folly — Engineering Performance
61 engineers all time · Jan 2025 – Nov 2026 · built 2026-11-22 · GitHub
Performance snapshot
Today's rolling 90-day reading for folly, compared with the start of the series. Pick a window to move that comparison point.
Avg. perf / dev / mo
+373.3%
0.25 → 1.20 ETV
Active engineers
+12.0%
25.0 → 28.0
Features
−11.1pp
31.0% → 19.9%
vs. Meta
0.35x
0.36x → 0.35x · −65% below
folly vs. Meta
Per-engineer ETV for folly against Meta as a whole. Both lines are 90-day rolling averages scaled to a 30-day month, so they share one axis and can be read against each other at any point. Pick a window to zoom the chart to it.
Performance Composition
Each month's output split by type of work: Features (new value), Maintenance (sustaining systems), Tests, Docs, and Fixes (rework). The yellow line is output per engineer, so when it rises each engineer is delivering more, whatever the team size did. Unit: Engineering Throughput Value (ETV).
Engineering capacity
Effective engineers behind folly, against its pre-AI baseline. Each subject has its own: folly's is 0.40 ETV / dev / mo, its first reading in Q1 2025. Per-engineer ETV divided by that gives a capacity multiple, and that multiple applied to the engineers active in the trailing 90 days turns it into engineer-equivalents. The line is the real headcount, so the gap between line and area is what the leverage is worth. Because each baseline is its own, every subject opens at 1.0x on its first day: multiples measure improvement and are not comparable between subjects.
Knowledge concentration
How dependent is this repo on a small number of engineers? Higher top-1 share = higher key-person risk.
Yedidya Feldblum owns 29.2 % of commits.
Behind the numbers
Written summary of the work completed each month.
No monthly reports available yet.
Top engineers
Most impactful commits
Top 10 by ETV in the all-time window.
- 2.8ETVrecover from malformed DWARF line tables Summary: Parse DWARF line-table headers, directory and file entries, and line-program opcodes with DwarfCursor. Propagate malformed data as recoverable failures, validate string offsets and value types, and check address and line arithmetic. Reviewed By: vitaut, dtolnay Differential Revision: D123635592 fbshipit-source-id: 24000e9aa6bc0895d31fb258a82d1f788270faa7Yedidya Feldblum · f70c5b9d · 2026-11-13
- 2.6ETVuint_divisor, lifted from thrift Fast64BitRemainderCalculator Summary: Lift `apache::thrift::frozen::detail::Fast64BitRemainderCalculator` into a new public header `folly/math/Division.h` as `folly::uint_divisor<Word>`, generalized to all unsigned integer types. Uses Lemire's constant-divisor technique (credited in doc-comments). Key changes vs. the thrift original: - Support all unsigned integer types, possibly falling back to integer division. - Support zero divisors since the optimized algorithm does not perform division. - Calculates composite division and remainder, division, or remainder. - `constexpr` construction and invocation. - Invocable-object usage with `operator()` in addition to `divrem`. Mathematical operator usage with `operator/` and `operator%` in addition to `div` and `rem`. - Thrift uses Lemire’s direct fast-remainder algorithm: multiply by a precomputed reciprocal, then perform another widened multiply to recover the remainder. uint_divisor uses the same reciprocal to compute the exact quotient and derives the remainder as dividend - quotient * divisor. For 64-bit words this reduces the hot path from roughly four multiplies to three while remaining branchless. Narrow types retain Lemire’s direct remainder path where it benchmarks faster. - Hold the original divisor in addition to the computed multiplier to support mathematical operator usage. This increases in-situ object size. - Add `uint_divisor<Word>::calc` that does not hold the original divisor; let both `uint_divisor` and `Fast64BitRemainderCalculator` delegate to it. Reviewed By: iahs Differential Revision: D115907214 fbshipit-source-id: de6d55b8dac71489f4f36b382eeb9e065380cad1Yedidya Feldblum · 653d0b3b · 2026-08-22
- 2.1ETV`benchmark_ab.py`: multi-binary benchmark A/B reports Summary: `benchmark_ab.py` is a new tool to simplify iterating on changes that affect several benchmark binaries. It handles three recurring chores: - Finds benchmark binaries from Buck target patterns. - Aggregates repeated A/B runs into one report. - Prioritizes wins and regressions with absolute and relative thresholds. Read the file docblock for more. ```text $ folly/tool/benchmark_ab.py measure --before=bc56e16776 --after=0097974dca \ //folly/result/... ... High-priority regressions: 15.2+11.2ns (+73.5%): try_to_result_error fbcode//folly/result/test:result_bench 15.2+11.2, 15.2+11.3, 15.2+11.1, 15.2+11.2, 15.4+11.2 ``` Reviewed By: janondrusek Differential Revision: D112568029 fbshipit-source-id: 544b93f75660e791285743188150a3eec38b1e1eAlexey Spiridonov · 206fc16e · 2026-07-28
