This is an archive of the discontinued LLVM Phabricator instance.

Differential D141558

[MemProf] Collect access density statistics during profiling
ClosedPublic

Authored by tejohnson on Jan 11 2023, 5:23 PM.

Details

Reviewers
snehasish
davidxl
Commits
rGee73d240ab1d: [MemProf] Collect access density statistics during profiling
Summary

Track min/max/avg access density (accesses per byte and accesses per
byte per lifetime second) metrics directly during profiling. This allows
more accurate use of these metrics in profile analysis and use, instead
of trying to compute them from already aggregated data in the profile.

This required regenerating some of the raw profile and executable inputs
for a few tests. While here, make the llvm-profdata memprof tests more
resilient to differences in things like memory mapping, timestamps and
cpu ids to make future test updates easier.

Diff Detail

Event Timeline

tejohnson created this revision.Jan 11 2023, 5:23 PM
Herald added a project: Restricted Project. · View Herald TranscriptJan 11 2023, 5:23 PM
Herald added subscribers: Enna1, wenlei. · View Herald Transcript
tejohnson requested review of this revision.Jan 11 2023, 5:23 PM
Herald added projects: Restricted Project, Restricted Project, Restricted Project. · View Herald TranscriptJan 11 2023, 5:23 PM
Herald added subscribers: Restricted Project, cfe-commits. · View Herald Transcript
Harbormaster completed remote builds in B207251: Diff 488434.Jan 11 2023, 8:06 PM
snehasish added a comment.Jan 12 2023, 11:21 AM

Thanks for updating the tests.

llvm/include/llvm/ProfileData/MemProfData.inc
35

Can you increment the version here? Then we won't accidentally read the old data with the new reader which expects more data per MemInfoBlock.

Also update the other copy of MemProfData.inc.

llvm/test/tools/llvm-profdata/memprof-multi.test
43–44

Use the same regex for NumSegments here?

tejohnson marked 2 inline comments as done.Jan 12 2023, 1:52 PM
tejohnson updated this revision to Diff 488766.Jan 12 2023, 1:52 PM

Address comments (version change necessitated updating the tests)

snehasish accepted this revision.Jan 12 2023, 2:02 PM

lgtm

This revision is now accepted and ready to land.Jan 12 2023, 2:02 PM
Harbormaster completed remote builds in B207478: Diff 488766.Jan 12 2023, 3:01 PM
This revision was landed with ongoing or failed builds.Jan 12 2023, 5:54 PM
Closed by commit rGee73d240ab1d: [MemProf] Collect access density statistics during profiling (authored by tejohnson). · Explain Why
This revision was automatically updated to reflect the committed changes.
tejohnson added a commit: rGee73d240ab1d: [MemProf] Collect access density statistics during profiling.

Revision Contents

PathSize
clang/
test/
CodeGen/
Inputs/
compiler-rt/
include/
profile/
6 lines
33 lines
llvm/
include/
llvm/
ProfileData/
6 lines
33 lines
test/
Transforms/
PGOProfile/
Inputs/
tools/
llvm-profdata/
Inputs/
82 lines
80 lines
4 lines

Diff 488829

clang/test/CodeGen/Inputs/memprof.exe

  • This is a binary file.

clang/test/CodeGen/Inputs/memprof.memprofraw

  • This is a binary file.

compiler-rt/include/profile/MIBEntryDef.inc

Show All 39 Lines
MIBEntryDef(MaxLifetime = 12, MaxLifetime, uint32_t) MIBEntryDef(MaxLifetime = 12, MaxLifetime, uint32_t)
MIBEntryDef(AllocCpuId = 13, AllocCpuId, uint32_t) MIBEntryDef(AllocCpuId = 13, AllocCpuId, uint32_t)
MIBEntryDef(DeallocCpuId = 14, DeallocCpuId, uint32_t) MIBEntryDef(DeallocCpuId = 14, DeallocCpuId, uint32_t)
MIBEntryDef(NumMigratedCpu = 15, NumMigratedCpu, uint32_t) MIBEntryDef(NumMigratedCpu = 15, NumMigratedCpu, uint32_t)
MIBEntryDef(NumLifetimeOverlaps = 16, NumLifetimeOverlaps, uint32_t) MIBEntryDef(NumLifetimeOverlaps = 16, NumLifetimeOverlaps, uint32_t)
MIBEntryDef(NumSameAllocCpu = 17, NumSameAllocCpu, uint32_t) MIBEntryDef(NumSameAllocCpu = 17, NumSameAllocCpu, uint32_t)
MIBEntryDef(NumSameDeallocCpu = 18, NumSameDeallocCpu, uint32_t) MIBEntryDef(NumSameDeallocCpu = 18, NumSameDeallocCpu, uint32_t)
MIBEntryDef(DataTypeId = 19, DataTypeId, uint64_t) MIBEntryDef(DataTypeId = 19, DataTypeId, uint64_t)
MIBEntryDef(TotalAccessDensity = 20, TotalAccessDensity, uint64_t)
MIBEntryDef(MinAccessDensity = 21, MinAccessDensity, uint32_t)
MIBEntryDef(MaxAccessDensity = 22, MaxAccessDensity, uint32_t)
MIBEntryDef(TotalLifetimeAccessDensity = 23, TotalLifetimeAccessDensity, uint64_t)
MIBEntryDef(MinLifetimeAccessDensity = 24, MinLifetimeAccessDensity, uint32_t)
MIBEntryDef(MaxLifetimeAccessDensity = 25, MaxLifetimeAccessDensity, uint32_t)

