Repository navigation
Smaller binaries and faster large transfers - #2
Merged
Merged
Conversation
Static Linux x86_64 binary: 50.3 MB -> 27.8 MB. opt-level and panic unwinding stay unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
tests/bench/bench.sh runs cp/pipe/cat/mirror/ls/find/du/rm scenarios against a throwaway tmpfs MinIO for mc, mx (glibc) and mx (musl) and prints a markdown table; RESULTS.md records the first run and the likely causes of the gaps. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
musl's malloc serializes allocations on one lock; parallel small-file work used 1.5-2.8x the CPU of the glibc build. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Uploads follow minio-go: single PUT up to 16 MiB (or --part-size), parts of OptimalPartInfo size (16 MiB multiples, 528 MiB for unknown lengths, as mc pipe), 4 parts in parallel unless -P/--concurrent say otherwise. - Local-file multipart parts stream from file ranges (ByteStream::read_from) instead of Vec buffers; --checksum/object-lock uploads keep buffered parts. Progress counts bytes as the part bodies are sent. - Downloads and local copies write through a 1 MiB buffer, so tokio does one blocking handoff per MiB instead of per 8 KiB. - BlockingReader zeroes at most 64 KiB per read (large part buffers made it quadratic). glibc release, 2 GiB file: cp up 5.84 s -> 0.94 s (mc 1.43 s), RSS 111 -> 49 MiB; cp down 2.02 s -> 0.76 s (mc 0.79 s). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ter perf fixes An empty SOURCE_DATE_EPOCH made GCC reject mimalloc's __DATE__ in local and CI Docker builds. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The reference mc can hang in sql (minio-go event stream); the harness now kills commands after MX_PARITY_TIMEOUT seconds (default 120) and the sql cases retry them. CI jobs get timeout-minutes so a hang no longer runs for six hours. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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
Smaller binaries and faster large transfers.
strip. Static Linux x86_64: 50.3 MB → 27.8 MB.cp2 GiB up: 5.9 s → 0.94 s (mc: 1.41 s). Memory: 111 → 47 MiB.cp2 GiB down: 2.06 s → 0.81 s (mc: 0.87 s).mirrorup: 1.03 s → 0.52 s.SOURCE_DATE_EPOCHbroke the C build).tests/bench/bench.sh, results intests/bench/RESULTS.md.Test plan
cargo test(1311+ tests), clippy, fmtlive_mc_parityagainst MinIOtests/container_smoke.shtests/bench/RESULTS.md)🤖 Generated with Claude Code