Add amd64 tarball artifact and guard packaging alternatives#1854
Merged
Conversation
4 tasks
oferchen
added a commit
that referenced
this pull request
May 5, 2026
Static-analysis half of #1854: predicts cache-line residency, branch behavior, and working-set size of BoundedReorderBuffer at 100K, 1M, and 10M items. Compares the BTreeMap-backed buffer with the ring-based alternative already shipped in engine::concurrent_delta::ReorderBuffer (#1734) and lists the perf counters and decision criteria the runtime half must capture.
oferchen
added a commit
that referenced
this pull request
May 5, 2026
…3707) Static-analysis half of #1854: predicts cache-line residency, branch behavior, and working-set size of BoundedReorderBuffer at 100K, 1M, and 10M items. Compares the BTreeMap-backed buffer with the ring-based alternative already shipped in engine::concurrent_delta::ReorderBuffer (#1734) and lists the perf counters and decision criteria the runtime half must capture.
This was referenced May 7, 2026
oferchen
added a commit
that referenced
this pull request
May 7, 2026
…#3827) Benchmark plan for cachegrind/perf profiling of BoundedReorderBuffer (crates/transfer/src/reorder_buffer.rs) at 1M files. Documents BTreeMap node layout, names the ring-buffer alternative from #1853, defines the three workloads (in-order, mostly-in-order with 5% gap, pathological), and pins L1d/LLC miss rate plus IPC as the decision metrics. Tracking task: oc-rsync #1854.
3 tasks
oferchen
added a commit
that referenced
this pull request
May 7, 2026
#3941) Specifies perf stat and cachegrind plan to test whether BTreeMap node scatter dominates ReorderBuffer dispatch cost, and the 2x cycle/miss threshold a VecDeque ring buffer must beat to justify replacement.
oferchen
added a commit
that referenced
this pull request
May 16, 2026
Add a Criterion bench that exercises ReorderBuffer's insert/drain path at 1M out-of-order indexed items across three payload sizes (32 B, 256 B, 4 KB) and three insertion patterns (fully reverse, random shuffle, near-in-order with 10% deltas). The bench is gated behind BENCH_REORDER_CACHE=1 so default cargo bench runs stay fast. The top-of-file documents how to drive perf stat (Linux) and cachegrind (any platform) against the produced binary, and how to interpret the numbers - favorable cache metrics keep the existing Box<[Option<T>]> layout, while unfavorable metrics motivate a flat Vec plus occupancy-bitmap layout or a hot/cold storage split. Payloads are pre-built outside the timed section via iter_batched so the measurement reflects ReorderBuffer ops, not allocation. Capacity is sized per pattern so no insert ever triggers grow().
oferchen
added a commit
that referenced
this pull request
May 16, 2026
Add a Criterion bench that exercises ReorderBuffer's insert/drain path at 1M out-of-order indexed items across three payload sizes (32 B, 256 B, 4 KB) and three insertion patterns (fully reverse, random shuffle, near-in-order with 10% deltas). The bench is gated behind BENCH_REORDER_CACHE=1 so default cargo bench runs stay fast. The top-of-file documents how to drive perf stat (Linux) and cachegrind (any platform) against the produced binary, and how to interpret the numbers - favorable cache metrics keep the existing Box<[Option<T>]> layout, while unfavorable metrics motivate a flat Vec plus occupancy-bitmap layout or a hot/cold storage split. Payloads are pre-built outside the timed section via iter_batched so the measurement reflects ReorderBuffer ops, not allocation. Capacity is sized per pattern so no insert ever triggers grow().
oferchen
added a commit
that referenced
this pull request
May 16, 2026
Add a Criterion bench that exercises ReorderBuffer's insert/drain path at 1M out-of-order indexed items across three payload sizes (32 B, 256 B, 4 KB) and three insertion patterns (fully reverse, random shuffle, near-in-order with 10% deltas). The bench is gated behind BENCH_REORDER_CACHE=1 so default cargo bench runs stay fast. The top-of-file documents how to drive perf stat (Linux) and cachegrind (any platform) against the produced binary, and how to interpret the numbers - favorable cache metrics keep the existing Box<[Option<T>]> layout, while unfavorable metrics motivate a flat Vec plus occupancy-bitmap layout or a hot/cold storage split. Payloads are pre-built outside the timed section via iter_batched so the measurement reflects ReorderBuffer ops, not allocation. Capacity is sized per pattern so no insert ever triggers grow().
oferchen
added a commit
that referenced
this pull request
May 16, 2026
Add a Criterion bench that exercises ReorderBuffer's insert/drain path at 1M out-of-order indexed items across three payload sizes (32 B, 256 B, 4 KB) and three insertion patterns (fully reverse, random shuffle, near-in-order with 10% deltas). The bench is gated behind BENCH_REORDER_CACHE=1 so default cargo bench runs stay fast. The top-of-file documents how to drive perf stat (Linux) and cachegrind (any platform) against the produced binary, and how to interpret the numbers - favorable cache metrics keep the existing Box<[Option<T>]> layout, while unfavorable metrics motivate a flat Vec plus occupancy-bitmap layout or a hot/cold storage split. Payloads are pre-built outside the timed section via iter_batched so the measurement reflects ReorderBuffer ops, not allocation. Capacity is sized per pattern so no insert ever triggers grow().
