09:02 snowycoder[d]: mhenning[d]: Thanks for the review!
09:02 snowycoder[d]: How can I measure compile-time in shader-db? Is there a metric already or do I need to manually wrap the code with `Instant::now`?
16:52 mhenning[d]: snowycoder[d]: `fossilize-replay --enable-pipeline-stats` will already give you a csv with some compile durations in it per-pipeline
16:53 mhenning[d]: What I've been doing is roughly:
16:53 mhenning[d]: 1) Build a release build of mesa and enter the devenv
16:53 mhenning[d]: 2) Run `MESA_SHADER_CACHE_DISABLE=true time fossilize-replay --enable-pipeline-stats stats/snowy_cross_block_1_vec.csv --num-threads 12 fossils/**/*.foz ../nvk-fossils-foss/fossils/*/*.foz` on both a baseline commit and on the modified branch (with different csv names)
16:53 mhenning[d]: 3) Run a script to compare the csv files and calculate summary statistics
16:54 mhenning[d]: https://cdn.discordapp.com/attachments/1034184951790305330/1417191784458489967/compare_time.py?ex=68c99633&is=68c844b3&hm=f9c806c58fda561edb086f92126f014b4db45cc699493ed9261501c7fd2ffc2e&
16:54 mhenning[d]: For 3) my script is this
22:01 mangodev[d]: how do you force nouveau gl on git mesa?
22:02 mangodev[d]: nothing on the mesa3d.org page
22:02 mangodev[d]: *nouveau FOR gl
22:06 mhenning[d]: assuming you mean the old gl driver, `NOUVEAU_USE_ZINK=0` same as before
22:07 mangodev[d]: mhenning[d]: ah, ty
22:07 mangodev[d]: i knew it was a thing, i just forgot the exact env var for it
22:07 mangodev[d]: currently running into some weird system issues and i think some new zink changes broke some things