinOps.h"
#include "mlir/IR/BuiltinTypes.h"
+#include "mlir/Transforms/GreedyPatternRewriteDriver.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/Support/Error.h"
@@ -522,6 +524,16 @@
DECLARE_CONVERT_OP_TO_GPU_RUNTIME_CALL_PATTERN(SetCsrPointersOp)
void GpuToLLVMConversionPass::runOnOperation() {
MLIRContext *context = &getContext();
+
+ // Perform progressive lowering of vector transfer operations.
+ {
+RewritePatternSet patterns(&getContext());
+// Vector transfer ops with rank > 1 should be lowered with VectorToSCF.
+vector::populateVectorTransferLoweringPatterns(patterns,
+ /*maxTransferRank=*/1);
+(void)applyPatternsAndFoldGreedily(getOperation(), std::move(patterns));
+ }
+
LowerToLLVMOptions options(context);
options.useBarePtrCallConv = hostBarePtrCallConv;
RewritePatternSet patterns(context);
diff --git a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
b/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
index a9a07c323c7358..577b74bb7e0c26 100644
--- a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
+++ b/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
@@ -1475,16 +1475,16 @@ class VectorTypeCastOpConversion
/// Conversion pattern for a `vector.create_mask` (1-D scalable vectors only).
/// Non-scalable versions of this operation are handled in Vector Transforms.
-class VectorCreateMaskOpRewritePattern
-: public OpRewritePattern {
+class VectorCreateMaskOpConversion
+: public OpConversionPattern {
public:
- explicit VectorCreateMaskOpRewritePattern(MLIRContext *context,
+ explicit VectorCreateMaskOpConversion(MLIRContext *context,
bool enableIndexOpt)
- : OpRewritePattern(context),
+ : OpConversionPattern(context),
force32BitVectorIndices(enableIndexOpt) {}
- LogicalResult matchAndRewrite(vector::CreateMaskOp op,
-PatternRewriter &rewriter) const override {
+ LogicalResult matchAndRewrite(vector::CreateMaskOp op, OpAdaptor adaptor,
+ConversionPatternRewriter &rewriter) const
override {
auto dstType = op.getType();
if (dstType.getRank() != 1 || !cast(dstType).isScalable())
return failure();
@@ -1495,7 +1495,7 @@ class VectorCreateMaskOpRewritePattern
loc, LLVM::getVectorType(idxType, dstType.getShape()[0],
/*isScalable=*/true));
auto bound = g
[llvm-branch-commits] [mlir] [mlir][Vector] Clean up `populateVectorToLLVMConversionPatterns` (PR #119975)
llvmbot wrote:
@llvm/pr-subscribers-mlir-vector
Author: Matthias Springer (matthias-springer)
Changes
Clean up `populateVectorToLLVMConversionPatterns` so that it populates only
conversion patterns. All rewrite patterns that do not lower to LLVM should be
populated into a separate greedy pattern rewrite.
The current combination of rewrite patterns and conversion patterns triggered
an edge case when merging the 1:1 and 1:N dialect conversions.
Depends on #119973.
---
Full diff: https://github.com/llvm/llvm-project/pull/119975.diff
6 Files Affected:
- (modified) mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
(+4)
- (modified) mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp (+12)
- (modified) mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp (+14-13)
- (modified) mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
(+5-1)
- (modified) mlir/test/Conversion/GPUCommon/lower-vector.mlir (+2-2)
- (modified) mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir (-5)
``diff
diff --git a/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
b/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
index 3d643c96b45008..c507b23c6d4de6 100644
--- a/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
+++ b/mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
@@ -292,6 +292,10 @@ void
populateVectorBitCastLoweringPatterns(RewritePatternSet &patterns,
int64_t targetRank = 1,
PatternBenefit benefit = 1);
+/// Populates a pattern that rank-reduces n-D FMAs into (n-1)-D FMAs where
+/// n > 1.
