Improves CI performance

[NickSdot]

WIP

tl/dr: ~45m -> ~21m; this is CI-only stuff -- everything in #22917 reduces further.

Results

f47a44e

Faster Windows CI through parallel builds/tests and overlapping database setup

Windows x64	Before	After	Improvement
Setup + build	10:53	3:22	69%
Tests	30:37	14:48	52%
Full job	41:55	18:48	55%

Cache hit rate: ~95%.

8e526f1

Benchmark checkout avoids downloading unnecessary historical file contents

Step	Before	After	Improvement
Benchmark checkout	~1:11	0:20	72%

d2b0d08

Keeps non-Windows compiler caches fresh

Before: the older the weekly cache, the lower the hit rate. 2026-07-27 was ~99% (0:30s) while 2026-08-02 was only ~25% (6:52). Rolling cache solves that. ¹

Compile job	Before 2	After	Improvement
Linux release	7:54	0:28	94%
Benchmark	2:33	0:17	89%
Unit tests	2:06	0:14	89%
Linux ASAN	1:50	0:27	75%
macOS	4:07	1:30	64%
Alpine ASAN	2:37	0:57	64%
Linux x32	2:10	0:57	56%

Cache hit rate: ~99%.

9de2a2e

Skip unused setup in the ordinary Linux ASAN job

Setup	Before	After	Improvement
Service containers	1:24	0:13	85%
MSSQL setup	0:14	Skipped	100%
Apt dependencies	0:54	0:42	22%
Total targeted setup	2:32	0:55	64%

Saves 1:37 per ASAN run. Depending on download speeds shaves off 1–3 min.

7507629

Run Function JIT tests concurrently

This will currently not reduce total CI time, because they are not the longest running job. However, runnign them in parallel gives earlier feedback. This will help when all the other non-CI PRs landed.

Footnotes

1. each new master commit rolls the cache forward. ↩

2. when cache not fresh. ↩