HybridVerifier: A Cross-Platform Verification Framework for Instruction Set Simulators

Address-encoded byte order

“…Again, when they are part of the left operand of an assignment, it is possible to write in an alien byte order. For example, suppose A is an one-dimensional array, B is a two-dimensional array, and the value 3*B [2][3] must be assigned to A [5]. If both arrays are stored using the native byte order then the following code should be used:…”

“…If B is stored in little-endian and A must be stored in big-endian, then the code could be: bea(A) [5]=3*lea(B) [2][3]; Again, the endianness conversion introduces no penalty. This can easily be extended to include any mixed-endian byte order.…”

“…Most architectures use one of the previous orders. For example, a picoJava processor uses the little-endian byte order [4], while the SPARC V8 architecture uses the big-endian byte order [5]. Several architectures, called bi-endian [6], use both.…”