+void populateVectorRankReducingFMAPattern(RewritePatternSet &patterns);
+
} // namespace vector
} // namespace mlir
#endif // MLIR_DIALECT_VECTOR_TRANSFORMS_LOWERINGPATTERNS_H
diff --git a/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
b/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
index 1497d662dcdbdd..2fe3b1302e5e5b 100644
--- a/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
+++ b/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
@@ -32,10 +32,12 @@
#include "mlir/Dialect/GPU/Transforms/Passes.h"
#include "mlir/Dialect/LLVMIR/LLVMDialect.h"
#include "mlir/Dialect/MemRef/IR/MemRef.h"
+#include "mlir/Dialect/Vector/Transforms/LoweringPatterns.h"
#include "mlir/IR/Attributes.h"
#include "mlir/IR/Builders.h"
#include "mlir/IR/BuiltinOps.h"
#include "mlir/IR/BuiltinTypes.h"
+#include "mlir/Transforms/GreedyPatternRewriteDriver.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/Support/Error.h"
@@ -522,6 +524,16 @@
DECLARE_CONVERT_OP_TO_GPU_RUNTIME_CALL_PATTERN(SetCsrPointersOp)
void GpuToLLVMConversionPass::runOnOperation() {
MLIRContext *context = &getContext();
+
+ // Perform progressive lowering of vector transfer operations.
+ {
+RewritePatternSet patterns(&getContext());
+// Vector transfer ops with rank > 1 should be lowered with VectorToSCF.
+vector::populateVectorTransferLoweringPatterns(patterns,
+ /*maxTransferRank=*/1);
+(void)applyPatternsAndFoldGreedily(getOperation(), std::move(patterns));
+ }
+
LowerToLLVMOptions options(context);
options.useBarePtrCallConv = hostBarePtrCallConv;
RewritePatternSet patterns(context);
diff --git a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
b/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
index a9a07c323c7358..577b74bb7e0c26 100644
--- a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
+++ b/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
@@ -1475,16 +1475,16 @@ class VectorTypeCastOpConversion
/// Conversion pattern for a `vector.create_mask` (1-D scalable vectors only).
/// Non-scalable versions of this operation are handled in Vector Transforms.
-class VectorCreateMaskOpRewritePattern
-: public OpRewritePattern {
+class VectorCreateMaskOpConversion
+: public OpConversionPattern {
public:
- explicit VectorCreateMaskOpRewritePattern(MLIRContext *context,
+ explicit VectorCreateMaskOpConversion(MLIRContext *context,
bool enableIndexOpt)
- : OpRewritePattern(context),
+ : OpConversionPattern(context),
force32BitVectorIndices(enableIndexOpt) {}
- LogicalResult matchAndRewrite(vector::CreateMaskOp op,
-PatternRewriter &rewriter) const override {
+ LogicalResult matchAndRewrite(vector::CreateMaskOp op, OpAdaptor adaptor,
+ConversionPatternRewriter &rewriter) const
override {
auto dstType = op.getType();
if (dstType.getRank() != 1 || !cast(dstType).isScalable())
return failure();
@@ -1495,7 +1495,7 @@ class VectorCreateMaskOpRewritePattern
loc, LLVM::getVectorType(idxType, dstType.getShape()[0],
/*isScalable=*/true));
auto bo
[llvm-branch-commits] [mlir] [mlir][Vector] Clean up `populateVectorToLLVMConversionPatterns` (PR #119975)
github-actions[bot] wrote:
:warning: C/C++ code formatter, clang-format found issues in your code.
:warning:
You can test this locally with the following command:
``bash
git-clang-format --diff bf0d13553b2bc2124a266e398976ba80a1114580
e5926b63835eec731c513d91ce9c451c429ca572 --extensions cpp,h --
mlir/include/mlir/Dialect/Vector/Transforms/LoweringPatterns.h
mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
``
View the diff from clang-format here.
``diff
diff --git a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
b/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
index 577b74bb7e..9657f583c3 100644
--- a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
+++ b/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
@@ -1479,12 +1479,13 @@ class VectorCreateMaskOpConversion
: public OpConversionPattern {
public:
explicit VectorCreateMaskOpConversion(MLIRContext *context,
-bool enableIndexOpt)
+bool enableIndexOpt)
: OpConversionPattern(context),
force32BitVectorIndices(enableIndexOpt) {}
- LogicalResult matchAndRewrite(vector::CreateMaskOp op, OpAdaptor adaptor,
-ConversionPatternRewriter &rewriter) const
override {
+ LogicalResult
+ matchAndRewrite(vector::CreateMaskOp op, OpAdaptor adaptor,
+ ConversionPatternRewriter &rewriter) const override {
auto dstType = op.getType();
if (dstType.getRank() != 1 || !cast(dstType).isScalable())
return failure();
``
https://github.com/llvm/llvm-project/pull/119975
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [mlir] [mlir][Vector] Clean up `populateVectorToLLVMConversionPatterns` (PR #119975)
https://github.com/banach-space approved this pull request. The non-GPU changes LGTM. The CHECK lines removed in tests were just dead code, so thanks for the clean-up! The GPU parts look reasonable, but it might be worth waiting a few days in case someone more experienced wants to take a look. If there are no comments, I would just land this. https://github.com/llvm/llvm-project/pull/119975 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [DirectX] Split resource info into type and binding info. NFC (PR #119773)
https://github.com/hekota edited https://github.com/llvm/llvm-project/pull/119773 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [DirectX] Split resource info into type and binding info. NFC (PR #119773)
https://github.com/hekota edited https://github.com/llvm/llvm-project/pull/119773 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [DirectX] Split resource info into type and binding info. NFC (PR #119773)
@@ -303,44 +289,157 @@ class ResourceInfo {
dxil::SamplerFeedbackType getFeedbackType() const;
uint32_t getMultiSampleCount() const;
- StringRef getName() const {
-// TODO: Get the name from the symbol once we include one here.
