Exploiting New Processors

Although 64-bit processors can run existing 32-bit applications, only native 64-bit applications can exploit the true power of the processor. In addition to the new 64-bit processors, new multicore processors, which put two or more processors into a single chip or die, may also require applications to be rewritten in order to fully exploit the power of the processors.

Exploiting 64 Bits

Microsoft says porting 32-bit applications to 64-bit Windows will be easier than porting 16-bit applications to 32-bit Windows. One of the biggest issues has to do with how developers store addresses. On 32-bit Windows, the size of a memory address, or pointer, and the default size of an integer are both 32 bits. However, on 64-bit Windows, a pointer becomes 64-bits wide while default integers remain 32-bits wide. The change in sizes is most likely to affect applications written in the C programming language because C permits developers to intermix pointers and integers with relative ease.

To gauge how much work will be involved, developers need to identify problem 32-bit code by compiling it with the 64-bit compiler, identifying all shared components that the application relies on, and finally identifying any legacy 16-bit or assembly code. It may be necessary to find substitutes for some technology, such as the .NET Framework (which won’t support 64 bits until the second half of 2005).