compiler-rt/include/profile/MemProfData.inc

Show All 26 Lines
#endif #endif
// A 64-bit magic number to uniquely identify the raw binary memprof profile file. // A 64-bit magic number to uniquely identify the raw binary memprof profile file.
#define MEMPROF_RAW_MAGIC_64 \ #define MEMPROF_RAW_MAGIC_64 \
((uint64_t)255 << 56 | (uint64_t)'m' << 48 | (uint64_t)'p' << 40 | (uint64_t)'r' << 32 | \ ((uint64_t)255 << 56 | (uint64_t)'m' << 48 | (uint64_t)'p' << 40 | (uint64_t)'r' << 32 | \
(uint64_t)'o' << 24 | (uint64_t)'f' << 16 | (uint64_t)'r' << 8 | (uint64_t)129) (uint64_t)'o' << 24 | (uint64_t)'f' << 16 | (uint64_t)'r' << 8 | (uint64_t)129)
// The version number of the raw binary format. // The version number of the raw binary format.
#define MEMPROF_RAW_VERSION 1ULL #define MEMPROF_RAW_VERSION 2ULL
namespace llvm { namespace llvm {
namespace memprof { namespace memprof {
// A struct describing the header used for the raw binary memprof profile format. // A struct describing the header used for the raw binary memprof profile format.
PACKED(struct Header { PACKED(struct Header {
uint64_t Magic; uint64_t Magic;
uint64_t Version; uint64_t Version;
uint64_t TotalSize; uint64_t TotalSize;
▲ Show 20 Lines • Show All 78 Lines • ▼ Show 20 LinesMemInfoBlock(uint32_t Size, uint64_t AccessCount, uint32_t AllocTs,
TotalSize = Size; TotalSize = Size;
MinSize = Size; MinSize = Size;
MaxSize = Size; MaxSize = Size;
AllocTimestamp = AllocTs; AllocTimestamp = AllocTs;
DeallocTimestamp = DeallocTs; DeallocTimestamp = DeallocTs;
TotalLifetime = DeallocTimestamp - AllocTimestamp; TotalLifetime = DeallocTimestamp - AllocTimestamp;
MinLifetime = TotalLifetime; MinLifetime = TotalLifetime;
MaxLifetime = TotalLifetime; MaxLifetime = TotalLifetime;
// Access density is accesses per byte. Multiply by 100 to include the
// fractional part.
TotalAccessDensity = AccessCount * 100 / Size;
MinAccessDensity = TotalAccessDensity;
MaxAccessDensity = TotalAccessDensity;
// Lifetime access density is the access density per second of lifetime.
// Multiply by 1000 to convert denominator lifetime to seconds (using a
// minimum lifetime of 1ms to avoid divide by 0. Do the multiplication first
// to reduce truncations to 0.
TotalLifetimeAccessDensity =
TotalAccessDensity * 1000 / (TotalLifetime ? TotalLifetime : 1);
MinLifetimeAccessDensity = TotalLifetimeAccessDensity;
MaxLifetimeAccessDensity = TotalLifetimeAccessDensity;
AllocCpuId = AllocCpu; AllocCpuId = AllocCpu;
DeallocCpuId = DeallocCpu; DeallocCpuId = DeallocCpu;
NumMigratedCpu = AllocCpuId != DeallocCpuId; NumMigratedCpu = AllocCpuId != DeallocCpuId;
} }
void Merge(const MemInfoBlock &newMIB) { void Merge(const MemInfoBlock &newMIB) {
AllocCount += newMIB.AllocCount; AllocCount += newMIB.AllocCount;
TotalAccessCount += newMIB.TotalAccessCount; TotalAccessCount += newMIB.TotalAccessCount;
MinAccessCount = newMIB.MinAccessCount < MinAccessCount ? newMIB.MinAccessCount : MinAccessCount; MinAccessCount = newMIB.MinAccessCount < MinAccessCount ? newMIB.MinAccessCount : MinAccessCount;
MaxAccessCount = newMIB.MaxAccessCount > MaxAccessCount ? newMIB.MaxAccessCount : MaxAccessCount; MaxAccessCount = newMIB.MaxAccessCount > MaxAccessCount ? newMIB.MaxAccessCount : MaxAccessCount;
TotalSize += newMIB.TotalSize; TotalSize += newMIB.TotalSize;
MinSize = newMIB.MinSize < MinSize ? newMIB.MinSize : MinSize; MinSize = newMIB.MinSize < MinSize ? newMIB.MinSize : MinSize;
MaxSize = newMIB.MaxSize > MaxSize ? newMIB.MaxSize : MaxSize; MaxSize = newMIB.MaxSize > MaxSize ? newMIB.MaxSize : MaxSize;
TotalLifetime += newMIB.TotalLifetime; TotalLifetime += newMIB.TotalLifetime;
MinLifetime = newMIB.MinLifetime < MinLifetime ? newMIB.MinLifetime : MinLifetime; MinLifetime = newMIB.MinLifetime < MinLifetime ? newMIB.MinLifetime : MinLifetime;
MaxLifetime = newMIB.MaxLifetime > MaxLifetime ? newMIB.MaxLifetime : MaxLifetime; MaxLifetime = newMIB.MaxLifetime > MaxLifetime ? newMIB.MaxLifetime : MaxLifetime;
TotalAccessDensity += newMIB.TotalAccessDensity;
MinAccessDensity = newMIB.MinAccessDensity < MinAccessDensity
? newMIB.MinAccessDensity
: MinAccessDensity;
MaxAccessDensity = newMIB.MaxAccessDensity > MaxAccessDensity
? newMIB.MaxAccessDensity
: MaxAccessDensity;
TotalLifetimeAccessDensity += newMIB.TotalLifetimeAccessDensity;
MinLifetimeAccessDensity =
newMIB.MinLifetimeAccessDensity < MinLifetimeAccessDensity
? newMIB.MinLifetimeAccessDensity
: MinLifetimeAccessDensity;
MaxLifetimeAccessDensity =
newMIB.MaxLifetimeAccessDensity > MaxLifetimeAccessDensity
? newMIB.MaxLifetimeAccessDensity
: MaxLifetimeAccessDensity;
// We know newMIB was deallocated later, so just need to check if it was // We know newMIB was deallocated later, so just need to check if it was
// allocated before last one deallocated. // allocated before last one deallocated.
NumLifetimeOverlaps += newMIB.AllocTimestamp < DeallocTimestamp; NumLifetimeOverlaps += newMIB.AllocTimestamp < DeallocTimestamp;
AllocTimestamp = newMIB.AllocTimestamp; AllocTimestamp = newMIB.AllocTimestamp;
DeallocTimestamp = newMIB.DeallocTimestamp; DeallocTimestamp = newMIB.DeallocTimestamp;
NumSameAllocCpu += AllocCpuId == newMIB.AllocCpuId; NumSameAllocCpu += AllocCpuId == newMIB.AllocCpuId;
NumSameDeallocCpu += DeallocCpuId == newMIB.DeallocCpuId; NumSameDeallocCpu += DeallocCpuId == newMIB.DeallocCpuId;
Show All 14 Lines