-return "";
- }
dxil::ResourceClass getResourceClass() const { return RC; }
dxil::ResourceKind getResourceKind() const { return Kind; }
+ bool operator==(const ResourceTypeInfo &RHS) const;
+ bool operator!=(const ResourceTypeInfo &RHS) const { return !(*this == RHS);
}
+ bool operator<(const ResourceTypeInfo &RHS) const;
+
+ void print(raw_ostream &OS, const DataLayout &DL) const;
+};
+
+//===--===//
+
+class ResourceBindingInfo {
+public:
+ struct ResourceBinding {
+uint32_t RecordID;
+uint32_t Space;
+uint32_t LowerBound;
+uint32_t Size;
+
+bool operator==(const ResourceBinding &RHS) const {
+ return std::tie(RecordID, Space, LowerBound, Size) ==
+ std::tie(RHS.RecordID, RHS.Space, RHS.LowerBound, RHS.Size);
+}
+bool operator!=(const ResourceBinding &RHS) const {
+ return !(*this == RHS);
+}
+bool operator<(const ResourceBinding &RHS) const {
+ return std::tie(RecordID, Space, LowerBound, Size) <
+ std::tie(RHS.RecordID, RHS.Space, RHS.LowerBound, RHS.Size);
+}
+ };
+
+private:
+ ResourceBinding Binding;
+ TargetExtType *HandleTy;
+
+public:
+ ResourceBindingInfo(uint32_t RecordID, uint32_t Space, uint32_t LowerBound,
+ uint32_t Size, TargetExtType *HandleTy)
+ : Binding{RecordID, Space, LowerBound, Size}, HandleTy(HandleTy) {}
+
void setBindingID(unsigned ID) { Binding.RecordID = ID; }
const ResourceBinding &getBinding() const { return Binding; }
+ TargetExtType *getHandleTy() const { return HandleTy; }
+ const StringRef getName() const {
+// TODO: Get the name from the symbol once we include one here.
+return "";
+ }
- MDTuple *getAsMetadata(Module &M) const;
- std::pair getAnnotateProps(Module &M) const;
+ MDTuple *getAsMetadata(Module &M, DXILResourceTypeMap &DRTM) const;
+ MDTuple *getAsMetadata(Module &M, dxil::ResourceTypeInfo RTI) const;
- bool operator==(const ResourceInfo &RHS) const;
- bool operator!=(const ResourceInfo &RHS) const { return !(*this == RHS); }
- bool operator<(const ResourceInfo &RHS) const;
+ std::pair
+ getAnnotateProps(Module &M, DXILResourceTypeMap &DRTM) const;
+ std::pair
+ getAnnotateProps(Module &M, dxil::ResourceTypeInfo RTI) const;
- void print(raw_ostream &OS, const DataLayout &DL) const;
+ bool operator==(const ResourceBindingInfo &RHS) const {
+return std::tie(Binding, HandleTy) == std::tie(RHS.Binding, RHS.HandleTy);
+ }
+ bool operator!=(const ResourceBindingInfo &RHS) const {
+return !(*this == RHS);
+ }
+ bool operator<(const ResourceBindingInfo &RHS) const {
+return Binding < RHS.Binding;
+ }
+
+ void print(raw_ostream &OS, DXILResourceTypeMap &DRTM,
+ const DataLayout &DL) const;
+ void print(raw_ostream &OS, dxil::ResourceTypeInfo RTI,
+ const DataLayout &DL) const;
};
} // namespace dxil
//===--===//
-class DXILResourceMap {
- SmallVector Infos;
+class DXILResourceTypeMap {
+ struct Info {
+dxil::ResourceClass RC;
+dxil::ResourceKind Kind;
+bool GloballyCoherent;
+bool HasCounter;
+ };
+ DenseMap Infos;
+
+public:
+ bool invalidate(Module &M, const PreservedAnalyses &PA,
+ ModuleAnalysisManager::Invalidator &Inv);
+
