LP64 application performance and program size

You can use the 64-bit address space to dramatically improve the performance of applications that manipulate large amounts of data, whether the data is be created within the application or obtained from files. Generally, the performance gain comes from the fact that the 64-bit application can contain the data in its 64-bit address space (either created in data structures or mapped into memory), when it would not have fit into a 32-bit address space. The data would need to be multiple GBs in size or larger to show this benefit.

If the same source code is used to create a 32-bit and a 64-bit application, the 64-bit application will typically be larger than the 32-bit application. The 64-bit application is unlikely to run faster than the 32-bit application unless it makes use of the larger 64-bit addressability. Because most C programs are pointer-intensive, a 64-bit application can be close to twice as large as a 32–bit application, depending on how many global pointers and longs are declared. A 64-bit C++ program uses almost twice the data as a 32-bit C++ program, due to the large number of pointers the compiler uses to implement virtual function tables, objects, templates, and so on. That is why the appropriate choice is to create a 32-bit application, unless 64-bit addressability is required by the application or can be used to dramatically improve its performance.