[llvm] r269131 - Reapply r266477 and r266488
Sean Silva via llvm-commits
llvm-commits at lists.llvm.org
Tue May 10 23:40:43 PDT 2016
On Tue, May 10, 2016 at 11:30 PM, Xinliang David Li <xinliangli at gmail.com>
wrote:
>
>
> On Tue, May 10, 2016 at 10:45 PM, Sean Silva <chisophugis at gmail.com>
> wrote:
>
>>
>>
>> On Tue, May 10, 2016 at 9:55 PM, Xinliang David Li <xinliangli at gmail.com>
>> wrote:
>>
>>>
>>>
>>> On Tue, May 10, 2016 at 5:19 PM, Sean Silva via llvm-commits <
>>> llvm-commits at lists.llvm.org> wrote:
>>>
>>>>
>>>>
>>>> On Tue, May 10, 2016 at 4:41 PM, Easwaran Raman <eraman at google.com>
>>>> wrote:
>>>>
>>>>> (+cc: vsk)
>>>>> Hi Sean,
>>>>> I have reverted the patch. First, could you clarify the "core
>>>>> libraries" part? Is it libCore alone or a larger set?
>>>>>
>>>>
>>>> Sorry, I used that term for the lack of a better description. More or
>>>> less I mean anything in lib/ of any LLVM subproject.
>>>>
>>>>
>>>>> I presume it is a larger set since this is part of libProfileData, but
>>>>> still wanted to check with you. I can think of a couple of ways to avoid
>>>>> the global: a) do not cache this at the ProfileSummary level and instead
>>>>> cache it at the client, b) move ProfileSummary to Support (or IR). What do
>>>>> you suggest is the best way forward?
>>>>>
>>>>
>>>> I have not done an in-depth analysis of the situation, but I think this
>>>> can be solved by taking the "compute from metadata" design to its logical
>>>> conclusion and splitting the solution into two parts:
>>>> - there is a part that lives in the IR library that is computed purely
>>>> based on the metadata and has no dependence on ProfileData
>>>> - there is a part that lives in the the ProfileData library that does
>>>> InstrProf/SampleProf specific annotation of the metadata.
>>>>
>>>
>>> I like the this suggestion of splitting.
>>>
>>>
>>>>
>>>> The annotating part needs the IR library, so it makes sense for some of
>>>> the shared routines to be in the IR library (such as the list of cutoffs).
>>>>
>>>
>>> I think it is better to put part 1 into Support lib -- both
>>> BranchProbability and BlockFrequency representation lives there today.
>>>
>>
>> So in the end we would have a data structure ProfileSummary in libSupport
>> and then a "ProfileSummaryInfo" analysis to complement it? (analogous to
>> BPI vs. BranchProbability and BFI vs. BlockFrequency).
>>
>
>
> Roughly like that but there is no need for a separate ProfileSummaryInfo
> analysis pass. This is because profile summary data is already computed by
> the profile-use/annotation pass and set as meta data into the module. The
> whole reason why Easwaran uses a static caching var is to make the access
> slightly faster (reading from meta data is slow).
>
>
> The current ProfileSummary has a `bool isFunctionHot(Function *F)` method
>> which seems like it would need a "ProfileSummaryInfo" analysis (or other
>> solution) so that it can reference `Function` without a layering violation.
>>
>
>
> It should only needs very low level interfaces such as
>
> 1) given a profile count, tell me if it is hot (with default or specfified
> cutoff) or cold ..
> 2) given a hotness cutoff, give me the profile count threshold
>
> etc.
>
>
>>
>> But I'm not sure about the exact details of the final interface you want
>> optimizations to have access to. Can you explain the final interface that
>> you want optimization passes to have access to?
>>
>
> Given the above low level interfaces, clients can define higher level
> interfaces that takes function, BB or callsite to query for the
> hotness/coldness info.
>
This sounds like the type of thing that a "ProfileSummaryInfo" pass would
be good for. Assuming that the clients are N>1 it would be redundant to
have multiple cases. Also, the analysis can do the caching.