+ dxil::ResourceTypeInfo operator[](TargetExtType *Ty) {
+Info I = Infos[Ty];
+return dxil::ResourceTypeInfo(Ty, I.RC, I.Kind, I.GloballyCoherent,
+ I.HasCounter);
+ }
+
+ void setGloballyCoherent(TargetExtType *Ty, bool GloballyCoherent) {
+Infos[Ty].GloballyCoherent = GloballyCoherent;
+ }
+
+ void setHasCounter(TargetExtType *Ty, bool HasCounter) {
+Infos[Ty].HasCounter = HasCounter;
+ }
+};
+
+class DXILResourceTypeAnalysis
+: public AnalysisInfoMixin {
+ friend AnalysisInfoMixin;
+
+ static AnalysisKey Key;
+
+public:
+ using Result = DXILResourceTypeMap;
+
+ DXILResourceTypeMap run(Module &M, ModuleAnalysisManager &AM) {
+return Result();
+ }
hekota wrote:
It would be nice to add a comment on this class explaining how will the
DXILResourceTypeMap get populated (that it happens in the
DXILResourceBindingAnalysis, if I understand it correctly).
https://github.com/llvm/llvm-project/pull/119773
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [DirectX] Split resource info into type and binding info. NFC (PR #119773)
https://github.com/hekota edited https://github.com/llvm/llvm-project/pull/119773 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [DirectX] Split resource info into type and binding info. NFC (PR #119773)
https://github.com/hekota edited https://github.com/llvm/llvm-project/pull/119773 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [DirectX] Create symbols for resource handles (PR #119775)
@@ -0,0 +1,48 @@
+; RUN: opt -S -passes=dxil-translate-metadata %s | FileCheck %s
+
+target triple = "dxil-pc-shadermodel6.6-compute"
+
+%struct.S = type { <4 x float>, <4 x i32> }
+
+define void @test() {
+ ; Buffer
+ %float4 = call target("dx.TypedBuffer", <4 x float>, 0, 0, 0)
+ @llvm.dx.handle.fromBinding(i32 0, i32 0, i32 1, i32 0, i1 false)
+ ; CHECK: %TypedBuffer = type { <4 x float> }
+
+ ; Buffer
+ %int = call target("dx.TypedBuffer", i32, 0, 0, 1)
+ @llvm.dx.handle.fromBinding(i32 0, i32 1, i32 1, i32 0, i1 false)
+ ; CHECK: %TypedBuffer.0 = type { i32 }
+
+ ; Buffer
+ %uint3 = call target("dx.TypedBuffer", <3 x i32>, 0, 0, 0)
+ @llvm.dx.handle.fromBinding(i32 0, i32 2, i32 1, i32 0, i1 false)
+ ; CHECK: %TypedBuffer.1 = type { <3 x i32> }
+
+ ; StructuredBuffer
+ %struct0 = call target("dx.RawBuffer", %struct.S, 0, 0)
+ @llvm.dx.handle.fromBinding(i32 0, i32 10, i32 1, i32 0, i1 true)
+ ; CHECK: %StructuredBuffer = type { %struct.S }
+
+ ; ByteAddressBuffer
+ %byteaddr = call target("dx.RawBuffer", i8, 0, 0)
+ @llvm.dx.handle.fromBinding(i32 0, i32 20, i32 1, i32 0, i1 false)
+ ; CHECK: %ByteAddressBuffer = type { i32 }
+
+ ret void
+}
+
+; CHECK: @[[T0:.*]] = external constant %TypedBuffer
+; CHECK-NEXT: @[[T1:.*]] = external constant %TypedBuffer.0
+; CHECK-NEXT: @[[T2:.*]] = external constant %TypedBuffer.1
+; CHECK-NEXT: @[[S0:.*]] = external constant %StructuredBuffer
+; CHECK-NEXT: @[[B0:.*]] = external constant %ByteAddressBuffer
+
+; CHECK: !{i32 0, ptr @[[T0]], !""
hekota wrote:
Is the name of the symbol expected to be empty string here?
