MozillaZine

Is it me or...

Discussion about official Mozilla Firefox builds
Piel

User avatar
 
Posts: 216
Joined: November 11th, 2002, 4:26 pm
Location: Boston, MA

Post Posted November 19th, 2002, 4:07 pm

I don't know if it's because I added a 4MB cache in user.js but today's Phoenix build feels *so* much faster and spritely than even yesterday's. I am running a p4 1.7ghz with 512 MB on WinXP.

Has anyone else experienced this performance improvement?

djst
Moderator

User avatar
 
Posts: 2826
Joined: November 5th, 2002, 1:34 am
Location: Sweden

Post Posted November 19th, 2002, 4:15 pm

Piel wrote:I don't know if it's because I added a 4MB cache in user.js but today's Phoenix build feels *so* much faster and spritely than even yesterday's. I am running a p4 1.7ghz with 512 MB on WinXP.

Has anyone else experienced this performance improvement?


I can confirm that the startup speed has improved *slightly*. It took almost exactly 1.5 seconds to start Phoenix before on my computer, and now it's aproximately 1.2 seconds. :) Something has happened, the file size of the whole build has decreased, but the memory footprint is the same (I'm keeping a log for that).

Blake
 
Posts: 198
Joined: November 4th, 2002, 4:12 pm
Location: Mountain View, CA

Post Posted November 19th, 2002, 6:12 pm

djst wrote:I can confirm that the startup speed has improved *slightly*. It took almost exactly 1.5 seconds to start Phoenix before on my computer, and now it's aproximately 1.2 seconds. :) Something has happened, the file size of the whole build has decreased, but the memory footprint is the same (I'm keeping a log for that).


Slightly? That's 20% faster!

djst
Moderator

User avatar
 
Posts: 2826
Joined: November 5th, 2002, 1:34 am
Location: Sweden

Post Posted November 19th, 2002, 6:20 pm

Blake wrote:
djst wrote:I can confirm that the startup speed has improved *slightly*. It took almost exactly 1.5 seconds to start Phoenix before on my computer, and now it's aproximately 1.2 seconds. :) Something has happened, the file size of the whole build has decreased, but the memory footprint is the same (I'm keeping a log for that).


Slightly? That's 20% faster!


Yeah, it's great! But since it started so quickly earlier on my machine, I'm having trouble measuring the difference. But as I said, it's about 0.3 seconds faster now! Anything changed in the startup code?

Blake
 
Posts: 198
Joined: November 4th, 2002, 4:12 pm
Location: Mountain View, CA

Post Posted November 19th, 2002, 6:24 pm

djst wrote:
Blake wrote:
djst wrote:I can confirm that the startup speed has improved *slightly*. It took almost exactly 1.5 seconds to start Phoenix before on my computer, and now it's aproximately 1.2 seconds. :) Something has happened, the file size of the whole build has decreased, but the memory footprint is the same (I'm keeping a log for that).


Slightly? That's 20% faster!


Yeah, it's great! But since it started so quickly earlier on my machine, I'm having trouble measuring the difference. But as I said, it's about 0.3 seconds faster now! Anything changed in the startup code?


Actually, I'm not sure why it's faster. The only thing I can think of is that some jars were disabled in Phoenix (which is why the build size decreased), and we have to read those in at startup. If that's the cause, that'd be good news, since we still have at least 5 jars to remove.

Piel

User avatar
 
Posts: 216
Joined: November 11th, 2002, 4:26 pm
Location: Boston, MA

Post Posted November 19th, 2002, 8:56 pm

Blake wrote:
djst wrote:
Blake wrote:
djst wrote:I can confirm that the startup speed has improved *slightly*. It took almost exactly 1.5 seconds to start Phoenix before on my computer, and now it's aproximately 1.2 seconds. :) Something has happened, the file size of the whole build has decreased, but the memory footprint is the same (I'm keeping a log for that).


Slightly? That's 20% faster!


Yeah, it's great! But since it started so quickly earlier on my machine, I'm having trouble measuring the difference. But as I said, it's about 0.3 seconds faster now! Anything changed in the startup code?


Actually, I'm not sure why it's faster. The only thing I can think of is that some jars were disabled in Phoenix (which is why the build size decreased), and we have to read those in at startup. If that's the cause, that'd be good news, since we still have at least 5 jars to remove.


What sort of beast are you using that gets you to have Phoenix start in 1.5 seconds? The 1st time I run Phoenix, it takes up to 14 seconds...have you configured it somehow to launch so much faster? I can however launch Phoenix again in about 2 seconds.

mesostinky

User avatar
 
Posts: 215
Joined: November 4th, 2002, 10:44 pm
Location: NJ

Post Posted November 19th, 2002, 11:13 pm

I'm sure he's referring to secondary launches. It takes a long time the first time Phoenix launches for everyone.

BTW in general Phoenix since last week does seem to have sped up.

Piel

User avatar
 
Posts: 216
Joined: November 11th, 2002, 4:26 pm
Location: Boston, MA

Post Posted November 19th, 2002, 11:44 pm

mesostinky wrote:I'm sure he's referring to secondary launches. It takes a long time the first time Phoenix launches for everyone.

BTW in general Phoenix since last week does seem to have sped up.


But why is that? It just seems odd to me that relatively small application would take so muchtime to load. Is it the XUL that is causing the delay? If so, are its cross platform capabilities really worth it if it is going to be such a burden on performance? I ask this last question as a wannabe geek who has never written anything other than HTML and CSS code.

djst
Moderator

User avatar
 
Posts: 2826
Joined: November 5th, 2002, 1:34 am
Location: Sweden

Post Posted November 20th, 2002, 5:25 am

mesostinky wrote:I'm sure he's referring to secondary launches. It takes a long time the first time Phoenix launches for everyone.

BTW in general Phoenix since last week does seem to have sped up.


Yes, I'm talking about second-time launches. The first time is pretty random, depending on what's currently in memory and what's currently in the virtual memory.

But if you close Phoenix completely and then start it again (and repeat that process a few times), you get the same startup speed every time. And that (short) startup time has increased a bit.

Return to Firefox Builds


Who is online

Users browsing this forum: streetwolf and 5 guests