Skip to content

Some bcsel of fsat stuff

Ian Romanick requested to merge idr/mesa:review/no-copy-prop-sat-to-sel into main

I saw this while working on something else, and I wondered, "Does this optimization actually do anything?" It turns out, no. There was also no shader-db stats when it originally landed. In the one fossil-db shader that was improved by removing it, I noticed the pattern in the follow-up algebraic optimization.

Merge request reports