llvm/include/llvm/ProfileData/MIBEntryDef.inc

Show All 39 Lines
MIBEntryDef(MaxLifetime = 12, MaxLifetime, uint32_t) MIBEntryDef(MaxLifetime = 12, MaxLifetime, uint32_t)
MIBEntryDef(AllocCpuId = 13, AllocCpuId, uint32_t) MIBEntryDef(AllocCpuId = 13, AllocCpuId, uint32_t)
MIBEntryDef(DeallocCpuId = 14, DeallocCpuId, uint32_t) MIBEntryDef(DeallocCpuId = 14, DeallocCpuId, uint32_t)
MIBEntryDef(NumMigratedCpu = 15, NumMigratedCpu, uint32_t) MIBEntryDef(NumMigratedCpu = 15, NumMigratedCpu, uint32_t)
MIBEntryDef(NumLifetimeOverlaps = 16, NumLifetimeOverlaps, uint32_t) MIBEntryDef(NumLifetimeOverlaps = 16, NumLifetimeOverlaps, uint32_t)
MIBEntryDef(NumSameAllocCpu = 17, NumSameAllocCpu, uint32_t) MIBEntryDef(NumSameAllocCpu = 17, NumSameAllocCpu, uint32_t)
MIBEntryDef(NumSameDeallocCpu = 18, NumSameDeallocCpu, uint32_t) MIBEntryDef(NumSameDeallocCpu = 18, NumSameDeallocCpu, uint32_t)
MIBEntryDef(DataTypeId = 19, DataTypeId, uint64_t) MIBEntryDef(DataTypeId = 19, DataTypeId, uint64_t)
MIBEntryDef(TotalAccessDensity = 20, TotalAccessDensity, uint64_t)
MIBEntryDef(MinAccessDensity = 21, MinAccessDensity, uint32_t)
MIBEntryDef(MaxAccessDensity = 22, MaxAccessDensity, uint32_t)
MIBEntryDef(TotalLifetimeAccessDensity = 23, TotalLifetimeAccessDensity, uint64_t)
MIBEntryDef(MinLifetimeAccessDensity = 24, MinLifetimeAccessDensity, uint32_t)
MIBEntryDef(MaxLifetimeAccessDensity = 25, MaxLifetimeAccessDensity, uint32_t)

llvm/include/llvm/ProfileData/MemProfData.inc

Show All 26 Lines
#endif #endif
// A 64-bit magic number to uniquely identify the raw binary memprof profile file. // A 64-bit magic number to uniquely identify the raw binary memprof profile file.
#define MEMPROF_RAW_MAGIC_64 \ #define MEMPROF_RAW_MAGIC_64 \
((uint64_t)255 << 56 | (uint64_t)'m' << 48 | (uint64_t)'p' << 40 | (uint64_t)'r' << 32 | \ ((uint64_t)255 << 56 | (uint64_t)'m' << 48 | (uint64_t)'p' << 40 | (uint64_t)'r' << 32 | \
(uint64_t)'o' << 24 | (uint64_t)'f' << 16 | (uint64_t)'r' << 8 | (uint64_t)129) (uint64_t)'o' << 24 | (uint64_t)'f' << 16 | (uint64_t)'r' << 8 | (uint64_t)129)
// The version number of the raw binary format. // The version number of the raw binary format.
#define MEMPROF_RAW_VERSION 1ULL #define MEMPROF_RAW_VERSION 2ULL
snehasishUnsubmitted
Done
ReplyInline Actions

Can you increment the version here? Then we won't accidentally read the old data with the new reader which expects more data per MemInfoBlock.