- 2.1ETVLet ambient request contexts own local lifetime Summary: Create thread-confined ambient contexts with `LocalRequestContextScopeGuard`. The ambient binding uniquely owns the lifetime wrapper and its FRC. Replace `RequestContextOwner` with a move-only `RequestContextHandle` that holds one non-atomic delayed-destruction pin. `saveContextLocal()` borrows the current local binding; it does not allocate a context or acquire shared ownership. Use the existing `DelayedDestructionUniquePtr` and `makeDelayedDestructionUniquePtr` for ambient ownership. Leaving an ambient scope requests destruction, while active or suspended scopes and caller handles postpone it. Releasing the last pin while the ambient owner remains must not delete FRC. A retained handle can reactivate the same context after the creating ambient scope ends. Allocate shared ownership when an owning API captures the local context. For CPU dispatch, call `capture()` or `saveContext()` on the creating thread before offload. Keep the promoted allocation alive until the ambient owner has requested destruction and all EVB pins have released. CPU captures can release concurrently without touching the thread-confined destruction count. Repeated captures preserve context identity and reuse the shared control block. Keep `RequestContext` layout unchanged, preserve scope restoration and callbacks, and reject local snapshots of shared bindings rather than adding a shared-pointer fallback to the handle. Existing owning APIs continue to return owning pointers. Reviewed By: yfeldblum Differential Revision: D127497243 fbshipit-source-id: bc2fb8f060091f5a1e890cc67205399c70f95f6aAnkit Kumar · 35b8f838 · 2026-11-03
- 2.0ETVrecover from malformed DWARF abbreviations and DIEs Summary: Parse abbreviation declarations and DIE codes with DwarfCursor, validate table structure and offsets, and propagate malformed data as structured errors. Existing symbolization APIs collapse these errors to unavailable optional metadata. Reviewed By: dtolnay Differential Revision: D123374802 fbshipit-source-id: 886f98350db90d96bbcbf9882ac3a079021fa67cYedidya Feldblum · 8ed89be9 · 2026-11-11
- 2.0ETVcli_apply_args_files Summary: In the context of CLI args-parsing, a facility to apply args-files to an args vector. An arg beginning with `@` but not with `@@` is treated as a filename, relative to the current directory if relative, after the `@`. That arg is then replaced by parsing the file into a sequence of arguments, recursively. Reviewed By: hchokshi Differential Revision: D92167066 fbshipit-source-id: 9652db02b04202fa258f6589d468568d5f51d3d5Yedidya Feldblum · 56a01e21 · 2026-02-10
- 1.9ETVFix F14Table UBSAN insufficient-object-size error Summary: UBSAN fires `insufficient-object-size` when F14Table's single-chunk optimization allocates less than `sizeof(F14Chunk)` bytes, then calls member functions on the `Chunk*`. For example, `F14Chunk<unsigned int>` had `sizeof` = 64 bytes, but a 2-element table allocates only 24 bytes (16-byte header + 2×4-byte items). Fix by removing the `rawItems_` member array from `F14Chunk` so that `sizeof(F14Chunk)` equals 16 (just the header: `tags_[14]` + `control_` + `outboundOverflowCount_`). Items are accessed via pointer arithmetic from `this + kItemsOffset` instead of through the member array. Multi-chunk tables use an explicit stride constant (`kChunkStride`, equal to the old `sizeof`) for chunk-to-chunk navigation. This preserves the single-chunk memory optimization while making all member calls on `Chunk*` UBSAN-clean: - Minimum real allocation = 16 + 2*sizeof(Item) >= 24 > 16 = sizeof(Chunk) - Empty instance (F14EmptyTagVector) = 16 bytes = sizeof(Chunk) - Multi-chunk allocations use kChunkStride * N (same total size) Reviewed By: ilvokhin Differential Revision: D94696039 fbshipit-source-id: 06ef5480be9a2fdf3f0beb1e2c87916667e72abeYedidya Feldblum · 381d0747 · 2026-03-03
- 1.9ETVAsyncSocketTest: parametrize with IoUringBackend Summary: Validate native AsyncSocket support with IoUringBackend. Parametrize most tests (that make sense) with the backend, testing both the standard libevent backend and the io_uring backend. Reviewed By: vishwanath1306 Differential Revision: D90996725 fbshipit-source-id: a60b1702f1b6efe8d089ef1f2cb8e5b2d8207f42David Wei · 193b53f3 · 2026-01-21
- 1.6ETVrecover from malformed ELF tables Summary: Validate `ElfFile` headers, section and program tables, bodies, string tables, entry sizes, indexes, and address arithmetic before dereferencing mapped data. Return existing null or empty sentinels for malformed structures, and make throwing `ElfFile::open` report invalid files with `std::runtime_error` instead of terminating. Reviewed By: dtolnay Differential Revision: D125909726 fbshipit-source-id: c9e74499b2c2d67b695ddd727ae1ff257cbf2b6eYedidya Feldblum · 80a13741 · 2026-11-19
- 1.5ETValgorithms for byte-sized find_first_of, find_first_not_of Summary: Includes a selection of scalar and vector algorithms all implementing `find_first_of` and `find_first_not_of`. May be useful to accelerate select parsers. Reviewed By: DenisYaroshevskiy Differential Revision: D61775260 fbshipit-source-id: 32fdd8e7c02c66a54764b0ae847aec1fa9564678Yedidya Feldblum · 8e49a654 · 2025-03-18