perf: store per-module constant prefix trees in oleans#14362
Draft
Kha wants to merge 15 commits into
Draft
Conversation
Member
Author
|
!bench |
|
Benchmark results for 7a68001 against 8006bb0 are in. There are significant results. @Kha Warning These warnings may indicate that the benchmark results are not directly comparable, for example due to changes in the runner configuration or hardware.
Large changes (55✅, 329🟥) Too many entries to display here. View the full report on radar instead. Medium changes (7✅, 1528🟥) Too many entries to display here. View the full report on radar instead. Small changes (5✅, 816🟥) Too many entries to display here. View the full report on radar instead. |
github-actions
Bot
added
toolchain-available
Collaborator
|
Reference manual CI status:
|
leanprover-bot
added
the
breaks-manual
Member
Author
|
!bench mathlib |
|
Benchmark results for leanprover-community/mathlib4-nightly-testing@48fbe78 against leanprover-community/mathlib4-nightly-testing@5edfd4c are in. There are significant results. @Kha
No significant changes detected. |
Member
Author
|
!bench |
|
Benchmark results for 736e8af against 8006bb0 are in. There are significant results. @Kha Warning These warnings may indicate that the benchmark results are not directly comparable, for example due to changes in the runner configuration or hardware.
Large changes (56✅, 232🟥) Too many entries to display here. View the full report on radar instead. Medium changes (5✅, 1101🟥) Too many entries to display here. View the full report on radar instead. Small changes (6✅, 1324🟥) Too many entries to display here. View the full report on radar instead. |
This PR replaces the eager construction of the imported-constants hash maps in `finalizeImport` (`privateConstantMap`, `publicConstantMap`, `const2ModIdx`; ~530k insertions for `import Lean`) with prefix trees over each module's constant names that are precomputed at olean write time and remain in the memory-mapped compacted regions. Import now only merges the per-module trees, allocating fresh nodes solely for name prefixes shared between modules (~3k nodes / ~81k edges for the stdlib), and lookups walk the merged view, disambiguating children by cached per-prefix `Name` hash. `import Lean` drops from ~527ms to ~393ms; olean sizes grow by ~6%. Details: * New `Lean.ConstTrie`: generic per-module tree (`ConstTrie α`), merged import view (`ImportedConsts α`) with single-module subtrees borrowed from the regions and module-index tagging, and the new `ConstMap` structure (`imported` view + `locals` map). * `ModuleData` gains `constTrie`/`extraConstTrie` fields, filled centrally in `saveModuleData(Parts)`. The extra-const tree replaces the 325k codegen-name entries of `const2ModIdx` (`Kernel.Environment` now holds `allImportedConsts`/`importedExtraConsts` instead, serving `getModuleIdxFor?`). * Duplicate imported constants are resolved (incl. `subsumesInfo` realizability subsumption and `throwAlreadyImported`) from the merge's reported collisions. * In `saveModuleDataParts`, all parts' data must stay alive across the chained compactor saves; the tries are therefore attached to all parts up front (the compactor tracks compacted objects by heap address, so per-part temporaries would corrupt later parts via address reuse). * `stage0/src/stdlib_flags.h` temporarily enables `interpreter.prefer_native` as this is an ABI-breaking change to `Kernel.Environment`/`ConstMap` (to be reverted by `update-stage0`). Stage 1 tests are meaningless on this commit since the stage 1 oleans lack the new `ModuleData` fields; validate against stage 2. Co-Authored-By: Claude
This PR speeds up constant lookups in the prefix-tree-backed environment by storing each trie node's child key hashes in a `ByteArray` alongside the children, so the per-level binary search runs over contiguous memory instead of chasing child and key pointers, in the memory-mapped per-module trees as well as the merged import spine. This supersedes the `indexed` hash-map wrappers on wide merged nodes, which could not cover wide nodes inside single-module region subtrees and made `Runtime.markPersistent` walk their entries on every import. Co-Authored-By: Claude
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/batteries
that referenced
this pull request
Jul 11, 2026
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/mathlib4-nightly-testing
that referenced
this pull request
Jul 11, 2026
leanprover-bot
added a commit
to leanprover/reference-manual
that referenced
this pull request
Jul 11, 2026
leanprover-bot
added
builds-manual
Member
Author
|
!bench |
|
Benchmark results for 2abb3e2 against 8006bb0 are in. There are significant results. @Kha Warning These warnings may indicate that the benchmark results are not directly comparable, for example due to changes in the runner configuration or hardware.