Also update the other copy of MemProfData.inc.

snehasish: Can you increment the version here? Then we won't accidentally read the old data with the new…
namespace llvm { namespace llvm {
namespace memprof { namespace memprof {
// A struct describing the header used for the raw binary memprof profile format. // A struct describing the header used for the raw binary memprof profile format.
PACKED(struct Header { PACKED(struct Header {
uint64_t Magic; uint64_t Magic;
uint64_t Version; uint64_t Version;
uint64_t TotalSize; uint64_t TotalSize;
▲ Show 20 Lines • Show All 78 Lines • ▼ Show 20 LinesMemInfoBlock(uint32_t Size, uint64_t AccessCount, uint32_t AllocTs,
TotalSize = Size; TotalSize = Size;
MinSize = Size; MinSize = Size;
MaxSize = Size; MaxSize = Size;
AllocTimestamp = AllocTs; AllocTimestamp = AllocTs;
DeallocTimestamp = DeallocTs; DeallocTimestamp = DeallocTs;
TotalLifetime = DeallocTimestamp - AllocTimestamp; TotalLifetime = DeallocTimestamp - AllocTimestamp;
MinLifetime = TotalLifetime; MinLifetime = TotalLifetime;
MaxLifetime = TotalLifetime; MaxLifetime = TotalLifetime;
// Access density is accesses per byte. Multiply by 100 to include the
// fractional part.
TotalAccessDensity = AccessCount * 100 / Size;
MinAccessDensity = TotalAccessDensity;
MaxAccessDensity = TotalAccessDensity;
// Lifetime access density is the access density per second of lifetime.
// Multiply by 1000 to convert denominator lifetime to seconds (using a
// minimum lifetime of 1ms to avoid divide by 0. Do the multiplication first
// to reduce truncations to 0.
TotalLifetimeAccessDensity =
TotalAccessDensity * 1000 / (TotalLifetime ? TotalLifetime : 1);
MinLifetimeAccessDensity = TotalLifetimeAccessDensity;
MaxLifetimeAccessDensity = TotalLifetimeAccessDensity;
AllocCpuId = AllocCpu; AllocCpuId = AllocCpu;
DeallocCpuId = DeallocCpu; DeallocCpuId = DeallocCpu;
NumMigratedCpu = AllocCpuId != DeallocCpuId; NumMigratedCpu = AllocCpuId != DeallocCpuId;
} }
void Merge(const MemInfoBlock &newMIB) { void Merge(const MemInfoBlock &newMIB) {
AllocCount += newMIB.AllocCount; AllocCount += newMIB.AllocCount;
TotalAccessCount += newMIB.TotalAccessCount; TotalAccessCount += newMIB.TotalAccessCount;
MinAccessCount = newMIB.MinAccessCount < MinAccessCount ? newMIB.MinAccessCount : MinAccessCount; MinAccessCount = newMIB.MinAccessCount < MinAccessCount ? newMIB.MinAccessCount : MinAccessCount;
MaxAccessCount = newMIB.MaxAccessCount > MaxAccessCount ? newMIB.MaxAccessCount : MaxAccessCount; MaxAccessCount = newMIB.MaxAccessCount > MaxAccessCount ? newMIB.MaxAccessCount : MaxAccessCount;
TotalSize += newMIB.TotalSize; TotalSize += newMIB.TotalSize;
MinSize = newMIB.MinSize < MinSize ? newMIB.MinSize : MinSize; MinSize = newMIB.MinSize < MinSize ? newMIB.MinSize : MinSize;
MaxSize = newMIB.MaxSize > MaxSize ? newMIB.MaxSize : MaxSize; MaxSize = newMIB.MaxSize > MaxSize ? newMIB.MaxSize : MaxSize;
TotalLifetime += newMIB.TotalLifetime; TotalLifetime += newMIB.TotalLifetime;
MinLifetime = newMIB.MinLifetime < MinLifetime ? newMIB.MinLifetime : MinLifetime; MinLifetime = newMIB.MinLifetime < MinLifetime ? newMIB.MinLifetime : MinLifetime;
MaxLifetime = newMIB.MaxLifetime > MaxLifetime ? newMIB.MaxLifetime : MaxLifetime; MaxLifetime = newMIB.MaxLifetime > MaxLifetime ? newMIB.MaxLifetime : MaxLifetime;
TotalAccessDensity += newMIB.TotalAccessDensity;
MinAccessDensity = newMIB.MinAccessDensity < MinAccessDensity
? newMIB.MinAccessDensity
: MinAccessDensity;
MaxAccessDensity = newMIB.MaxAccessDensity > MaxAccessDensity
? newMIB.MaxAccessDensity
: MaxAccessDensity;
TotalLifetimeAccessDensity += newMIB.TotalLifetimeAccessDensity;
MinLifetimeAccessDensity =
newMIB.MinLifetimeAccessDensity < MinLifetimeAccessDensity
? newMIB.MinLifetimeAccessDensity
: MinLifetimeAccessDensity;
MaxLifetimeAccessDensity =
newMIB.MaxLifetimeAccessDensity > MaxLifetimeAccessDensity
? newMIB.MaxLifetimeAccessDensity
: MaxLifetimeAccessDensity;
// We know newMIB was deallocated later, so just need to check if it was // We know newMIB was deallocated later, so just need to check if it was
// allocated before last one deallocated. // allocated before last one deallocated.
NumLifetimeOverlaps += newMIB.AllocTimestamp < DeallocTimestamp; NumLifetimeOverlaps += newMIB.AllocTimestamp < DeallocTimestamp;
AllocTimestamp = newMIB.AllocTimestamp; AllocTimestamp = newMIB.AllocTimestamp;
DeallocTimestamp = newMIB.DeallocTimestamp; DeallocTimestamp = newMIB.DeallocTimestamp;
NumSameAllocCpu += AllocCpuId == newMIB.AllocCpuId; NumSameAllocCpu += AllocCpuId == newMIB.AllocCpuId;
NumSameDeallocCpu += DeallocCpuId == newMIB.DeallocCpuId; NumSameDeallocCpu += DeallocCpuId == newMIB.DeallocCpuId;
Show All 14 Lines

llvm/test/Transforms/PGOProfile/Inputs/memprof.exe

  • This is a binary file.