-- Sean Silva
>
> David
>
>
>> -- Sean Silva
>>
>>
>>>
>>> David
>>>
>>>
>>>> -- Sean Silva
>>>>
>>>>
>>>>>
>>>>> Thanks,
>>>>> Easwaran
>>>>>
>>>>>
>>>>> On Tue, May 10, 2016 at 4:17 PM, Sean Silva <chisophugis at gmail.com>
>>>>> wrote:
>>>>>
>>>>>>
>>>>>>
>>>>>> On Tue, May 10, 2016 at 3:03 PM, Easwaran Raman via llvm-commits <
>>>>>> llvm-commits at lists.llvm.org> wrote:
>>>>>>
>>>>>>> Author: eraman
>>>>>>> Date: Tue May 10 17:03:23 2016
>>>>>>> New Revision: 269131
>>>>>>>
>>>>>>> URL: http://llvm.org/viewvc/llvm-project?rev=269131&view=rev
>>>>>>> Log:
>>>>>>> Reapply r266477 and r266488
>>>>>>>
>>>>>>> Added:
>>>>>>> llvm/trunk/unittests/ProfileData/ProfileSummaryTest.cpp
>>>>>>> - copied unchanged from r266618,
>>>>>>> llvm/trunk/unittests/ProfileData/ProfileSummaryTest.cpp
>>>>>>> Modified:
>>>>>>> llvm/trunk/include/llvm/ProfileData/ProfileCommon.h
>>>>>>> llvm/trunk/lib/Analysis/InlineCost.cpp
>>>>>>> llvm/trunk/lib/Analysis/LLVMBuild.txt
>>>>>>> llvm/trunk/lib/ProfileData/ProfileSummary.cpp
>>>>>>> llvm/trunk/test/Transforms/Inline/inline-cold-callee.ll
>>>>>>> llvm/trunk/test/Transforms/Inline/inline-hot-callee.ll
>>>>>>> llvm/trunk/unittests/ProfileData/CMakeLists.txt
>>>>>>>
>>>>>>> Modified: llvm/trunk/include/llvm/ProfileData/ProfileCommon.h
>>>>>>> URL:
>>>>>>> http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/ProfileData/ProfileCommon.h?rev=269131&r1=269130&r2=269131&view=diff
>>>>>>>
>>>>>>> ==============================================================================
>>>>>>> --- llvm/trunk/include/llvm/ProfileData/ProfileCommon.h (original)
>>>>>>> +++ llvm/trunk/include/llvm/ProfileData/ProfileCommon.h Tue May 10
>>>>>>> 17:03:23 2016
>>>>>>> @@ -21,6 +21,8 @@
>>>>>>> #include <vector>
>>>>>>>
>>>>>>> #include "llvm/Support/Casting.h"
>>>>>>> +#include "llvm/Support/ManagedStatic.h"
>>>>>>> +#include "llvm/Support/Mutex.h"
>>>>>>>
>>>>>>> namespace llvm {
>>>>>>> class Function;
>>>>>>> @@ -35,6 +37,7 @@ class LLVMContext;
>>>>>>> class Metadata;
>>>>>>> class MDTuple;
>>>>>>> class MDNode;
>>>>>>> +class Module;
>>>>>>>
>>>>>>> inline const char *getHotSectionPrefix() { return ".hot"; }
>>>>>>> inline const char *getUnlikelySectionPrefix() { return ".unlikely";
>>>>>>> }
>>>>>>> @@ -67,6 +70,14 @@ private:
>>>>>>> // appears in the profile. The map is kept sorted in the
>>>>>>> descending order of
>>>>>>> // counts.
>>>>>>> std::map<uint64_t, uint32_t, std::greater<uint64_t>>
>>>>>>> CountFrequencies;
>>>>>>> + // Compute profile summary for a module.
>>>>>>> + static ProfileSummary *computeProfileSummary(Module *M);
>>>>>>> + // Cache of last seen module and its profile summary.
>>>>>>> + static ManagedStatic<std::pair<Module *,
>>>>>>> std::unique_ptr<ProfileSummary>>>
>>>>>>> + CachedSummary;
>>>>>>> + // Mutex to access summary cache
>>>>>>> + static ManagedStatic<sys::SmartMutex<true>> CacheMutex;
>>>>>>>
>>>>>>
>>>>>> Hi Easwaran,
>>>>>>
>>>>>> In the LLVM core libraries, we don't use globals like this.
