Re: PGO and Windows

2012-08-03 Thread Neil
Jean-Marc Desperrier wrote: BTW the last time I had Seamonkey use all the 32-bits adress space, the result was *really* bad. A system down to it's knees, and not able to close the Seamonkey process anymore. It's harder to get the same thing with Firefox, the lazy tab opening helps :-) You ca

Re: PGO and Windows

2012-08-02 Thread Jean-Marc Desperrier
Zack Weinberg a écrit : Does MS have any plan to support the x32 memory model? It needs kernel as well as compiler support. Could a shim layer possibly bring x32 to a x64 process under Windows ? Or else have most of the code handle memory as "offset inside a bucket", and reduce most pointer

Re: PGO and Windows

2012-07-18 Thread Henri Sivonen
On Tue, Jul 17, 2012 at 7:24 PM, Jean-Marc Desperrier wrote: > OTOH if there's something like that, the Linux users should have realized it > earlier. Or they're just used to gcc compiled code performing badly ;-) When distros were shipping 64-bit 3.6, it was without a JIT. Many people, including