AFAIK it's mainly that 64-bit aupports heaps and gobs more memory. If you are getting close to the limits of a 32-bit address space (a mere(!) 4 GB), then bigger addresses help. I don't think Linux is as much of a memory hog as Windows, but with multitasking, program bloat, huge data files, and dropping memory prices you'll use it at some point. I don't think 64 bit, by itself, brings higher speed as sometimes claimed. Floating point and most data paths are already 64 bits or more, this change only concerns integer word sizes. Little integer arithmetic needs more than 32 bits, other than RSA encryptions and address calcs.
The disadvantages would mainly be cost and possible incompatibility, but backward compatibility seems good. Even if your apps are 32-bit, you could run more instances at once without the slowness of virtual memory once you have enough real memory.
|