>>>>>>
>>>>>> Can you please revert and redesign this to avoid the global? For
>>>>>> example, keeping it in the Module itself? I understand that there are some
>>>>>> layering issues here, and I'm glad to help with finding a nice solution,
>>>>>> but putting a global in the LLVM core libraries like this is not acceptable.
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>> -- Sean Silva
>>>>>>
>>>>>>
>>>>>>> +
>>>>>>> protected:
>>>>>>> SummaryEntryVector DetailedSummary;
>>>>>>> std::vector<uint32_t> DetailedSummaryCutoffs;
>>>>>>> @@ -85,14 +96,12 @@ protected:
>>>>>>> : PSK(K), DetailedSummary(DetailedSummary),
>>>>>>> TotalCount(TotalCount),
>>>>>>> MaxCount(MaxCount), MaxFunctionCount(MaxFunctionCount),
>>>>>>> NumCounts(NumCounts), NumFunctions(NumFunctions) {}
>>>>>>> - ~ProfileSummary() = default;
>>>>>>> inline void addCount(uint64_t Count);
>>>>>>> /// \brief Return metadata specific to the profile format.
>>>>>>> /// Derived classes implement this method to return a vector of
>>>>>>> Metadata.
>>>>>>> virtual std::vector<Metadata *> getFormatSpecificMD(LLVMContext
>>>>>>> &Context) = 0;
>>>>>>> /// \brief Return detailed summary as metadata.
>>>>>>> Metadata *getDetailedSummaryMD(LLVMContext &Context);
>>>>>>> -
>>>>>>> public:
>>>>>>> static const int Scale = 1000000;
>>>>>>> Kind getKind() const { return PSK; }
>>>>>>> @@ -111,6 +120,10 @@ public:
>>>>>>> static ProfileSummary *getFromMD(Metadata *MD);
>>>>>>> uint32_t getNumFunctions() { return NumFunctions; }
>>>>>>> uint64_t getMaxFunctionCount() { return MaxFunctionCount; }
>>>>>>> + /// \brief Get profile summary associated with module \p M
>>>>>>> + static inline ProfileSummary *getProfileSummary(Module *M);
>>>>>>> + virtual ~ProfileSummary() = default;
>>>>>>> + virtual bool operator==(ProfileSummary &Other);
>>>>>>> };
>>>>>>>
>>>>>>> class InstrProfSummary final : public ProfileSummary {
>>>>>>> @@ -140,6 +153,7 @@ public:
>>>>>>> uint64_t getTotalCount() { return TotalCount; }
>>>>>>> uint64_t getMaxBlockCount() { return MaxCount; }
>>>>>>> uint64_t getMaxInternalBlockCount() { return
>>>>>>> MaxInternalBlockCount; }
>>>>>>> + bool operator==(ProfileSummary &Other) override;
>>>>>>> };
>>>>>>>
>>>>>>> class SampleProfileSummary final : public ProfileSummary {
>>>>>>> @@ -180,5 +194,24 @@ SummaryEntryVector &ProfileSummary::getD
>>>>>>> return DetailedSummary;
>>>>>>> }
>>>>>>>
>>>>>>> +ProfileSummary *ProfileSummary::getProfileSummary(Module *M) {
>>>>>>> + if (!M)
>>>>>>> + return nullptr;
>>>>>>> + sys::SmartScopedLock<true> Lock(*CacheMutex);
>>>>>>> + // Computing profile summary for a module involves parsing a
>>>>>>> fairly large
>>>>>>> + // metadata and could be expensive. We use a simple cache of the
>>>>>>> last seen
>>>>>>> + // module and its profile summary.
>>>>>>> + if (CachedSummary->first != M) {
>>>>>>> + auto *Summary = computeProfileSummary(M);
>>>>>>> + // Do not cache if the summary is empty. This is because a
>>>>>>> later pass
>>>>>>> + // (sample profile loader, for example) could attach the
>>>>>>> summary metadata on
>>>>>>> + // the module.