oferchen
added a commit
that referenced
this pull request
May 16, 2026
… (#4180) * chore(bench): profile ReorderBuffer cache behavior at 1M items (#1854) Add a Criterion bench that exercises ReorderBuffer's insert/drain path at 1M out-of-order indexed items across three payload sizes (32 B, 256 B, 4 KB) and three insertion patterns (fully reverse, random shuffle, near-in-order with 10% deltas). The bench is gated behind BENCH_REORDER_CACHE=1 so default cargo bench runs stay fast. The top-of-file documents how to drive perf stat (Linux) and cachegrind (any platform) against the produced binary, and how to interpret the numbers - favorable cache metrics keep the existing Box<[Option<T>]> layout, while unfavorable metrics motivate a flat Vec plus occupancy-bitmap layout or a hot/cold storage split. Payloads are pre-built outside the timed section via iter_batched so the measurement reflects ReorderBuffer ops, not allocation. Capacity is sized per pattern so no insert ever triggers grow(). * style(engine): fix clippy doc-lazy-continuation and hex literal grouping * style(engine): break paragraph out of nested list in bench doc
oferchen
added a commit
that referenced
this pull request
May 18, 2026
…3707) Static-analysis half of #1854: predicts cache-line residency, branch behavior, and working-set size of BoundedReorderBuffer at 100K, 1M, and 10M items. Compares the BTreeMap-backed buffer with the ring-based alternative already shipped in engine::concurrent_delta::ReorderBuffer (#1734) and lists the perf counters and decision criteria the runtime half must capture.
oferchen
added a commit
that referenced
this pull request
May 18, 2026
…#3827) Benchmark plan for cachegrind/perf profiling of BoundedReorderBuffer (crates/transfer/src/reorder_buffer.rs) at 1M files. Documents BTreeMap node layout, names the ring-buffer alternative from #1853, defines the three workloads (in-order, mostly-in-order with 5% gap, pathological), and pins L1d/LLC miss rate plus IPC as the decision metrics. Tracking task: oc-rsync #1854.
oferchen
added a commit
that referenced
this pull request
May 18, 2026
#3941) Specifies perf stat and cachegrind plan to test whether BTreeMap node scatter dominates ReorderBuffer dispatch cost, and the 2x cycle/miss threshold a VecDeque ring buffer must beat to justify replacement.
oferchen
added a commit
that referenced
this pull request
May 18, 2026
… (#4180) * chore(bench): profile ReorderBuffer cache behavior at 1M items (#1854) Add a Criterion bench that exercises ReorderBuffer's insert/drain path at 1M out-of-order indexed items across three payload sizes (32 B, 256 B, 4 KB) and three insertion patterns (fully reverse, random shuffle, near-in-order with 10% deltas). The bench is gated behind BENCH_REORDER_CACHE=1 so default cargo bench runs stay fast. The top-of-file documents how to drive perf stat (Linux) and cachegrind (any platform) against the produced binary, and how to interpret the numbers - favorable cache metrics keep the existing Box<[Option<T>]> layout, while unfavorable metrics motivate a flat Vec plus occupancy-bitmap layout or a hot/cold storage split. Payloads are pre-built outside the timed section via iter_batched so the measurement reflects ReorderBuffer ops, not allocation. Capacity is sized per pattern so no insert ever triggers grow(). * style(engine): fix clippy doc-lazy-continuation and hex literal grouping * style(engine): break paragraph out of nested list in bench doc
oferchen
added a commit
that referenced
this pull request
May 18, 2026
… (#4180) * chore(bench): profile ReorderBuffer cache behavior at 1M items (#1854) Add a Criterion bench that exercises ReorderBuffer's insert/drain path at 1M out-of-order indexed items across three payload sizes (32 B, 256 B, 4 KB) and three insertion patterns (fully reverse, random shuffle, near-in-order with 10% deltas). The bench is gated behind BENCH_REORDER_CACHE=1 so default cargo bench runs stay fast. The top-of-file documents how to drive perf stat (Linux) and cachegrind (any platform) against the produced binary, and how to interpret the numbers - favorable cache metrics keep the existing Box<[Option<T>]> layout, while unfavorable metrics motivate a flat Vec plus occupancy-bitmap layout or a hot/cold storage split. Payloads are pre-built outside the timed section via iter_batched so the measurement reflects ReorderBuffer ops, not allocation. Capacity is sized per pattern so no insert ever triggers grow(). * style(engine): fix clippy doc-lazy-continuation and hex literal grouping * style(engine): break paragraph out of nested list in bench doc
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
1 participant
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.
Summary
Testing
https://chatgpt.com/codex/tasks/task_e_690575c1982c83239fbdea27b8191981