Large changes (49✅, 433🟥) Too many entries to display here. View the full report on radar instead. Medium changes (9✅, 1889🟥) Too many entries to display here. View the full report on radar instead. Small changes (4✅, 369🟥) Too many entries to display here. View the full report on radar instead. |
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/batteries
that referenced
this pull request
Jul 11, 2026
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/mathlib4-nightly-testing
that referenced
this pull request
Jul 11, 2026
leanprover-bot
added a commit
to leanprover/reference-manual
that referenced
this pull request
Jul 11, 2026
This PR restores the intended final form of the trie child index, which the previous commit accidentally carried in an interim form due to a workspace snapshot race: instead of a binary search that composes each probed hash from eight byte loads, the child index stores `[k][2^k + 1 u16 offsets][one fingerprint byte per child]`; the query hash's top `k` bits select a window of on average at most one child, a fingerprint byte filters it, and the final name component comparison (sound on its own, as siblings always differ in their final component) confirms the match. Co-Authored-By: Claude
Member
Author
|
!bench |
|
Benchmark results for e957cab against 8006bb0 are in. There are significant results. @Kha Warning These warnings may indicate that the benchmark results are not directly comparable, for example due to changes in the runner configuration or hardware.
Large changes (52✅, 126🟥) Too many entries to display here. View the full report on radar instead. Medium changes (7✅, 560🟥) Too many entries to display here. View the full report on radar instead. Small changes (6✅, 1925🟥) Too many entries to display here. View the full report on radar instead. |
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/batteries
that referenced
this pull request
Jul 11, 2026
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/mathlib4-nightly-testing
that referenced
this pull request
Jul 11, 2026
leanprover-bot
added a commit
to leanprover/reference-manual
that referenced
this pull request
Jul 11, 2026
The compat accessor recomputed the map by folding over all imported constants on every call. `Batteries.Tactic.Lint.getDeclsInPackage` reads it inside a per-declaration fold, turning linting into a quadratic walk over millions of constants on mathlib — `lake lint` effectively never terminated (the bench and mathlib CI lint timeouts). Store the mapping as a `Thunk` on `Kernel.Environment`, filled in `finalizeImport` from the merged views: the runtime forces thunks exactly once, safely across threads (`lean_thunk_get_core` claims the closure atomically and `mark_mt`s the value), so downstream per-declaration reads are constant time again after a single fold, matching the eagerly-built field this accessor replaced. Code that manually reconstructs a `Kernel.Environment` with different views carries the old thunk, which matches how the precomputed field behaved under the same surgery. Verified: `lake lint` on a from-scratch batteries build previously hung indefinitely in `getDeclsInPackage` and now completes. Co-Authored-By: Claude
|
Benchmark results for leanprover-community/mathlib4-nightly-testing@a04b6d1 against leanprover-community/mathlib4-nightly-testing@5edfd4c are in. There are significant results. @Kha
Large changes (1🟥)
Small changes (3978🟥)
|
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/batteries
that referenced
this pull request
Jul 13, 2026
github-actions
Bot
added
the
mathlib4-nightly-available
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/mathlib4-nightly-testing
that referenced
this pull request
Jul 13, 2026
leanprover-bot
added a commit
to leanprover/reference-manual
that referenced
this pull request
Jul 13, 2026
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/batteries
that referenced
this pull request
Jul 15, 2026
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/mathlib4-nightly-testing
that referenced
this pull request
Jul 15, 2026
leanprover-bot
added a commit
to leanprover/reference-manual
that referenced
this pull request
Jul 15, 2026
The extra-constants view (codegen names consulted only when a name misses the constants view) was merged eagerly at import like the constants view, although its duplicate report is not consumed. Merge it lazily instead: a lazy node holds the candidate subtrees sharing its prefix and computes its children on first descent via a thunk, one level at a time; entry lookups at the prefix itself answer directly from the candidates. Thunk forcing is thread-safe and supported on persistent objects, so lazy nodes are safe to share across elaboration threads, and the process-global lookup cache keeps each forced walk to once per name. This removes the extras merge (over half the total merge work) from the import path: `import Mathlib` drops 7% task-clock locally, and even extras-heavy elaboration workloads come out slightly ahead since only touched prefixes are ever merged. Also ports the leanchecker `ReplaceAxiom` helper to the `modifyAt`/`mkMerged` API, which the new constructor's exhaustiveness requirement would otherwise break. Co-Authored-By: Claude
Member
Author
|
!bench |
|
Benchmark results for de7e9f4 against 8006bb0 are in. There are significant results. @Kha Warning These warnings may indicate that the benchmark results are not directly comparable, for example due to changes in the runner configuration or hardware.