>>>>>>> + if (!Summary)
>>>>>>> + return nullptr;
>>>>>>> + CachedSummary->first = M;
>>>>>>> + CachedSummary->second.reset(Summary);
>>>>>>> + }
>>>>>>> + return CachedSummary->second.get();
>>>>>>> +}
>>>>>>> } // end namespace llvm
>>>>>>> #endif
>>>>>>>
>>>>>>> Modified: llvm/trunk/lib/Analysis/InlineCost.cpp
>>>>>>> URL:
>>>>>>> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Analysis/InlineCost.cpp?rev=269131&r1=269130&r2=269131&view=diff
>>>>>>>
>>>>>>> ==============================================================================
>>>>>>> --- llvm/trunk/lib/Analysis/InlineCost.cpp (original)
>>>>>>> +++ llvm/trunk/lib/Analysis/InlineCost.cpp Tue May 10 17:03:23 2016
>>>>>>> @@ -30,6 +30,7 @@
>>>>>>> #include "llvm/IR/InstVisitor.h"
>>>>>>> #include "llvm/IR/IntrinsicInst.h"
>>>>>>> #include "llvm/IR/Operator.h"
>>>>>>> +#include "llvm/ProfileData/ProfileCommon.h"
>>>>>>> #include "llvm/Support/Debug.h"
>>>>>>> #include "llvm/Support/raw_ostream.h"
>>>>>>>
>>>>>>> @@ -630,10 +631,11 @@ void CallAnalyzer::updateThreshold(CallS
>>>>>>> // a well-tuned heuristic based on *callsite* hotness and not
>>>>>>> callee hotness.
>>>>>>> uint64_t FunctionCount = 0, MaxFunctionCount = 0;
>>>>>>> bool HasPGOCounts = false;
>>>>>>> - if (Callee.getEntryCount() &&
>>>>>>> Callee.getParent()->getMaximumFunctionCount()) {
>>>>>>> + ProfileSummary *PS =
>>>>>>> ProfileSummary::getProfileSummary(Callee.getParent());
>>>>>>> + if (Callee.getEntryCount() && PS) {
>>>>>>> HasPGOCounts = true;
>>>>>>> FunctionCount = Callee.getEntryCount().getValue();
>>>>>>> - MaxFunctionCount =
>>>>>>> Callee.getParent()->getMaximumFunctionCount().getValue();
>>>>>>> + MaxFunctionCount = PS->getMaxFunctionCount();
>>>>>>> }
>>>>>>>
>>>>>>> // Listen to the inlinehint attribute or profile based hotness
>>>>>>> information
>>>>>>>
>>>>>>> Modified: llvm/trunk/lib/Analysis/LLVMBuild.txt
>>>>>>> URL:
>>>>>>> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Analysis/LLVMBuild.txt?rev=269131&r1=269130&r2=269131&view=diff
>>>>>>>
>>>>>>> ==============================================================================
>>>>>>> --- llvm/trunk/lib/Analysis/LLVMBuild.txt (original)
>>>>>>> +++ llvm/trunk/lib/Analysis/LLVMBuild.txt Tue May 10 17:03:23 2016
>>>>>>> @@ -19,4 +19,4 @@
>>>>>>> type = Library
>>>>>>> name = Analysis
>>>>>>> parent = Libraries
>>>>>>> -required_libraries = Core Support
>>>>>>> +required_libraries = Core ProfileData Support
>>>>>>>
>>>>>>> Modified: llvm/trunk/lib/ProfileData/ProfileSummary.cpp
>>>>>>> URL:
>>>>>>> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/ProfileData/ProfileSummary.cpp?rev=269131&r1=269130&r2=269131&view=diff
>>>>>>>
>>>>>>> ==============================================================================
>>>>>>> --- llvm/trunk/lib/ProfileData/ProfileSummary.cpp (original)
>>>>>>> +++ llvm/trunk/lib/ProfileData/ProfileSummary.cpp Tue May 10
>>>>>>> 17:03:23 2016
>>>>>>> @@ -15,6 +15,7 @@
>>>>>>> #include "llvm/IR/Constants.h"
>>>>>>> #include "llvm/IR/Function.h"
>>>>>>> #include "llvm/IR/Metadata.h"
>>>>>>> +#include "llvm/IR/Module.h"
>>>>>>> #include "llvm/IR/Type.h"
>>>>>>> #include "llvm/ProfileData/InstrProf.h"
>>>>>>> #include "llvm/ProfileData/ProfileCommon.h"
>>>>>>> @@ -32,6 +33,10 @@ const std::vector<uint32_t> ProfileSumma
>>>>>>> 900000, 950000, 990000, 999000, 999900, 999990, 999999});
>>>>>>> const char *ProfileSummary::KindStr[2] = {"InstrProf",
>>>>>>> "SampleProfile"};
>>>>>>>
>>>>>>> +ManagedStatic<std::pair<Module *, std::unique_ptr<ProfileSummary>>>
>>>>>>> + ProfileSummary::CachedSummary;
>>>>>>> +ManagedStatic<sys::SmartMutex<true>> ProfileSummary::CacheMutex;
>>>>>>> +
>>>>>>> void InstrProfSummary::addRecord(const InstrProfRecord &R) {
>>>>>>> // The first counter is not necessarily an entry count for IR
>>>>>>> // instrumentation profiles.
