raphael
dae6bfd747
Tone pass across the non-tutorial markdown — README, PROOFS,
ARCHITECTURE, BENCHMARKS, COMPLIANCE, FAQ, MES_INTEROP, SECURITY,
and interop/README. Three patterns came out:
- Bug-history war stories ("Past interop sweeps surfaced…",
"What these harnesses caught: 1. Strict U-width parsing…").
- Chat-with-reader framing ("Don't skip TLS unless…", "Treat as a
punch list", "If you're running in a pod…", "Misconfiguration
incidents drop dramatically").
- Self-referential narration ("we ship", "our codec", "the
codebase's most-tested layer", "three orders of magnitude above
fab load", "the gift that keeps giving").
README also drops the standalone ThreadSanitizer subsection under
Build details (now a single line under the new Testing section).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
66 lines
2.7 KiB
Markdown
66 lines
2.7 KiB
Markdown
# Performance baseline
|
||
|
||
Numbers from `build/secs_bench --requests 20000 --concurrency 16` on
|
||
Docker / Ubuntu 24.04 inside Docker Desktop on macOS (M-series), single
|
||
io_context thread. A rough capacity-planning envelope, not a
|
||
lab-grade benchmark — re-run on target hardware before sizing.
|
||
|
||
## Round-trip throughput / latency
|
||
|
||
| Scenario | Ops | Elapsed | Ops/sec | p50 µs | p95 µs | p99 µs |
|
||
|----------------------------------|--------:|--------:|-----------:|--------:|--------:|--------:|
|
||
| S1F1/F2 (header-only) | 20000 | 0.14 | ~140000 | 74 | 103 | 161 |
|
||
| S1F3/F4 (32 SVIDs) | 20000 | 0.25 | ~79000 | 165 | 186 | 260 |
|
||
| S6F11 push (W=0) | 20000 | 0.03 | ~572000 | n/a | n/a | n/a |
|
||
|
||
A real fab tool sustains tens to a few hundred S6F11 events/second.
|
||
Throughput is not the bottleneck on this stack; latency tail under
|
||
contention is.
|
||
|
||
## Memory footprint
|
||
|
||
A `ProcessJob` + `ControlJob` pair (no persistence enabled) is around
|
||
**~450 bytes** of heap (1000 pairs ≈ 0.45 MiB, measured on a fresh
|
||
process). With persistence enabled add ~200 bytes of in-memory journal
|
||
path tracking per record.
|
||
|
||
| Active entity | Approx bytes / instance |
|
||
|----------------------|------------------------:|
|
||
| PJ + CJ pair | ~450 |
|
||
| Carrier (no slots) | ~80 |
|
||
| Carrier slot | ~24 |
|
||
| Substrate | ~120 |
|
||
| Spool entry | ~40 + encoded body size |
|
||
|
||
A busy fab tool tracking 50 carriers × 25 slots, 200 substrates, 20
|
||
active PJ+CJ pairs comes in well under 1 MiB of model state. RSS will
|
||
be dominated by the binary itself + asio's buffers (~10-20 MiB),
|
||
not the model.
|
||
|
||
## How to re-run
|
||
|
||
```sh
|
||
docker compose run --rm builder /app/build/secs_bench \
|
||
--requests 50000 \
|
||
--concurrency 32 \
|
||
--svid-count 32 \
|
||
--store-pairs 10000
|
||
```
|
||
|
||
Output is markdown.
|
||
|
||
## What this does NOT measure
|
||
|
||
- **Real network**. Loopback TCP has no MTU fragmentation, no
|
||
retransmits, no jitter. Production HSMS over a fab control LAN will
|
||
see higher tail latency.
|
||
- **Persistence write amplification**. The bench runs with persistence
|
||
disabled. Each store mutation with persistence enabled is one
|
||
atomic-rename to disk; on rotational media that limits you to a few
|
||
hundred mutations/sec. SSD-backed deployments are fine.
|
||
- **Concurrent S6F11 enable filtering**. Real CEID emission gates on
|
||
the host's enable/disable list — this bench fires raw S6F11s.
|
||
- **Multi-session HSMS-GS** dispatch overhead — single-session only.
|
||
- **TLS-tunneled sockets** (via stunnel/sidecar) — these add ~50 µs
|
||
per round-trip on modern hardware.
|