What are the OpenVMS differences between VAX and AXP?
Very few. As of OpenVMS V6.1, the VAX and AXP platforms are very close to “feature parity”. Most applications can just be recompiled and run. Some differences to be aware of: • The default double-precision floating type on OpenVMS AXP is VAX G_float, whereas on VAX it is usually D_float. D_float is available on AXP, but D_float values are converted to G_float for computations and then converted back to D_float when stored. Because the G_float type has three fewer fraction bits than D_float, some applications may get different results. IEEE float types are also available on OpenVMS AXP. • Data alignment is extremely important for best performance on AXP. This means that data items should be allocated at addresses which are exact multiples of their sizes. Quadword alignment will offer the best performance, especially for character values and those smaller than 32 bits. Compilers will naturally align variables where they can and will issue warnings if they detect unaligned data items. • D