>>>>>>> @@ -86,6 +91,39 @@ void ProfileSummary::computeDetailedSumm
>>>>>>> }
>>>>>>> }
>>>>>>>
>>>>>>> +bool ProfileSummary::operator==(ProfileSummary &Other) {
>>>>>>> + if (getKind() != Other.getKind())
>>>>>>> + return false;
>>>>>>> + if (TotalCount != Other.TotalCount)
>>>>>>> + return false;
>>>>>>> + if (MaxCount != Other.MaxCount)
>>>>>>> + return false;
>>>>>>> + if (MaxFunctionCount != Other.MaxFunctionCount)
>>>>>>> + return false;
>>>>>>> + if (NumFunctions != Other.NumFunctions)
>>>>>>> + return false;
>>>>>>> + if (NumCounts != Other.NumCounts)
>>>>>>> + return false;
>>>>>>> + std::vector<ProfileSummaryEntry> DS1 = getDetailedSummary();
>>>>>>> + std::vector<ProfileSummaryEntry> DS2 = Other.getDetailedSummary();
>>>>>>> + auto CompareSummaryEntry = [](ProfileSummaryEntry &E1,
>>>>>>> + ProfileSummaryEntry &E2) {
>>>>>>> + return E1.Cutoff == E2.Cutoff && E1.MinCount == E2.MinCount &&
>>>>>>> + E1.NumCounts == E2.NumCounts;
>>>>>>> + };
>>>>>>> + if (!std::equal(DS1.begin(), DS1.end(), DS2.begin(),
>>>>>>> CompareSummaryEntry))
>>>>>>> + return false;
>>>>>>> + return true;
>>>>>>> +}
>>>>>>> +
>>>>>>> +bool InstrProfSummary::operator==(ProfileSummary &Other) {
>>>>>>> + InstrProfSummary *OtherIPS = dyn_cast<InstrProfSummary>(&Other);
>>>>>>> + if (!OtherIPS)
>>>>>>> + return false;
>>>>>>> + return MaxInternalBlockCount == OtherIPS->MaxInternalBlockCount &&
>>>>>>> + ProfileSummary::operator==(Other);
>>>>>>> +}
>>>>>>> +
>>>>>>> // Returns true if the function is a hot function.
>>>>>>> bool ProfileSummary::isFunctionHot(const Function *F) {
>>>>>>> // FIXME: update when summary data is stored in module's metadata.
