Record preparation validation timings and measurement limits

Co-Authored-By: Codex <noreply@openai.com>
This commit is contained in:
rdbtCVS
2026-10-25 23:15:40 +00:00
parent 03190ec724
commit 3adff150cf
+14 -6
View File
@@ -45,9 +45,17 @@ for every file stable across the two isolated runs. For the three known noisy JS
it also requires semantic equality after normalizing object key order, opaque JVM lambda
addresses and creative-tab tag set order. Other variation fails the release check.
On Linux with Java 25, two comparisons produced 23,204 matching resource files each.
The first isolated runs took 178.95 s and 104.25 s, versus 45.05 s batched; the second took
100.03 s and 176.06 s, versus 120.09 s batched. All used precompiled exporters. These are fresh output directories with a warm OS file cache, on a
machine with other workloads; the spread between isolated runs prevents treating this as a
controlled speedup estimate. The compiler saving is separate: release preparation removes
the previously measured 15 compiler invocations (29–52 s).
On Linux with Java 25, the output comparisons used 23,204 resource files per run.
All runs used precompiled exporters:
| Comparison | Isolated A | Isolated B | Batched |
|---|---:|---:|---:|
| Initial | 178.95 s | 104.25 s | 45.05 s |
| Repeat | 100.03 s | 176.06 s | 120.09 s |
| Final, immutable collision cache | 133.80 s | 123.79 s | 56.19 s |
These use fresh output directories with a warm OS file cache, on a machine with other
workloads. The first two rows include output hashing; the final row measures preparation
before comparison. The variability prevents treating these as a controlled speedup estimate.
The compiler saving is separate: release preparation removes the previously measured
15 compiler invocations (29–52 s).