8a52fdc8

By: Kaapo Seppänen <seppanenkaapo@gmail.com>

load the builtin SM in paired performance benchmarks

The SM JSON model now stores rational and optional hypercharges, which the
baseline binary cannot parse from the candidate's shared model files. The
GL15 and GL1 cards import sm-default so each revision loads its own
embedded model; results must still agree between revisions.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Total

7m29s

Time to Start Worker time Duration Time to finish Idle
Config 0s 1s 1s 2s 0s
Eval 2s 1m28s 1m28s 1m30s 0s
Build 1m29s 0s 0s 1m29s 0s
Test 1m30s 16m33s 5m36s 7m06s 0s
Deploy 7m12s 17s 17s 7m29s 5s
Suite 0s 18m21s 7m29s 7m29s 5s

Timeline

Config Eval Build Test Deploy Compressed gap
0s1m2m3m4m5m6m7m