>>>>>>> @@ -366,3 +404,9 @@ ProfileSummary *ProfileSummary::getFromM
>>>>>>> else
>>>>>>> return nullptr;
>>>>>>> }
>>>>>>> +
>>>>>>> +ProfileSummary *ProfileSummary::computeProfileSummary(Module *M) {
>>>>>>> + if (Metadata *MD = M->getProfileSummary())
>>>>>>> + return getFromMD(MD);
>>>>>>> + return nullptr;
>>>>>>> +}
>>>>>>>
>>>>>>> Modified: llvm/trunk/test/Transforms/Inline/inline-cold-callee.ll
>>>>>>> URL:
>>>>>>> http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/Inline/inline-cold-callee.ll?rev=269131&r1=269130&r2=269131&view=diff
>>>>>>>
>>>>>>> ==============================================================================
>>>>>>> --- llvm/trunk/test/Transforms/Inline/inline-cold-callee.ll
>>>>>>> (original)
>>>>>>> +++ llvm/trunk/test/Transforms/Inline/inline-cold-callee.ll Tue May
>>>>>>> 10 17:03:23 2016
>>>>>>> @@ -5,7 +5,7 @@
>>>>>>> ; A callee with identical body does gets inlined because cost fits
>>>>>>> within the
>>>>>>> ; inline-threshold
>>>>>>>
>>>>>>> -define i32 @callee1(i32 %x) !prof !1 {
>>>>>>> +define i32 @callee1(i32 %x) !prof !21 {
>>>>>>> %x1 = add i32 %x, 1
>>>>>>> %x2 = add i32 %x1, 1
>>>>>>> %x3 = add i32 %x2, 1
>>>>>>> @@ -13,7 +13,7 @@ define i32 @callee1(i32 %x) !prof !1 {
>>>>>>> ret i32 %x3
>>>>>>> }
>>>>>>>
>>>>>>> -define i32 @callee2(i32 %x) !prof !2 {
>>>>>>> +define i32 @callee2(i32 %x) !prof !22 {
>>>>>>> ; CHECK-LABEL: @callee2(
>>>>>>> %x1 = add i32 %x, 1
>>>>>>> %x2 = add i32 %x1, 1
>>>>>>> @@ -22,7 +22,7 @@ define i32 @callee2(i32 %x) !prof !2 {
>>>>>>> ret i32 %x3
>>>>>>> }
>>>>>>>
>>>>>>> -define i32 @caller2(i32 %y1) !prof !2 {
>>>>>>> +define i32 @caller2(i32 %y1) !prof !22 {
>>>>>>> ; CHECK-LABEL: @caller2(
>>>>>>> ; CHECK: call i32 @callee2
>>>>>>> ; CHECK-NOT: call i32 @callee1
>>>>>>> @@ -32,8 +32,19 @@ define i32 @caller2(i32 %y1) !prof !2 {
>>>>>>> ret i32 %y3
>>>>>>> }
>>>>>>>
>>>>>>> -!llvm.module.flags = !{!0}
>>>>>>> -!0 = !{i32 1, !"MaxFunctionCount", i32 1000}
>>>>>>> -!1 = !{!"function_entry_count", i64 100}
>>>>>>> -!2 = !{!"function_entry_count", i64 1}
>>>>>>> -
>>>>>>> +!llvm.module.flags = !{!1}
>>>>>>> +!21 = !{!"function_entry_count", i64 100}
>>>>>>> +!22 = !{!"function_entry_count", i64 1}
>>>>>>> +
>>>>>>> +!1 = !{i32 1, !"ProfileSummary", !2}
>>>>>>> +!2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
>>>>>>> +!3 = !{!"ProfileFormat", !"InstrProf"}
>>>>>>> +!4 = !{!"TotalCount", i64 10000}
>>>>>>> +!5 = !{!"MaxBlockCount", i64 1000}
>>>>>>> +!6 = !{!"MaxInternalBlockCount", i64 1}
>>>>>>> +!7 = !{!"MaxFunctionCount", i64 1000}
>>>>>>> +!8 = !{!"NumBlocks", i64 3}
>>>>>>> +!9 = !{!"NumFunctions", i64 3}
>>>>>>> +!10 = !{!"DetailedSummary", !11}
>>>>>>> +!11 = !{!12}
>>>>>>> +!12 = !{i32 10000, i64 0, i32 0}
>>>>>>>
>>>>>>> Modified: llvm/trunk/test/Transforms/Inline/inline-hot-callee.ll
>>>>>>> URL:
>>>>>>> http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/Inline/inline-hot-callee.ll?rev=269131&r1=269130&r2=269131&view=diff
>>>>>>>
>>>>>>> ==============================================================================
>>>>>>> --- llvm/trunk/test/Transforms/Inline/inline-hot-callee.ll (original)
