BYTEmark Bug Bitten?

In regard to "BYTEmark Bug Bashed" (December), it appears to me that the bug is with the compiler vendor's malloc() rather than with your code. The C Standard section on memory management functions states that "the pointer returned if the allocation succeeds is suitably aligned so that it may be assigned to a pointer to any type of object and then used to access such an object or an array of such objects in the space allocated (until the space is explicitly freed or reallocated)." If double requires 8-byte alignment to work properly, and double has the strictest alignment requirements, then malloc() should return memory aligned to at least 8-byte boundaries.

Fred J. Tydeman
Member, X3J11 (ANSI C Committee)
tydeman@tybor.com

You may be rig ht. We have received several letters suggesting that the problem lies with the C library's implementation of malloc() , not with our benchmarks; however, we have not yet found any documentation that explicitly states that a malloc() returning data not optimally aligned is in error. It's not quite accurate to say that "double requires 8-byte alignment to work." The algorithm works even if the data is not aligned. It just works more slowly.--Rick Grehan, senior technical editor