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
While performance may not be relevant to all users, I'm a little curious why perf of this build is poor relative to other builds.
For reference I have run small comparisons on this build vs a fresh build using VS2017 + v141_xp toolchain built with /O2 (similar to this one ), and also Olaf's cairo/sqlite mashup from http://www.vbrichclient.com/#/en/Downloads.htm (which can be used as a "drop-in replacement" for VBcairo.dll).
Testing something simple like rendering a large radial gradient (8 mp) yields times roughly on the range of 800 ms for this build, 600 ms for Olaf's build, and 400 ms for a default /O2 build. Linear gradients show a similar perf spread.
These timings seem particularly odd since Olaf's cairo is based off a very old build (1.9.4 as reported by the dll) and I would expect at least some pixman improvements since then.
Given that an /O2 build adds barely 200 kb to the final dll, would it be worthwhile to revisit current optimization settings?
The text was updated successfully, but these errors were encountered:
While performance may not be relevant to all users, I'm a little curious why perf of this build is poor relative to other builds.
For reference I have run small comparisons on this build vs a fresh build using VS2017 + v141_xp toolchain built with /O2 (similar to this one ), and also Olaf's cairo/sqlite mashup from http://www.vbrichclient.com/#/en/Downloads.htm (which can be used as a "drop-in replacement" for VBcairo.dll).
Testing something simple like rendering a large radial gradient (8 mp) yields times roughly on the range of 800 ms for this build, 600 ms for Olaf's build, and 400 ms for a default /O2 build. Linear gradients show a similar perf spread.
These timings seem particularly odd since Olaf's cairo is based off a very old build (1.9.4 as reported by the dll) and I would expect at least some pixman improvements since then.
Given that an /O2 build adds barely 200 kb to the final dll, would it be worthwhile to revisit current optimization settings?
The text was updated successfully, but these errors were encountered: