You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A part of #8243. PRs #5073 followed by #5258, aiming to fix an obscure Broadcom GPU bug, also increased the amount of data we upload to the GPU for line layers by 50% — from 8 to 12 bytes. This directly affects rendering performance.
We should explore alternative packing approaches to get this back to 8 bytes while avoiding the GPU bug trigger. One thing to try would be to try packing the round/up attributes into most significant bit rather than least significant. cc @ansis
The text was updated successfully, but these errors were encountered:
A part of #8243. PRs #5073 followed by #5258, aiming to fix an obscure Broadcom GPU bug, also increased the amount of data we upload to the GPU for line layers by 50% — from 8 to 12 bytes. This directly affects rendering performance.
We should explore alternative packing approaches to get this back to 8 bytes while avoiding the GPU bug trigger. One thing to try would be to try packing the
round
/up
attributes into most significant bit rather than least significant. cc @ansisThe text was updated successfully, but these errors were encountered: