time frame for a better x64 browser?

[stravinsky]

is there any time frame for a native x64 browser that is significantly better than the x86 version?

by significant i mean at least 30% better.
As of now, the x64 has the potential to be faster than the x86 version. any particular time when PM shifts to a recommended x64 version ? or when the performance of x64 is significantly better?

[Moonchild]

Performance is already significantly better as far as large data blocks goes - but benchmarks won't show it because benchmarks don't test on it. See also my explanation of comparing benchmark results in the FAQ post.

You can't expect 30% better performance, ever, of a 64-bit application that does most of its work in 8-bit space (text). So that is never going to happen.

Native 64-bit applications are only really a huge step forward from the user's experience side if what the application does is actually using the 64-bit address space or large data blocks. The only advantage there is for Pale Moon x64 in most of what it does is that it cuts away the need for the WoW64 layer in Windows, which is why you see lower CPU strain on complex pages in it. So basically you are using a more efficient browser, from a machine perspective, but not necessarily from a user perspective because of the tasks it is set to perform.

[stravinsky]

WoW64 is hardly 1%-2% performance loss. probably lesser.

any timeframe for when mozilla source code will get multiprocessing?

[Moonchild]

WoW64 is hardly 1%-2% performance loss. probably lesser.

According to whom?

Any timeframe for when mozilla source code will get multiprocessing?

You're asking the wrong person

[dv8cowboy]

@Moonchild...

So what is your best hunch about the next IMPROVED Pale Moon 64bit,which I love?

Kind regards
-Cowboy

[Ryrynz]

Probably worth asking now since 12.2 has hit..
Since benchmarks aren't really indicative of the TRUE performance we're likely to experience with a browser, do you consider the 64bit version to be a better choice as far as performance goes now?

[lobocursor]

is there any time frame for a native x64 browser that is significantly better than the x86 version?

by significant i mean at least 30% better.
As of now, the x64 has the potential to be faster than the x86 version. any particular time when PM shifts to a recommended x64 version ? or when the performance of x64 is significantly better?

Sometimes I miss the old Silicon Graphics IRIX Os, it was really 64 bits and the old Netscape could handle a heavy use. Actually Windows is a mess for support legacy code from all places, traces of MSDOS, Windows 3.1, Windows 95/98... Windows is a mess inside the core.

Multiprocessing is more than just yank code to several processors, you must "cut" in pieces the process and make sure that all processors do their part and assemble the work again in the correct order. That is a complex trick to do, so why I'm not surprised that Mozilla takes ages for create a version multicore of FireFox.

[Moonchild]

Probably worth asking now since 12.2 has hit..
Since benchmarks aren't really indicative of the TRUE performance we're likely to experience with a browser, do you consider the 64bit version to be a better choice as far as performance goes now?

It depends on your use.

Pale Moon 12.2 x64 should be faster and more efficient than its 32-bit counterpart, BUT it's not necessarily a better choice because of very limited plugin support, 64-bit graphics driver issues that occur on a number of systems, etc.
If you just want raw performance and don't care about the plugin compatibility issues and have modern hardware (including dedicated full feature graphics, ruling out a lot of laptops) with updated and properly supported drivers, then it's a better choice. Otherwise take the cautious approach.

Parallel multi-core support is a long way off. And it seems Mozilla is more interested in "polish" than "mechanics" at the moment, so progress in that respect doesn't seem to be on the menu.

[stravinsky]

And it seems Mozilla is more interested in "polish" than "mechanics" at the moment, so progress in that respect doesn't seem to be on the menu.

+1
from reading a lot of bugzilla bugs and the discussions, i could see that the devs arent really trying for better performance. They are looking for "appearance of speed" , not the actual speed..

for a really good browser, take a look at Opera 12 RC.

Parallel multi-core support is a long way off.

is it even being looked into? i dont think there is any work being done on this at all. they are too busy releasing new versions every 6 weeks.

[stravinsky]

slightly OT, but anyway :

i tried to open 800 tabs together on a number of browsers. the tabs were quite simple, they basically had ~5 hyperlinks and 1 image. ONLY PM was able to open all of them. . I have 4GB ram. i used 32bit versions of all browsers.

1 FF nightly 15 : ate > 4GB RAM . So had to end-task it. would have opened the tabs but used all of RAM first.

2. Chrome Canary 21 : ate >4GB RAM + Complete hang.

3. Opera nightly : Hang.

4. PM 12.1 : ate about 2.5GB RAM but opened all the tabs without missing a single tab

5. IE9 : complete hang + hang-crash-restart-hang cycle.

6. IE8 :


link of the tabs i opened : http://www.redmeat.com/redmeat/2012-05-08/index.html