On 5/25/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> -----Original Message----- > From: Randy Barlow [mailto:[EMAIL PROTECTED] <Snip> > If you are using a lot of memory in your computations, then > the 64-bit environment will be much friendlier to you :) > Also, if I understand correctly, you will get higher > precision on floating point calculations (someone correct me > if I am wrong here!) I also believe that the 64 bit > processors are able to perform more instructions per second > on average when executing 64 bit code vs. 32 bit code if I am > not mistaken... I am not sure, but that makes sense. If nothing else, things executed directly usually run more smoothly than those who are run through emulation. 64bit code on 64bit processor good...
There is no emulation involved when running 32bit code in either core2 nor amd64 processors. The difference when running in 32bit mode is that some instructions are unavailable, you have a smaller number of registers, and the registers are only 32bit. More registers speed up most kind of code, 64bit registers speed up 64bit calculations, and the extra instructions are good for array calculations and similar (more sse instructions for example). The only bad thing with 64bit code is that the programs get bigger, which may effect memory performance negatively. But if you have a lot of memory then you will benefit from not needing any special addressing modes. Since you (Denis) are doing a lot of mathematical calculations you will probably benefit from running in 64bit mode. -- [EMAIL PROTECTED] mailing list