llvm/test/Transforms/PGOProfile/Inputs/memprof.memprofraw

  • This is a binary file.

llvm/test/tools/llvm-profdata/Inputs/basic.memprofexe

  • This is a binary file.

llvm/test/tools/llvm-profdata/Inputs/basic.memprofraw

  • This is a binary file.

llvm/test/tools/llvm-profdata/Inputs/inline.memprofexe

  • This is a binary file.

llvm/test/tools/llvm-profdata/Inputs/inline.memprofraw

  • This is a binary file.
PropertyOld ValueNew Value
File Mode100644100755

llvm/test/tools/llvm-profdata/Inputs/multi.memprofexe

  • This is a binary file.

llvm/test/tools/llvm-profdata/Inputs/multi.memprofraw

  • This is a binary file.

llvm/test/tools/llvm-profdata/Inputs/pic.memprofexe

  • This is a binary file.

llvm/test/tools/llvm-profdata/Inputs/pic.memprofraw

  • This is a binary file.

llvm/test/tools/llvm-profdata/memprof-basic.test

Show All 33 Lines
RUN: llvm-profdata show --memory %p/Inputs/basic.memprofraw --profiled-binary %p/Inputs/basic.memprofexe -o - | FileCheck %s RUN: llvm-profdata show --memory %p/Inputs/basic.memprofraw --profiled-binary %p/Inputs/basic.memprofexe -o - | FileCheck %s
We expect 2 MIB entries, 1 each for the malloc calls in the program. Any We expect 2 MIB entries, 1 each for the malloc calls in the program. Any
additional allocations which do not originate from the main binary are pruned. additional allocations which do not originate from the main binary are pruned.
CHECK: MemprofProfile: CHECK: MemprofProfile:
CHECK-NEXT: Summary: CHECK-NEXT: Summary:
CHECK-NEXT: Version: 1 CHECK-NEXT: Version: 2
CHECK-NEXT: NumSegments: 9 CHECK-NEXT: NumSegments: {{[0-9]+}}
CHECK-NEXT: NumMibInfo: 2 CHECK-NEXT: NumMibInfo: 2
CHECK-NEXT: NumAllocFunctions: 1 CHECK-NEXT: NumAllocFunctions: 1
CHECK-NEXT: NumStackOffsets: 2 CHECK-NEXT: NumStackOffsets: 2
CHECK-NEXT: Segments: CHECK-NEXT: Segments:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: BuildId: <None> CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x200000 CHECK-NEXT: Start: 0x{{[0-9]+}}
CHECK-NEXT: End: 0x298000 CHECK-NEXT: End: 0x{{[0-9]+}}
CHECK-NEXT: Offset: 0x0 CHECK-NEXT: Offset: 0x{{[0-9]+}}
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7FFFF7C7C000 CHECK: Records:
CHECK-NEXT: End: 0x7FFFF7DC5000
CHECK-NEXT: Offset: 0x26000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7FFFF7E1E000
CHECK-NEXT: End: 0x7FFFF7E30000
CHECK-NEXT: Offset: 0x3000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7FFFF7E36000
CHECK-NEXT: End: 0x7FFFF7E38000
CHECK-NEXT: Offset: 0x1000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7FFFF7E4A000
CHECK-NEXT: End: 0x7FFFF7EE5000
CHECK-NEXT: Offset: 0xF000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7FFFF7F83000
CHECK-NEXT: End: 0x7FFFF7F87000
CHECK-NEXT: Offset: 0x3000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7FFFF7F92000
CHECK-NEXT: End: 0x7FFFF7FA1000
CHECK-NEXT: Offset: 0x7000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7FFFF7FD0000
CHECK-NEXT: End: 0x7FFFF7FD2000
CHECK-NEXT: Offset: 0x0
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7FFFF7FD3000
CHECK-NEXT: End: 0x7FFFF7FF3000
CHECK-NEXT: Offset: 0x1000
CHECK-NEXT: Records:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: FunctionGUID: {{[0-9]+}} CHECK-NEXT: FunctionGUID: {{[0-9]+}}
CHECK-NEXT: AllocSites: CHECK-NEXT: AllocSites:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: Callstack: CHECK-NEXT: Callstack:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: Function: {{[0-9]+}} CHECK-NEXT: Function: {{[0-9]+}}