Large changes (51✅, 21🟥)
Medium changes (33✅, 11🟥)
Small changes (27✅, 2476🟥)
|
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/batteries
that referenced
this pull request
Jul 15, 2026
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/mathlib4-nightly-testing
that referenced
this pull request
Jul 15, 2026
leanprover-bot
added a commit
to leanprover/reference-manual
that referenced
this pull request
Jul 15, 2026
Member
Author
|
!bench mathlib |
|
Benchmark results for leanprover-community/mathlib4-nightly-testing@5843b6f against leanprover-community/mathlib4-nightly-testing@5edfd4c are in. There are significant results. @Kha
Large changes (2🟥)
Small changes (68🟥)
|
Member
Author
|
!bench |
|
Benchmark results for 6f3486c against 8006bb0 are in. There are significant results. @Kha Warning These warnings may indicate that the benchmark results are not directly comparable, for example due to changes in the runner configuration or hardware.
Large changes (50✅, 22🟥)
Medium changes (4✅, 44🟥)
Small changes (15✅, 2472🟥)
|
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/batteries
that referenced
this pull request
Jul 15, 2026
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/mathlib4-nightly-testing
that referenced
this pull request
Jul 15, 2026
leanprover-bot
added a commit
to leanprover/reference-manual
that referenced
this pull request
Jul 15, 2026
The imported-constant cache marked its values multi-threaded so reader threads could reference them atomically, which makes the reference count of hot values (`Eq`, `Nat`, common instances) a cache line bouncing between cores on every lookup from every elaboration thread. Add `lean_mark_persistent_unshared`, a `lean_mark_persistent` variant that skips multi-threaded objects and their subgraphs: it only writes thread-owned reference counts, so it is safe where the plain variant races, and the skipped fringe stays alive below persistent parents since `dec` never traverses persistent objects. Marking cache values with it makes hits free of atomic reference counting and of coherence traffic: 1-2% fewer cycles on elaboration workloads and 3% less CPU time on parallel-elaboration-heavy ones, at unchanged instruction counts. Co-Authored-By: Claude
Member
Author
|
!bench |
|
Benchmark results for 5460173 against 8006bb0 are in. There are significant results. @Kha Warning These warnings may indicate that the benchmark results are not directly comparable, for example due to changes in the runner configuration or hardware.
Large changes (51✅, 22🟥)
Medium changes (30✅, 10🟥)
Small changes (30✅, 2475🟥)
|
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/batteries
that referenced
this pull request
Jul 15, 2026
mathlib-nightly-testing Bot
pushed a commit
to leanprover-community/mathlib4-nightly-testing
that referenced
this pull request
Jul 15, 2026
leanprover-bot
added a commit
to leanprover/reference-manual
that referenced
this pull request
Jul 15, 2026
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
3 participants
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.
This PR replaces the eager construction of the imported-constants hash maps in
finalizeImport(privateConstantMap,publicConstantMap,const2ModIdx; ~530k insertions forimport Lean) with prefix trees over each module's constant names that are precomputed at olean write time and remain in the memory-mapped compacted regions. Import now only merges the per-module trees, allocating fresh nodes solely for name prefixes shared between modules (~3k nodes / ~81k edges for the stdlib), and lookups walk the merged view, disambiguating children by cached per-prefixNamehash.import Leandrops from ~527ms to ~393ms; olean sizes grow by ~6%.Details:
Lean.ConstTrie: generic per-module tree (ConstTrie α), merged import view (ImportedConsts α) with single-module subtrees borrowed from the regions and module-index tagging, and the newConstMapstructure (importedview +localsmap).ModuleDatagainsconstTrie/extraConstTriefields, filled centrally insaveModuleData(Parts). The extra-const tree replaces the 325k codegen-name entries ofconst2ModIdx(Kernel.Environmentnow holdsallImportedConsts/importedExtraConstsinstead, servinggetModuleIdxFor?).subsumesInforealizability subsumption andthrowAlreadyImported) from the merge's reported collisions.saveModuleDataParts, all parts' data must stay alive across the chained compactor saves; the tries are therefore attached to all parts up front (the compactor tracks compacted objects by heap address, so per-part temporaries would corrupt later parts via address reuse).stage0/src/stdlib_flags.htemporarily enablesinterpreter.prefer_nativeas this is an ABI-breaking change toKernel.Environment/ConstMap(to be reverted byupdate-stage0). Stage 1 tests are meaningless on this commit since the stage 1 oleans lack the newModuleDatafields; validate against stage 2.