I will try it in new Firefox3.2 nightly build after it downloads.
brb
Name:
Athlon 4000+2009-01-03 10:36
658ms (in x100 more loops)=6.58ms
3.2prealpha is 14 times faster(JIT)!
the code:
function tmp444(){ var d = new Date;
var std = d.getTime();
for(var i=10000000;i>0;i--){
var res = Math.log(i);}
var d2 = new Date;
var std2 = d2.getTime();
var diff = std2 - std;
return diff}
Name:
Anonymous2009-01-03 11:15
How fast would this run in C++ or assembly?
Name:
Anonymous2009-01-03 11:29
>>8
assuming a function call takes 20 CPU cycles
and CPU executes 3billion of ins. per seconds
3e9/20=150000000 loops/S=150000 loops per ms
this is 987000 loops per 6.58ms >>7 is 100000 loops per 6.58ms
9.87 times faster
Name:
Anonymous2009-01-03 11:45
>>7
Are the 3.2 nightlies worth using? I am intrigued by this JIT compilation and would like to try it without getting assraped by alpha-quality software.
Name:
Anonymous2009-01-03 11:49
>>10
according to forums:Addons sometimes crash or not work.
but Firefox 2 crashes every week by itself.
Right now nothing crashed, just few extensions don't work(using nightly tester tools).
Name:
Anonymous2009-01-03 13:38
there is however a maddenign thign which closes tabs if you drag them even 1 mm.
Implementations may choose any appropriate value for each limit, provided it is not more restrictive than the Minimum Acceptable Values Applications should not assume any particular value for a limit.
In Anonix, many of the values in limits.h will be INT_MAX or UINT_MAX. Everything that can vary in size will be dynamically allocated up to available memory or the ulimit for the process, whichever is smaller.