Originally posted by: Nothinman
Exactly. I don't see how x32 and x64 are different except that x64 is a bigger number and except that x64 allows for (more easily) running more memory. Unless you've got 3.5GB of RAM (or more) I don't (currently) see a point to x64.
The main difference is the amount of VM each process can address but there are other differences like in long mode there are 8 more GPRs and 8 more SSE registers available which can make a noticable difference depending on the type of work the process is doing.
But when we look at performance differences (say, at Toms Hardware or similar sites) we see almost no difference...
Sure --- if you actually have that memory over and above 3.5GB (or so). If you just have 2.0GB in both machines, both computers (x32 and x64) will handle that memory in identical fashion - no difference.
Not totally correct. On Windows each process is limited to 2G of VM (the other 2G is reserved for the kernel) and that includes memory taken up by the process itself, mapped libraries, mmap()'d files, etc so you end up running out of VM around 1.5-1.8G of VM usually. On a 64-bit system each 32-bit process is allowed up to 4G of VM since 32-bit processes can't come close to touching the amount of VM available there's no need to steal any from them for the kernel.
Correct. I should have mentioned this. However, this is something most people simply won't see. Also, with 2G of memory, it's just not significant.
Boot up times shouldn't change at all. If anything, X32 should be marginally faster since no compatibility layer needs to load, but, regardless, they shouldn't be significantly different.
The compatibility layer in XP64 is virtually nonexistant since the differences between the two operating modes is so small. And the same compatibility layer exists in XP32, just there it lets you run 16-bit apps.
Summary: I just don't see a significant difference. Fair to say, for most applications?