CHECK-NEXT: SymbolName: main CHECK-NEXT: SymbolName: main
CHECK-NEXT: LineOffset: 1 CHECK-NEXT: LineOffset: 1
CHECK-NEXT: Column: 21 CHECK-NEXT: Column: 21
CHECK-NEXT: Inline: 0 CHECK-NEXT: Inline: 0
CHECK-NEXT: MemInfoBlock: CHECK-NEXT: MemInfoBlock:
CHECK-NEXT: AllocCount: 1 CHECK-NEXT: AllocCount: 1
CHECK-NEXT: TotalAccessCount: 2 CHECK-NEXT: TotalAccessCount: 2
CHECK-NEXT: MinAccessCount: 2 CHECK-NEXT: MinAccessCount: 2
CHECK-NEXT: MaxAccessCount: 2 CHECK-NEXT: MaxAccessCount: 2
CHECK-NEXT: TotalSize: 10 CHECK-NEXT: TotalSize: 10
CHECK-NEXT: MinSize: 10 CHECK-NEXT: MinSize: 10
CHECK-NEXT: MaxSize: 10 CHECK-NEXT: MaxSize: 10
CHECK-NEXT: AllocTimestamp: 986 CHECK-NEXT: AllocTimestamp: {{[0-9]+}}
CHECK-NEXT: DeallocTimestamp: 986 CHECK-NEXT: DeallocTimestamp: {{[0-9]+}}
CHECK-NEXT: TotalLifetime: 0 CHECK-NEXT: TotalLifetime: 0
CHECK-NEXT: MinLifetime: 0 CHECK-NEXT: MinLifetime: 0
CHECK-NEXT: MaxLifetime: 0 CHECK-NEXT: MaxLifetime: 0
CHECK-NEXT: AllocCpuId: 56 CHECK-NEXT: AllocCpuId: {{[0-9]+}}
CHECK-NEXT: DeallocCpuId: 56 CHECK-NEXT: DeallocCpuId: {{[0-9]+}}
CHECK-NEXT: NumMigratedCpu: 0 CHECK-NEXT: NumMigratedCpu: 0
CHECK-NEXT: NumLifetimeOverlaps: 0 CHECK-NEXT: NumLifetimeOverlaps: 0
CHECK-NEXT: NumSameAllocCpu: 0 CHECK-NEXT: NumSameAllocCpu: 0
CHECK-NEXT: NumSameDeallocCpu: 0 CHECK-NEXT: NumSameDeallocCpu: 0
CHECK-NEXT: DataTypeId: {{[0-9]+}} CHECK-NEXT: DataTypeId: {{[0-9]+}}
CHECK-NEXT: TotalAccessDensity: 20
CHECK-NEXT: MinAccessDensity: 20
CHECK-NEXT: MaxAccessDensity: 20
CHECK-NEXT: TotalLifetimeAccessDensity: 20000
CHECK-NEXT: MinLifetimeAccessDensity: 20000
CHECK-NEXT: MaxLifetimeAccessDensity: 20000
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: Callstack: CHECK-NEXT: Callstack:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: Function: {{[0-9]+}} CHECK-NEXT: Function: {{[0-9]+}}
CHECK-NEXT: SymbolName: main CHECK-NEXT: SymbolName: main
CHECK-NEXT: LineOffset: 5 CHECK-NEXT: LineOffset: 4
CHECK-NEXT: Column: 15 CHECK-NEXT: Column: 15
CHECK-NEXT: Inline: 0 CHECK-NEXT: Inline: 0
CHECK-NEXT: MemInfoBlock: CHECK-NEXT: MemInfoBlock:
CHECK-NEXT: AllocCount: 1 CHECK-NEXT: AllocCount: 1
CHECK-NEXT: TotalAccessCount: 2 CHECK-NEXT: TotalAccessCount: 2
CHECK-NEXT: MinAccessCount: 2 CHECK-NEXT: MinAccessCount: 2
CHECK-NEXT: MaxAccessCount: 2 CHECK-NEXT: MaxAccessCount: 2
CHECK-NEXT: TotalSize: 10 CHECK-NEXT: TotalSize: 10
CHECK-NEXT: MinSize: 10 CHECK-NEXT: MinSize: 10
CHECK-NEXT: MaxSize: 10 CHECK-NEXT: MaxSize: 10
CHECK-NEXT: AllocTimestamp: 987 CHECK-NEXT: AllocTimestamp: {{[0-9]+}}
CHECK-NEXT: DeallocTimestamp: 987 CHECK-NEXT: DeallocTimestamp: {{[0-9]+}}
CHECK-NEXT: TotalLifetime: 0 CHECK-NEXT: TotalLifetime: 0
CHECK-NEXT: MinLifetime: 0 CHECK-NEXT: MinLifetime: 0
CHECK-NEXT: MaxLifetime: 0 CHECK-NEXT: MaxLifetime: 0
CHECK-NEXT: AllocCpuId: 56 CHECK-NEXT: AllocCpuId: {{[0-9]+}}
CHECK-NEXT: DeallocCpuId: 56 CHECK-NEXT: DeallocCpuId: {{[0-9]+}}
CHECK-NEXT: NumMigratedCpu: 0 CHECK-NEXT: NumMigratedCpu: 0
CHECK-NEXT: NumLifetimeOverlaps: 0 CHECK-NEXT: NumLifetimeOverlaps: 0
CHECK-NEXT: NumSameAllocCpu: 0 CHECK-NEXT: NumSameAllocCpu: 0
CHECK-NEXT: NumSameDeallocCpu: 0 CHECK-NEXT: NumSameDeallocCpu: 0
CHECK-NEXT: DataTypeId: {{[0-9]+}} CHECK-NEXT: DataTypeId: {{[0-9]+}}
CHECK-NEXT: TotalAccessDensity: 20
CHECK-NEXT: MinAccessDensity: 20
CHECK-NEXT: MaxAccessDensity: 20
CHECK-NEXT: TotalLifetimeAccessDensity: 20000
CHECK-NEXT: MinLifetimeAccessDensity: 20000
CHECK-NEXT: MaxLifetimeAccessDensity: 20000

