64-bit architecture is not something new, but there are still a lot of 32-bit computers on the market. While high-end servers almost always have 64-bit Xeon, AMD64, Itanium, or PowerPC processors, when starting out, you may be faced with the choice between 32-bit or 64-bit. Which should you choose?
Before you decide, you should first know what the difference is and how it interacts with your operating system. A 32-bit processor and a 64-bit processor can have the same clock speed, cache, etc., but the 64-bit processor will have a data bus that is 64 bits wide, while the 32-bit processor’s is only 32 bits wide. Generally, this should mean faster processing and more power.
In practice, most users may not notice much of a difference in desktop computers and low-traffic servers. When a server gets into more complex computing, there can be a significant increase in performance, especially if the server has 4GB or more RAM.
The other issue with 64-bit hardware is having the software to support it. Most Linux distributions have supported 64-bit processors for a long time, so it should not be a problem. An x86_64 processor (64-bit) can run x86 (32-bit) software because of the way the architecture is built on x86 technology. For servers, however, all of the software you run will be compiled for X86_64.
The bottom line is that you will want to evaluate overall performance when looking for a server. Processor architecture is only one area to consider, but it is an important one. If it will not cost you more or much more to run a server with a 64-bit processor, you should definitely choose it over 32-bit.