docs(bench): record the measured pinned-Linux speed verdict + root cause#268
Merged
Conversation
The pinned-Linux head-to-head (headtohead.yml) was run and OVERTURNED FINDINGS.md's "throughput parity is probably a measurement artifact, confirm before optimizing" guidance: on a clean pinned box IronCache is ~9k qps/core vs redis ~65k/core (~7x), while memory stays a clear win (0.92x). The gap was then isolated to the per-request I/O datapath (NOT cross-shard, NOT thread oversubscription - both proven by measurement: per-core throughput is flat vs shard count, and 1->2 shards on the same cores nearly doubles qps). A CPU profile under load shows ~80% of self-time in syscalls + the async reactor + thread parking (kevent/semwait/sendto/recvfrom) vs ~20% in cache compute. So the throughput lever is the I/O datapath (#28 io_uring / batched event loop), not memory/sharding/threading. Records this so the perf docs reflect the measured truth rather than the now-disproven "confirm later" guidance. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Signed-off-by: Zeke <ezequiel.lares@outlook.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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Records the measured pinned-Linux throughput verdict and root cause in FINDINGS.md (docs-only). The run overturned the prior 'parity is probably a measurement artifact' guidance: ~9k qps/core vs redis ~65k/core (~7x), memory still a clear win (0.92x). Isolated to the per-request I/O datapath (NOT cross-shard, NOT thread oversubscription - both disproven by measurement); CPU profile is ~80% syscalls+reactor+parking. Lever = #28 (io_uring / batched event loop).
🤖 Generated with Claude Code