https://github.com/llvm/llvm-project/pull/119775
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [DirectX] Create symbols for resource handles (PR #119775)
hekota wrote: If there already is a global variable for the resource in the module, shouldn't we be using that instead of creating a new symbol? https://github.com/llvm/llvm-project/pull/119775 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [NFC][BoundsChecking] Add TrapBB local variable (PR #119983)
llvmbot wrote:
@llvm/pr-subscribers-llvm-transforms
Author: Vitaly Buka (vitalybuka)
Changes
---
Full diff: https://github.com/llvm/llvm-project/pull/119983.diff
1 Files Affected:
- (modified) llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp (+4-2)
``diff
diff --git a/llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp
b/llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp
index c86d967716a5a0..c4511d574f2185 100644
--- a/llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp
+++ b/llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp
@@ -126,16 +126,18 @@ static void insertBoundsCheck(Value *Or, BuilderTy &IRB,
GetTrapBBT GetTrapBB) {
BasicBlock *Cont = OldBB->splitBasicBlock(SplitI);
OldBB->getTerminator()->eraseFromParent();
+ BasicBlock * TrapBB = GetTrapBB(IRB);
+
if (C) {
// If we have a constant zero, unconditionally branch.
// FIXME: We should really handle this differently to bypass the splitting
// the block.
-BranchInst::Create(GetTrapBB(IRB), OldBB);
+BranchInst::Create(TrapBB, OldBB);
return;
}
// Create the conditional branch.
- BranchInst::Create(GetTrapBB(IRB), Cont, Or, OldBB);
+ BranchInst::Create(TrapBB, Cont, Or, OldBB);
}
static bool addBoundsChecking(Function &F, TargetLibraryInfo &TLI,
``
https://github.com/llvm/llvm-project/pull/119983
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [NFC][BoundsChecking] Add TrapBB local variable (PR #119983)
https://github.com/vitalybuka converted_to_draft https://github.com/llvm/llvm-project/pull/119983 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [NFC][BoundsChecking] Add TrapBB local variable (PR #119983)
github-actions[bot] wrote:
:warning: C/C++ code formatter, clang-format found issues in your code.
:warning:
You can test this locally with the following command:
``bash
git-clang-format --diff 8bfb53a797a0d1cb2290f188ef23418c90b7d254
daf365e3e7fdc383e18e8171f1aa89a6ce8f72ec --extensions cpp --
llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp
``
View the diff from clang-format here.
``diff
diff --git a/llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp
b/llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp
index c4511d574f..769fbe92b0 100644
--- a/llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp
+++ b/llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp
@@ -126,7 +126,7 @@ static void insertBoundsCheck(Value *Or, BuilderTy &IRB,
GetTrapBBT GetTrapBB) {
BasicBlock *Cont = OldBB->splitBasicBlock(SplitI);
OldBB->getTerminator()->eraseFromParent();
- BasicBlock * TrapBB = GetTrapBB(IRB);
+ BasicBlock *TrapBB = GetTrapBB(IRB);
if (C) {
// If we have a constant zero, unconditionally branch.
``
https://github.com/llvm/llvm-project/pull/119983
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [NFC][BoundsChecking] Add TrapBB local variable (PR #119983)
https://github.com/vitalybuka created https://github.com/llvm/llvm-project/pull/119983 None ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Fix counts aggregation in merge-fdata (PR #119652)
@@ -330,14 +347,14 @@ void mergeLegacyProfiles(const
SmallVectorImpl &Filenames) {
ProfileTy MergedProfile;
for (const auto &[Thread, Profile] : ParsedProfiles)
for (const auto &[Key, Value] : Profile) {
- uint64_t Count = MergedProfile.lookup(Key) + Value;
+ auto Count = MergedProfile.lookup(Key) + Value;
aaupov wrote:
The type of `Count` is changed, made explicit in updated version
https://github.com/llvm/llvm-project/pull/119652
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Fix counts aggregation in merge-fdata (PR #119652)
https://github.com/aaupov edited https://github.com/llvm/llvm-project/pull/119652 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Fix counts aggregation in merge-fdata (PR #119652)
aaupov wrote: > Since you are switching to regex, are there any changes in parsing > performance? With updated version (on top of #119942), it's neutral/slightly faster than current merge-fdata > 1.04 ± 0.07 times faster https://github.com/llvm/llvm-project/pull/119652 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Fix counts aggregation in merge-fdata (PR #119652)
https://github.com/aaupov edited https://github.com/llvm/llvm-project/pull/119652 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [llvm] [BOLT] Fix counts aggregation in merge-fdata (PR #119652)
https://github.com/maksfb approved this pull request. LGTM https://github.com/llvm/llvm-project/pull/119652 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