llvm/test/tools/llvm-profdata/memprof-inline.test

Show All 34 Lines
env MEMPROF_OPTIONS=log_path=stdout ./inline.memprofexe > inline.memprofraw env MEMPROF_OPTIONS=log_path=stdout ./inline.memprofexe > inline.memprofraw
``` ```
RUN: llvm-profdata show --memory %p/Inputs/inline.memprofraw --profiled-binary %p/Inputs/inline.memprofexe | FileCheck %s RUN: llvm-profdata show --memory %p/Inputs/inline.memprofraw --profiled-binary %p/Inputs/inline.memprofexe | FileCheck %s
CHECK: MemprofProfile: CHECK: MemprofProfile:
CHECK-NEXT: Summary: CHECK-NEXT: Summary:
CHECK-NEXT: Version: 1 CHECK-NEXT: Version: 2
CHECK-NEXT: NumSegments: 9 CHECK-NEXT: NumSegments: {{[0-9]+}}
CHECK-NEXT: NumMibInfo: 2 CHECK-NEXT: NumMibInfo: 2
CHECK-NEXT: NumAllocFunctions: 2 CHECK-NEXT: NumAllocFunctions: 2
CHECK-NEXT: NumStackOffsets: 1 CHECK-NEXT: NumStackOffsets: 1
CHECK-NEXT: Segments: CHECK-NEXT: Segments:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: BuildId: <None> CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x200000 CHECK-NEXT: Start: 0x{{[0-9]+}}
CHECK-NEXT: End: 0x29B000 CHECK-NEXT: End: 0x{{[0-9]+}}
CHECK-NEXT: Offset: 0x0 CHECK-NEXT: Offset: 0x{{[0-9]+}}
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7F5871485000 CHECK: Records:
CHECK-NEXT: End: 0x7F58715CD000
CHECK-NEXT: Offset: 0x26000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7F587162D000
CHECK-NEXT: End: 0x7F587163F000
CHECK-NEXT: Offset: 0x3000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7F5871646000
CHECK-NEXT: End: 0x7F5871648000
CHECK-NEXT: Offset: 0x2000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7F587165A000
CHECK-NEXT: End: 0x7F58716F4000
CHECK-NEXT: Offset: 0xF000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7F5871791000
CHECK-NEXT: End: 0x7F5871795000
CHECK-NEXT: Offset: 0x3000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7F58717A0000
CHECK-NEXT: End: 0x7F58717AF000
CHECK-NEXT: Offset: 0x7000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7F58717D6000
CHECK-NEXT: End: 0x7F58717FA000
CHECK-NEXT: Offset: 0x1000
CHECK-NEXT: -
CHECK-NEXT: BuildId: <None>
CHECK-NEXT: Start: 0x7FFFC77BD000
CHECK-NEXT: End: 0x7FFFC77BF000
CHECK-NEXT: Offset: 0x0
CHECK-NEXT: Records:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: FunctionGUID: 15505678318020221912 CHECK-NEXT: FunctionGUID: 15505678318020221912
CHECK-NEXT: AllocSites: CHECK-NEXT: AllocSites:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: Callstack: CHECK-NEXT: Callstack:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: Function: 15505678318020221912 CHECK-NEXT: Function: 15505678318020221912
CHECK-NEXT: SymbolName: qux CHECK-NEXT: SymbolName: qux
Show All 21 Lines
CHECK-NEXT: MemInfoBlock: CHECK-NEXT: MemInfoBlock:
CHECK-NEXT: AllocCount: 1 CHECK-NEXT: AllocCount: 1
CHECK-NEXT: TotalAccessCount: 1 CHECK-NEXT: TotalAccessCount: 1
CHECK-NEXT: MinAccessCount: 1 CHECK-NEXT: MinAccessCount: 1
CHECK-NEXT: MaxAccessCount: 1 CHECK-NEXT: MaxAccessCount: 1
CHECK-NEXT: TotalSize: 1 CHECK-NEXT: TotalSize: 1
CHECK-NEXT: MinSize: 1 CHECK-NEXT: MinSize: 1
CHECK-NEXT: MaxSize: 1 CHECK-NEXT: MaxSize: 1
CHECK-NEXT: AllocTimestamp: 894 CHECK-NEXT: AllocTimestamp: {{[0-9]+}}
CHECK-NEXT: DeallocTimestamp: 894 CHECK-NEXT: DeallocTimestamp: {{[0-9]+}}
CHECK-NEXT: TotalLifetime: 0 CHECK-NEXT: TotalLifetime: 0
CHECK-NEXT: MinLifetime: 0 CHECK-NEXT: MinLifetime: 0
CHECK-NEXT: MaxLifetime: 0 CHECK-NEXT: MaxLifetime: 0
CHECK-NEXT: AllocCpuId: 23 CHECK-NEXT: AllocCpuId: {{[0-9]+}}
CHECK-NEXT: DeallocCpuId: 23 CHECK-NEXT: DeallocCpuId: {{[0-9]+}}
CHECK-NEXT: NumMigratedCpu: 0 CHECK-NEXT: NumMigratedCpu: 0
CHECK-NEXT: NumLifetimeOverlaps: 0 CHECK-NEXT: NumLifetimeOverlaps: 0
CHECK-NEXT: NumSameAllocCpu: 0 CHECK-NEXT: NumSameAllocCpu: 0
CHECK-NEXT: NumSameDeallocCpu: 0 CHECK-NEXT: NumSameDeallocCpu: 0
CHECK-NEXT: DataTypeId: {{[0-9]+}} CHECK-NEXT: DataTypeId: {{[0-9]+}}
