I am not an expert on this, but this may help in the why?
The primary advantage to using more bits is in gaining addressability for more memory. Most UNIX operating systems were 32 bit systems. allowing memory up to 4 GB to be adressed. Performance of very large applications can be improved by having even more bits addressable. When you go to 64 bit addressability, the theoretical limit on your memory goes up to 16 million TB
Lets assume you are a manufacturing firm and you have a large database of say 30 GB...If you have a 16 GB main memory, you can load very large segments of your database into real memory, increasing the likelihood of the data you require being in main memory. You avoid a performance hit of going to disks for information, swapping or paging this data... REMEMBER though you need a 64 bit database, OS and hardware to allow this...
However, even if an individual database is not large, performance advantages can still be realized by having a 64 bit operation system and larger memory availalbe. SEveral applications and many processes running concurrently on your system can be achieved...Greater amounts of total memory will reduce memory contention between processes that are running.
Aix 4.33 will run all existing 32 bit and 64 bit apps ... If you have 32 bit apps it may be recommended that you recompile them to 64 to gain further performance...But this is only if your app is under address space constraint that it is required... And to really get the benefit, as I said you need a 64bit database, OS, and hardware to accomodate this.
Hope this helps?
#-)
