Speed improvements in Firefox 3.7a1pre
15 posts
• Page 1 of 1
Man anyone feeling the speed difference in page loads between Firefox 3.5.3 and 3.7a1pre? Digg seems to load just as fast in firefox as in chrome. I am not sure if its because I am using a new profile or because of the tracemonkey improvements in the nightlies. Anyone have benchmarks?
There are a LOT of improvements on Trunk. JavaScript speed is just one of many.
Latest: Firefox/84.0 *ESR/78.6.0 - Mobile/81.1.5 - Thunderbird/78.5.0
Nightly: Nightly/86.0a1 - Mobile/GV 86.0a1 - Daily/85.0a1 Don't see much improvement on JavaScript speed. But it seems that the rendering speed is faster than before.
A few improvements were mentioned in this topic: viewtopic.php?f=23&t=1528565 On the V8 Benchmark I've noticed a 200% to 400% performance increase since the recursion tracing landed on trunk.
There have always been ghosts in the machine... random segments of code that have grouped together to form unexpected protocols. Unanticipated, these free radicals engender questions of free will, creativity, and even the nature of what we might call the soul...
Yeah I've been seeing some increases in dromaeo as well.
Everytime I tell myself I am going to stick to Firefox releases, I always end up shifting back to Minefield again. ![]()
Little idea of how FF scored on V8 Benchmark before. I heard that the score was bad mainly because of recursion tracing. Tried to run V8 benchmark just now. Quite sad to see that FF trunk scores only 398, while Chromium scores 2771 on the same machine. Has the recursion tracing landed for Win? I am getting 538 in the V8 benchmark on a Intel Core 2 Duo 2.53ghz
No, full recursion tracing has not landed yet. Only partial. Recursion tracing it the only thing stopping Firefox from scoring high on the V8 test. Latest: Firefox/84.0 *ESR/78.6.0 - Mobile/81.1.5 - Thunderbird/78.5.0
Nightly: Nightly/86.0a1 - Mobile/GV 86.0a1 - Daily/85.0a1
Not really. The V8 tests were likely designed with V8's features in mind. It was the only JS engine that implemented recursion tracing, and that is what the tests heavily rely on. There have always been ghosts in the machine... random segments of code that have grouped together to form unexpected protocols. Unanticipated, these free radicals engender questions of free will, creativity, and even the nature of what we might call the soul...
I can't go back to using 3.5.3. I've got Minefield installed on all of the machines I might need to browse the web on.
![]()
Unfortunately no, though it would help a lot. It was (and still somewhat is) something needed, but not the only one. A lot of others work has been done (on guards and a few others one; see bug 504587 for details) and some other still need to be done like Polymorphic Inline Caching (bug 518487). Its critical regardless of other tweaks. Recursion tracing is the big win for benchmarketing.
Latest: Firefox/84.0 *ESR/78.6.0 - Mobile/81.1.5 - Thunderbird/78.5.0
Nightly: Nightly/86.0a1 - Mobile/GV 86.0a1 - Daily/85.0a1 Score: 1040 on V8 test.
Intel i7 Core Quad @ 4.25Ghz | Cooler Master Hyper 212 Plus | 12GB PC3-16000 RAM | ASUS HD6970 2GB - 12.3 CATS | ASUS PA246Q IPS 24" LCD 16:10 | OCZ AGILITY-EX SSD SLC | Windows 7 SP1 x64 Ultimate
You must have a beast of a computer. I am only getting 538 on a Intel Core 2 Duo 2.53 ghz. Yes, he does. Poser.
![]() Latest: Firefox/84.0 *ESR/78.6.0 - Mobile/81.1.5 - Thunderbird/78.5.0
Nightly: Nightly/86.0a1 - Mobile/GV 86.0a1 - Daily/85.0a1
15 posts
• Page 1 of 1
Who is onlineUsers browsing this forum: No registered users and 3 guests |
![]() |