[PGO] Ensure non-zero entry-count after populateCounters
#112029
Conversation
|
Thank you for submitting a Pull Request (PR) to the LLVM Project! This PR will be automatically labeled and the relevant teams will be notified. If you wish to, you can add reviewers by using the "Reviewers" section on this page. If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers. If you have further questions, they may be answered by the LLVM GitHub User Guide. You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums. |
|
@llvm/pr-subscribers-llvm-transforms Author: Michael O'Farrell (mofarrell) ChangesWith sampled instrumentation (#69535), profile counts may appear corrupt and Full diff: https://github.com/llvm/llvm-project/pull/112029.diff 1 Files Affected:
diff --git a/llvm/lib/Transforms/Instrumentation/PGOInstrumentation.cpp b/llvm/lib/Transforms/Instrumentation/PGOInstrumentation.cpp
index e6e474ed376069..23d34016173165 100644
--- a/llvm/lib/Transforms/Instrumentation/PGOInstrumentation.cpp
+++ b/llvm/lib/Transforms/Instrumentation/PGOInstrumentation.cpp
@@ -1970,11 +1970,9 @@ static void fixFuncEntryCount(PGOUseFunc &Func, LoopInfo &LI,
BranchProbabilityInfo &NBPI) {
Function &F = Func.getFunc();
BlockFrequencyInfo NBFI(F, NBPI, LI);
-#ifndef NDEBUG
auto BFIEntryCount = F.getEntryCount();
- assert(BFIEntryCount && (BFIEntryCount->getCount() > 0) &&
- "Invalid BFI Entrycount");
-#endif
+ if (!BFIEntryCount || BFIEntryCount->getCount() == 0)
+ return;
auto SumCount = APFloat::getZero(APFloat::IEEEdouble());
auto SumBFICount = APFloat::getZero(APFloat::IEEEdouble());
for (auto &BBI : F) {
|
|
@llvm/pr-subscribers-pgo Author: Michael O'Farrell (mofarrell) ChangesWith sampled instrumentation (#69535), profile counts may appear corrupt and Full diff: https://github.com/llvm/llvm-project/pull/112029.diff 1 Files Affected:
diff --git a/llvm/lib/Transforms/Instrumentation/PGOInstrumentation.cpp b/llvm/lib/Transforms/Instrumentation/PGOInstrumentation.cpp
index e6e474ed376069..23d34016173165 100644
--- a/llvm/lib/Transforms/Instrumentation/PGOInstrumentation.cpp
+++ b/llvm/lib/Transforms/Instrumentation/PGOInstrumentation.cpp
@@ -1970,11 +1970,9 @@ static void fixFuncEntryCount(PGOUseFunc &Func, LoopInfo &LI,
BranchProbabilityInfo &NBPI) {
Function &F = Func.getFunc();
BlockFrequencyInfo NBFI(F, NBPI, LI);
-#ifndef NDEBUG
auto BFIEntryCount = F.getEntryCount();
- assert(BFIEntryCount && (BFIEntryCount->getCount() > 0) &&
- "Invalid BFI Entrycount");
-#endif
+ if (!BFIEntryCount || BFIEntryCount->getCount() == 0)
+ return;
auto SumCount = APFloat::getZero(APFloat::IEEEdouble());
auto SumBFICount = APFloat::getZero(APFloat::IEEEdouble());
for (auto &BBI : F) {
|
mofarrell
force-pushed
the
zero-prof-counts
branch
from
d469548 to
4d1221e
Compare
mofarrell
force-pushed
the
zero-prof-counts
branch
from
4494a6d to
4536acd
Compare
mofarrell
force-pushed
the
zero-prof-counts
branch
from
4536acd to
abfdcb6
Compare
mofarrell
changed the title
fixFuncEntryCountpopulateCounters
|
Changed the fix to guarantee the function entry count is non zero after |
With sampled instrumentation (llvm#69535), profile counts can appear corrupt. In particular a function can have a 0 block counts for all its blocks, while having some non-zero counters for select instrumentation. This is only possible for colder functions, and a reasonable modification to ensure the entry is non-zero (required by `fixFuncEntryCounts`) is to set the counter to one. This is only likely to happen for colder functions, so it is reasonable to take any action that does not crash.
mofarrell
force-pushed
the
zero-prof-counts
branch
from
abfdcb6 to
86e0a1f
Compare
| @@ -0,0 +1,28 @@ | |||
| ; RUN: llvm-profdata merge %S/Inputs/fix_entry_count_sampled.proftext -o %t.profdata | |||
There was a problem hiding this comment.
Since this input .proftext file is only used for this test, can we use split-file and inline it here?
There was a problem hiding this comment.
These testing oddities were mimicking the approach taken in other nearby tests covering the same functionality. In particular fix_entry_count.ll was used as a base.
In llvm/test/Transforms/PGOProfile/ no other test makes use of split-file for proftext data, does this make it reasonable to follow this localized convention despite the proftext only being used by one test?
There was a problem hiding this comment.
I think split-file was introduced relatively recently and fix_entry_count.ll was last updated 2 years ago. More recent proftext tests do use split-file.
I like to use split-file because otherwise we have lots of similar .proftext files in Inputs/ and it is slightly inconvenient to find their uses. It also means only one file needs to be updated if we want to change the test.
| @@ -0,0 +1,28 @@ | |||
| ; RUN: llvm-profdata merge %S/Inputs/fix_entry_count_sampled.proftext -o %t.profdata | |||
| ; RUN: opt < %s -passes=pgo-instr-use -pgo-test-profile-file=%t.profdata -S | FileCheck %s --check-prefix=USE | |||
There was a problem hiding this comment.
Since this is the only FileCheck command, can we use the default check prefix?
There was a problem hiding this comment.
Certainly. This followed the neighboring tests fix_entry_count.ll's style, which is testing similar functionality.
This crashed before making fixFuncEntryCount gracefully handle zero counts.
mofarrell
force-pushed
the
zero-prof-counts
branch
from
86e0a1f to
bcc71ab
Compare
|
@mofarrell Congratulations on having your first Pull Request (PR) merged into the LLVM Project! Your changes will be combined with recent changes from other authors, then tested by our build bots. If there is a problem with a build, you may receive a report in an email or a comment on this PR. Please check whether problems have been caused by your change specifically, as the builds can include changes from many authors. It is not uncommon for your change to be included in a build that fails due to someone else's changes, or infrastructure issues. How to do this, and the rest of the post-merge process, is covered in detail here. If your change does cause a problem, it may be reverted, or you can revert it yourself. This is a normal part of LLVM development. You can fix your changes and open a new PR to merge them again. If you don't get any reports, no action is required from you. Your changes are working as expected, well done! |
With sampled instrumentation (#69535), profile counts may appear corrupt and
fixFuncEntryCountmay assert. In particular a function can have a 0 block count for its entry, while later blocks are non zero. This is only likely to happen for colder functions, so it is reasonable to take any action that does not crash. Here we simply bail from fixing the entry count.