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

docs(perf): record combined risk and series storage Backtest release

Browse Source
This commit is contained in:
boris
2026-09-14 11:13:50 +08:00
committed by boris
parent 665653c3fe
commit 93de28d369
2 changed files with 48 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/series-column-storage-performance-20260914.md
+23 -4
View File
@@ -8,9 +8,10 @@ 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.
now published to Backtest only in the combined 665653c/501f6d0 release below;
existing measurements retain their original versions.
Candidate not deployed. Twelve real long replays preserve their independent
The original 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
as evidence of a smaller working set, not as a proved general speedup or closure
@@ -114,8 +115,9 @@ 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Do not publish this as the main performance fix.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 and HTTP publication remain outstanding for this
storage candidate. The original clock issue was fixed by subsequent business
real minute-mode acceptance remain outstanding for this storage change. The
combined version subsequently passed daily HTTP publication below. 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.
@@ -126,3 +128,20 @@ remain outside this completed storage experiment.
- Evidence: /srv/fidc/canonical/run/research/series-column-storage-20260914.
[Verified summary](evidence/series-column-storage-20260914/acceptance.json).
## Combined Release
Engine 665653c / service 501f6d0 passed 860 core, 448 runner and 119 API tests,
six additional independent-process replays and nine post-publication HTTP runs.
Complete canonical/store results remain equal to each strategy's own baseline.
An adjacent original/new rotation pair measures 15.279/14.579 seconds and
6,923,640/6,369,900 KiB peak RSS, but the default-window HTTP means are essentially
unchanged (13.646/13.740 seconds). Reduced memory and conditional cross-window
reuse are not promoted to a universal latency improvement.
The official Backtest-only publication preserves Source d5, paused research,
trading services and execution permissions. Its active build root
/srv/fidc/canonical/build/factor-reserve-20260913 must not be overwritten or
reused. Shared DataSet acceptance proves input reuse while distinct strategies
execute independently; no results are cached. Identities and original receipts:
`/Users/boris/WorkSpace/fidc-backtest-service/docs/cache-boundary-planning-performance-20260914.md`.