>>>>>>> +++ llvm/trunk/test/Transforms/Inline/inline-hot-callee.ll Tue May
>>>>>>> 10 17:03:23 2016
>>>>>>> @@ -5,7 +5,7 @@
>>>>>>> ; A cold callee with identical body does not get inlined because
>>>>>>> cost exceeds the
>>>>>>> ; inline-threshold
>>>>>>>
>>>>>>> -define i32 @callee1(i32 %x) !prof !1 {
>>>>>>> +define i32 @callee1(i32 %x) !prof !20 {
>>>>>>> %x1 = add i32 %x, 1
>>>>>>> %x2 = add i32 %x1, 1
>>>>>>> %x3 = add i32 %x2, 1
>>>>>>> @@ -13,7 +13,7 @@ define i32 @callee1(i32 %x) !prof !1 {
>>>>>>> ret i32 %x3
>>>>>>> }
>>>>>>>
>>>>>>> -define i32 @callee2(i32 %x) !prof !2 {
>>>>>>> +define i32 @callee2(i32 %x) !prof !21 {
>>>>>>> ; CHECK-LABEL: @callee2(
>>>>>>> %x1 = add i32 %x, 1
>>>>>>> %x2 = add i32 %x1, 1
>>>>>>> @@ -22,7 +22,7 @@ define i32 @callee2(i32 %x) !prof !2 {
>>>>>>> ret i32 %x3
>>>>>>> }
>>>>>>>
>>>>>>> -define i32 @caller2(i32 %y1) !prof !2 {
>>>>>>> +define i32 @caller2(i32 %y1) !prof !21 {
>>>>>>> ; CHECK-LABEL: @caller2(
>>>>>>> ; CHECK: call i32 @callee2
>>>>>>> ; CHECK-NOT: call i32 @callee1
>>>>>>> @@ -32,8 +32,19 @@ define i32 @caller2(i32 %y1) !prof !2 {
>>>>>>> ret i32 %y3
>>>>>>> }
>>>>>>>
>>>>>>> -!llvm.module.flags = !{!0}
>>>>>>> -!0 = !{i32 1, !"MaxFunctionCount", i32 10}
>>>>>>> -!1 = !{!"function_entry_count", i64 10}
>>>>>>> -!2 = !{!"function_entry_count", i64 1}
>>>>>>> -
>>>>>>> +!llvm.module.flags = !{!1}
>>>>>>> +!20 = !{!"function_entry_count", i64 10}
>>>>>>> +!21 = !{!"function_entry_count", i64 1}
>>>>>>> +
>>>>>>> +!1 = !{i32 1, !"ProfileSummary", !2}
>>>>>>> +!2 = !{!3, !4, !5, !6, !7, !8, !9, !10}
>>>>>>> +!3 = !{!"ProfileFormat", !"InstrProf"}
>>>>>>> +!4 = !{!"TotalCount", i64 10000}
>>>>>>> +!5 = !{!"MaxBlockCount", i64 10}
>>>>>>> +!6 = !{!"MaxInternalBlockCount", i64 1}
>>>>>>> +!7 = !{!"MaxFunctionCount", i64 10}
>>>>>>> +!8 = !{!"NumBlocks", i64 3}
>>>>>>> +!9 = !{!"NumFunctions", i64 3}
>>>>>>> +!10 = !{!"DetailedSummary", !11}
>>>>>>> +!11 = !{!12}
>>>>>>> +!12 = !{i32 10000, i64 0, i32 0}
>>>>>>>
>>>>>>> Modified: llvm/trunk/unittests/ProfileData/CMakeLists.txt
>>>>>>> URL:
>>>>>>> http://llvm.org/viewvc/llvm-project/llvm/trunk/unittests/ProfileData/CMakeLists.txt?rev=269131&r1=269130&r2=269131&view=diff
>>>>>>>
>>>>>>> ==============================================================================
>>>>>>> --- llvm/trunk/unittests/ProfileData/CMakeLists.txt (original)
>>>>>>> +++ llvm/trunk/unittests/ProfileData/CMakeLists.txt Tue May 10
>>>>>>> 17:03:23 2016
>>>>>>> @@ -8,5 +8,6 @@ set(LLVM_LINK_COMPONENTS
>>>>>>> add_llvm_unittest(ProfileDataTests
>>>>>>> CoverageMappingTest.cpp
>>>>>>> InstrProfTest.cpp
>>>>>>> + ProfileSummaryTest.cpp
>>>>>>> SampleProfTest.cpp
>>>>>>> )
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> llvm-commits mailing list
>>>>>>> llvm-commits at lists.llvm.org
>>>>>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> llvm-commits mailing list
>>>> llvm-commits at lists.llvm.org
>>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>>>>
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160510/69c6d9f0/attachment.html>
More information about the llvm-commits
mailing list