C++: Fix two joins in IR construction - #22503
Merged
Merged
Conversation
Before (on an internal Microsoft repo):
Evaluated relational algebra for predicate AliasedSSA::AllocationMemoryLocation.getVirtualVariable/0#dispred#8debd926@f62ba08v with tuple counts:
5606472 ~0% {2} r1 = AliasedSSA::AllocationMemoryLocation#57439a9b_10#join_rhs AND NOT `AliasAnalysis::allocationEscapes/1#93dc9772`(FIRST 1)
5534569 ~0% {2} r2 = r1 AND NOT `project#AliasedSSA::getGroupedMemoryLocation/3#14ef79fc#ffbf`(FIRST 1)
10613026097 ~0% {2} | JOIN WITH AliasedSSA::AllocationMemoryLocation#57439a9b_10#join_rhs ON FIRST 1 OUTPUT Rhs.1, Lhs.1
5534569 ~6% {2} | JOIN WITH AliasedSSA::VirtualVariable#5712df39 ON FIRST 1 OUTPUT Lhs.1, Lhs.0
759836 ~0% {2} r3 = JOIN AliasedSSA::AllocationMemoryLocation#57439a9b_10#join_rhs WITH `AliasAnalysis::allocationEscapes/1#93dc9772` ON FIRST 1 OUTPUT Lhs.0, Lhs.1
759832 ~6% {3} | JOIN WITH `AliasConfiguration::Allocation.getEnclosingIRFunction/0#dispred#3254a7ee` ON FIRST 1 OUTPUT Rhs.1, _, Lhs.1
759832 ~0% {3} | REWRITE WITH Out.1 := false
759832 ~1% {2} | JOIN WITH num#AliasedSSA::TAllAliasedMemory#4bb632db ON FIRST 2 OUTPUT Rhs.2, Lhs.2
759832 ~1% {2} | JOIN WITH AliasedSSA::VirtualVariable#5712df39 ON FIRST 1 OUTPUT Lhs.1, Lhs.0
5606472 ~0% {4} r4 = SCAN r1 OUTPUT In.0, _, _, In.1
5606472 ~0% {4} | REWRITE WITH Out.1 := false, Out.2 := false
71903 ~5% {2} | JOIN WITH `AliasedSSA::getGroupedMemoryLocation/3#14ef79fc#ffbf` ON FIRST 3 OUTPUT Rhs.3, Lhs.3
71903 ~3% {2} | JOIN WITH `AliasedSSA::GroupedMemoryLocation.getVirtualVariable/0#dispred#143f2d1b` ON FIRST 1 OUTPUT Rhs.1, Lhs.1
71903 ~2% {2} | JOIN WITH AliasedSSA::VirtualVariable#5712df39 ON FIRST 1 OUTPUT Lhs.1, Lhs.0
6366304 ~5% {2} r5 = r2 UNION r3 UNION r4
return r5
After:
[2026-09-03 14:39:34] Evaluated non-recursive predicate AliasedSSA::getAllocationMemoryLocation/1#171552fa@42e30agr in 15ms (size: 505315).
Evaluated relational algebra for predicate AliasedSSA::getAllocationMemoryLocation/1#171552fa@42e30agr with tuple counts:
505315 ~2% {2} r1 = JOIN AliasedSSA::VirtualVariable#5712df39 WITH `AliasedSSA::MemoryLocation0.getAnAllocation/0#dispred#f0047858` ON FIRST 1 OUTPUT Rhs.1, Lhs.0
return r1
[2026-09-03 14:39:39] Evaluated non-recursive predicate AliasedSSA::AllocationMemoryLocation.getVirtualVariable/0#dispred#8debd926@21855asb in 1811ms (size: 6366304).
