zsb/fidc-backtest-engine
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

docs(perf): reconcile storage acceptance with current business main

Browse Source
This commit is contained in:
boris
2026-09-14 10:03:27 +08:00
committed by boris
parent 4c96d0c31f
commit a29c434be9
2 changed files with 21 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/selection-risk-plan-performance-20260914.md
+5 -3
View File
@@ -9,9 +9,11 @@ slightly in the measured cases; this is not the solution to the main remaining
data construction cost and is not a general whole-backtest speedup claim.
The independent [intraday-clock counterexample](intraday-clock-causality-20260914.md)
remains open. This work does not remove that test or its evidence, change the
execution clock, or turn day-level parity into full framework acceptance.
The published service stays at e81bf47/c98bcc3. Source d5b682c6 remains frozen;
was subsequently resolved by business-main work and published in the clock and
81acc54 callback releases. That correction is not attributed to this candidate.
Current published Backtest uses e81bf47/81acc54; the measurements below retain
their earlier c98 baseline. Complete manual-replay integration and publication
of this performance candidate remain open. Source d5b682c6 remains frozen;
research and signal work stay paused. No trading operation was submitted.
## Evidence Leading to the Change
docs/series-column-storage-performance-20260914.md
+16 -5
View File
@@ -2,6 +2,14 @@
## Status
The subsequent business-main merge includes the separately published 81acc54
clock/callback fixes and 5e11f3d manual-replay foundation. The combined version
passed 857 core, 448 runner and 119 API tests, six long reference replays and
three additional strategy replays;
see `/Users/boris/WorkSpace/fidc-backtest-service/docs/arrow-factor-scratch-rejection-20260914.md`.
The scratch candidate from that experiment was removed. Series storage is
still not deployed; existing measurements below retain their original versions.
Candidate not deployed. Twelve real long replays preserve their independent
business baselines and reduce peak RSS by about 9.5%. Construction latency is
mixed, including a reversed pair where the control is faster. This is accepted
@@ -9,9 +17,11 @@ as evidence of a smaller working set, not as a proved general speedup or closure
of the main performance objective. Original results are retained unchanged.
The Source implementation remains d5b682c6d09704ff23d725a8dd8b155db3eb6967.
Research/signal work remains paused. Current service remains the e81bf47/c98
release; no Source, trading or other-user process was restarted or modified.
The independent intraday-clock counterexample remains open.
Research/signal work remains paused. The initial experiments ran while e81bf47/c98
was published; later business work published e81bf47/81acc54. This performance
task did not restart Source, trading or another user's process. The original
clock counterexample was resolved by that business work; complete manual-replay
integration remains open and is not proved by these performance tests.
## Separate DayOpen Correction
@@ -104,8 +114,9 @@ compact verified summary is stored here to avoid duplicating input inventories.
Do not publish this as the main performance fix. Next, target the remaining
daily snapshot/factor construction and direct typed-column reuse, avoiding
new per-access branches or post-hoc compression passes. Cold-query acceptance,
real minute-mode acceptance, HTTP publication and the intraday-clock issue are
still outstanding. Signal lifecycle, financial PIT and UI factor conditions
real minute-mode acceptance and HTTP publication remain outstanding for this
storage candidate. The original clock issue was fixed by subsequent business
work, not this experiment. Signal lifecycle, financial PIT and UI factor conditions
remain outside this completed storage experiment.
- Engine candidate: 996b909608589fb1987f33c0cfb4c62099f69617.