-
Re: Fx 125.0.3
2024年4月30日 2:20 AMTeteHello Rodri-san,
thank you for letting me know the infomation.
I have uploaded my build.Re: Fx 125.0.3
2024年4月30日 3:30 AMUliThank you for 125.0.3 – already running fine here! 🙂
Re: Fx 125.0.3
2024年4月30日 1:42 PMTeteHello Uli-san, you are welcome! 😉
Re: Fx 125.0.3
2024年5月5日 8:19 PMUliHello Tete-san,
i just found your new test build and only want to give positive feedback!
It runs fine on Core2Quad Q9550s and it feels considerably snappier! 🙂Best regards
Re: Fx 125.0
2024年5月6日 11:23 AMTeteHello Uli-san,
thank you for your feedback on new build! 🙂
I am considering creating a new test version soon.Re: Fx 125.0
2024年5月6日 10:41 PMUliHello Tete-san,
meanwhile i’m using TEST3 without problems, but it doesn’t feel snappier than the first TEST version.
Maybe, my CPU is too old to recognize any further improvements, but from the standard version there is an impressive improvement.Re: Fx 125.0
2024年5月7日 1:47 PMTeteHello Uli-san,
thank you for the report.In TEST1, by changing the Rust optimization level and the type of LTO to the same settings as the official builds, it seems that the performance has improved. I had overlooked these settings until now.
After comparing the loading times of several sites and the results of Speedmeter 3.0, it appears that TEST2 often produces the best results in my environment.
TEST3 showed a slight decrease in performance compared to TEST2. For non-PGO builds, raising the threshold for inlining seemed to improve performance, but applying PGO seems to turn out to be counterproductive.Re: Fx 125.0
2024年5月8日 5:54 AMUliHello Tete-san,
i didn’t run any artificial speed tests and from the “feeling” of the snappiness, i would say TEST1 and TEST2 are nearly equal and as i already mentioned, they both feel easily snappier than the standard build.
Thanks for these optimizing experiments! 🙂
Re: Fx 125.0
2024年5月9日 3:14 PMTeteHello Uli-san,
thank you for writing your impressions of using my builds. 🙂