Evaluated relational algebra for predicate AliasedSSA::AllocationMemoryLocation.getVirtualVariable/0#dispred#8debd926@21855asb with tuple counts:
5606472 ~0% {2} r1 = AliasedSSA::AllocationMemoryLocation#57439a9b_10#join_rhs AND NOT `AliasAnalysis::allocationEscapes/1#93dc9772`(FIRST 1)
5534569 ~0% {2} r2 = r1 AND NOT `project#AliasedSSA::getGroupedMemoryLocation/3#14ef79fc#ffbf`(FIRST 1)
5534569 ~1% {2} | JOIN WITH `AliasedSSA::getAllocationMemoryLocation/1#171552fa` ON FIRST 1 OUTPUT Rhs.1, Lhs.1
5534569 ~6% {2} | JOIN WITH AliasedSSA::VirtualVariable#5712df39 ON FIRST 1 OUTPUT Lhs.1, Lhs.0
759836 ~0% {2} r3 = JOIN AliasedSSA::AllocationMemoryLocation#57439a9b_10#join_rhs WITH `AliasAnalysis::allocationEscapes/1#93dc9772` ON FIRST 1 OUTPUT Lhs.0, Lhs.1
759832 ~6% {3} | JOIN WITH `AliasConfiguration::Allocation.getEnclosingIRFunction/0#dispred#3254a7ee` ON FIRST 1 OUTPUT Rhs.1, _, Lhs.1
759832 ~0% {3} | REWRITE WITH Out.1 := false
759832 ~1% {2} | JOIN WITH num#AliasedSSA::TAllAliasedMemory#4bb632db ON FIRST 2 OUTPUT Rhs.2, Lhs.2
759832 ~1% {2} | JOIN WITH AliasedSSA::VirtualVariable#5712df39 ON FIRST 1 OUTPUT Lhs.1, Lhs.0
5606472 ~0% {4} r4 = SCAN r1 OUTPUT In.0, _, _, In.1
5606472 ~0% {4} | REWRITE WITH Out.1 := false, Out.2 := false
71903 ~5% {2} | JOIN WITH `AliasedSSA::getGroupedMemoryLocation/3#14ef79fc#ffbf` ON FIRST 3 OUTPUT Rhs.3, Lhs.3
71903 ~3% {2} | JOIN WITH `AliasedSSA::GroupedMemoryLocation.getVirtualVariable/0#dispred#143f2d1b` ON FIRST 1 OUTPUT Rhs.1, Lhs.1
71903 ~2% {2} | JOIN WITH AliasedSSA::VirtualVariable#5712df39 ON FIRST 1 OUTPUT Lhs.1, Lhs.0
6366304 ~5% {2} r5 = r2 UNION r3 UNION r4
return r5
Before (on an internal Microsoft repo):
[2026-08-28 14:36:46] Evaluated non-recursive predicate TranslatedElement::TranslatedElement.getInstructionSuccessorInternal/2#dispred#a6e054ca@2248725r in 470505ms (size: 136995987).
Evaluated relational algebra for predicate TranslatedElement::TranslatedElement.getInstructionSuccessorInternal/2#dispred#a6e054ca@2248725r with tuple counts:
437 ~0% {2} r158 = JOIN EdgeKind::EdgeKindImpl#8ed21aeb WITH num#InstructionTag::CallTargetTag#8c4ab419 CARTESIAN PRODUCT OUTPUT Lhs.0, Rhs.0
437 ~0% {3} | JOIN WITH num#InstructionTag::CallTag#a77d4021 CARTESIAN PRODUCT OUTPUT Rhs.0, Lhs.0, Lhs.1
1070556919 ~0% {4} | JOIN WITH `TranslatedElement::TranslatedElement.getInstruction/1#dispred#f6df9482_102#join_rhs` ON FIRST 1 OUTPUT Rhs.1, Lhs.1, Lhs.2, Rhs.2
316388 ~1% {4} | JOIN WITH TranslatedInitialization::TranslatedDefaultFieldInitialization#d9c761ed ON FIRST 1 OUTPUT Lhs.0, Lhs.2, Lhs.1, Lhs.3
After:
[2026-09-03 13:56:28] Evaluated non-recursive predicate TranslatedInitialization::getCallInstruction/1#c88d849b@3e960544 in 40ms (size: 724).
Evaluated relational algebra for predicate TranslatedInitialization::getCallInstruction/1#c88d849b@3e960544 with tuple counts:
724 ~3% {2} r1 = JOIN `_TranslatedElement::TranslatedElement.getInstruction/1#dispred#f6df9482_102#join_rhs_num#Instruction__#shared#1` WITH TranslatedInitialization::TranslatedDefaultFieldInitialization#d9c761ed ON FIRST 1 OUTPUT Lhs.0, Lhs.1
return r1
437 ~0% {2} r5 = JOIN EdgeKind::EdgeKindImpl#8ed21aeb WITH num#InstructionTag::CallTargetTag#8c4ab419 CARTESIAN PRODUCT OUTPUT Lhs.0, Rhs.0
316388 ~1% {4} | JOIN WITH `TranslatedInitialization::getCallInstruction/1#c88d849b` CARTESIAN PRODUCT OUTPUT Rhs.0, Lhs.1, Lhs.0, Rhs.1
Contributor
There was a problem hiding this comment.
Copilot review overview
🟢 Approval recommended
The changes are narrowly scoped and preserve existing query semantics.
Review tier: Balanced
Findings: None
What changed in this PR
Optimizes two expensive joins in C++ IR construction using pragma[nomagic] helper predicates.
Changes:
- Isolates default-field call-instruction lookup.
- Isolates allocation-to-virtual-variable lookup.
| File | Description |
|---|---|
TranslatedInitialization.qll |
Optimizes call-instruction lookup. |
AliasedSSA.qll |
Optimizes allocation memory-location lookup. |
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Contributor
Author
|
Only a 1% performance improvement on the DCA projects, but what a beautiful reduction in repored bad join orders! 😍 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See the individual commits for before/after tuple counts.