When using memset or memcpy within an Obj-C program, will the compiler optimise the setting (memset) or copying (memcpy) of data into 32-bit writes or will it do it byte by byte?
Memset will come as part of your standard C library so it depends on the implementation you are using. I would guess most implementations will copy in blocks of the native CPU size (32/64 bits) and then the remainder byte-by-byte.
Here is glibc's version of memcpy for an example implementation:
void *
memcpy (dstpp, srcpp, len)
void *dstpp;
const void *srcpp;
size_t len;
{
unsigned long int dstp = (long int) dstpp;
unsigned long int srcp = (long int) srcpp;
/* Copy from the beginning to the end. */
/* If there not too few bytes to copy, use word copy. */
if (len >= OP_T_THRES)
{
/* Copy just a few bytes to make DSTP aligned. */
len -= (-dstp) % OPSIZ;
BYTE_COPY_FWD (dstp, srcp, (-dstp) % OPSIZ);
/* Copy whole pages from SRCP to DSTP by virtual address manipulation,
as much as possible. */
PAGE_COPY_FWD_MAYBE (dstp, srcp, len, len);
/* Copy from SRCP to DSTP taking advantage of the known alignment of
DSTP. Number of bytes remaining is put in the third argument,
i.e. in LEN. This number may vary from machine to machine. */
WORD_COPY_FWD (dstp, srcp, len, len);
/* Fall out and copy the tail. */
}
/* There are just a few bytes to copy. Use byte memory operations. */
BYTE_COPY_FWD (dstp, srcp, len);
return dstpp;
}
So you can see it copies a few bytes first to get aligned, then copies in words, then finally in bytes again. It does some optimized page copying using some kernel operations.
You can see the libc implementations of these methods in the Darwin source. In 10.6.3, memset works at the word level. I didn't check memcpy, but probably it's the same.
You are correct that it's possible for the compiler to do the work inline instead of calling these functions. I suppose I'll let someone who knows better answer what it will do, though I would not expect a problem.