CHECK-NEXT: TotalAccessDensity: 100
CHECK-NEXT: MinAccessDensity: 100
CHECK-NEXT: MaxAccessDensity: 100
CHECK-NEXT: TotalLifetimeAccessDensity: 100000
CHECK-NEXT: MinLifetimeAccessDensity: 100000
CHECK-NEXT: MaxLifetimeAccessDensity: 100000
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: FunctionGUID: 6699318081062747564 CHECK-NEXT: FunctionGUID: 6699318081062747564
CHECK-NEXT: AllocSites: CHECK-NEXT: AllocSites:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: Callstack: CHECK-NEXT: Callstack:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: Function: 15505678318020221912 CHECK-NEXT: Function: 15505678318020221912
CHECK-NEXT: SymbolName: qux CHECK-NEXT: SymbolName: qux
Show All 21 Lines
CHECK-NEXT: MemInfoBlock: CHECK-NEXT: MemInfoBlock:
CHECK-NEXT: AllocCount: 1 CHECK-NEXT: AllocCount: 1
CHECK-NEXT: TotalAccessCount: 1 CHECK-NEXT: TotalAccessCount: 1
CHECK-NEXT: MinAccessCount: 1 CHECK-NEXT: MinAccessCount: 1
CHECK-NEXT: MaxAccessCount: 1 CHECK-NEXT: MaxAccessCount: 1
CHECK-NEXT: TotalSize: 1 CHECK-NEXT: TotalSize: 1
CHECK-NEXT: MinSize: 1 CHECK-NEXT: MinSize: 1
CHECK-NEXT: MaxSize: 1 CHECK-NEXT: MaxSize: 1
CHECK-NEXT: AllocTimestamp: 894 CHECK-NEXT: AllocTimestamp: {{[0-9]+}}
CHECK-NEXT: DeallocTimestamp: 894 CHECK-NEXT: DeallocTimestamp: {{[0-9]+}}
CHECK-NEXT: TotalLifetime: 0 CHECK-NEXT: TotalLifetime: 0
CHECK-NEXT: MinLifetime: 0 CHECK-NEXT: MinLifetime: 0
CHECK-NEXT: MaxLifetime: 0 CHECK-NEXT: MaxLifetime: 0
CHECK-NEXT: AllocCpuId: 23 CHECK-NEXT: AllocCpuId: {{[0-9]+}}
CHECK-NEXT: DeallocCpuId: 23 CHECK-NEXT: DeallocCpuId: {{[0-9]+}}
CHECK-NEXT: NumMigratedCpu: 0 CHECK-NEXT: NumMigratedCpu: 0
CHECK-NEXT: NumLifetimeOverlaps: 0 CHECK-NEXT: NumLifetimeOverlaps: 0
CHECK-NEXT: NumSameAllocCpu: 0 CHECK-NEXT: NumSameAllocCpu: 0
CHECK-NEXT: NumSameDeallocCpu: 0 CHECK-NEXT: NumSameDeallocCpu: 0
CHECK-NEXT: DataTypeId: {{[0-9]+}} CHECK-NEXT: DataTypeId: {{[0-9]+}}
CHECK-NEXT: TotalAccessDensity: 100
CHECK-NEXT: MinAccessDensity: 100
CHECK-NEXT: MaxAccessDensity: 100
CHECK-NEXT: TotalLifetimeAccessDensity: 100000
CHECK-NEXT: MinLifetimeAccessDensity: 100000
CHECK-NEXT: MaxLifetimeAccessDensity: 100000
CHECK-NEXT: CallSites: CHECK-NEXT: CallSites:
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: - CHECK-NEXT: -
CHECK-NEXT: Function: 15505678318020221912 CHECK-NEXT: Function: 15505678318020221912
CHECK-NEXT: SymbolName: qux CHECK-NEXT: SymbolName: qux
CHECK-NEXT: LineOffset: 1 CHECK-NEXT: LineOffset: 1
CHECK-NEXT: Column: 15 CHECK-NEXT: Column: 15
CHECK-NEXT: Inline: 1 CHECK-NEXT: Inline: 1
Show All 27 Lines

llvm/test/tools/llvm-profdata/memprof-multi.test

Show All 34 Lines
``` ```
RUN: llvm-profdata show --memory %p/Inputs/multi.memprofraw --profiled-binary %p/Inputs/multi.memprofexe -o - | FileCheck %s RUN: llvm-profdata show --memory %p/Inputs/multi.memprofraw --profiled-binary %p/Inputs/multi.memprofexe -o - | FileCheck %s
We expect 2 MIB entries, 1 each for the malloc calls in the program. We expect 2 MIB entries, 1 each for the malloc calls in the program.
CHECK: MemprofProfile: CHECK: MemprofProfile:
CHECK-NEXT: Summary: CHECK-NEXT: Summary:
CHECK-NEXT: Version: 1 CHECK-NEXT: Version: 2
CHECK-NEXT: NumSegments: 9 CHECK-NEXT: NumSegments: {{[0-9]+}}
snehasishUnsubmitted
Done
ReplyInline Actions

Use the same regex for NumSegments here?

snehasish: Use the same regex for NumSegments here?
CHECK-NEXT: NumMibInfo: 2 CHECK-NEXT: NumMibInfo: 2
CHECK-NEXT: NumAllocFunctions: 1 CHECK-NEXT: NumAllocFunctions: 1
CHECK-NEXT: NumStackOffsets: 2 CHECK-NEXT: NumStackOffsets: 2
CHECK: SymbolName: main CHECK: SymbolName: main
CHECK-NEXT: LineOffset: 1 CHECK-NEXT: LineOffset: 1
CHECK-NEXT: Column: 21 CHECK-NEXT: Column: 21
CHECK: SymbolName: main CHECK: SymbolName: main
CHECK-NEXT: LineOffset: 5 CHECK-NEXT: LineOffset: 5
CHECK-NEXT: Column: 15 CHECK-NEXT: Column: 15