Variable sized i386 string operations

Richard Henderson
Thu Jan 27 11:27:00 GMT 2000

On Thu, Jan 27, 2000 at 08:16:51PM +0100, Jan Hubicka wrote:
> Speedups for other pixel sizes are basically by avoiding function call
> overhead in library version ...

How big are these copies?  I would imagine that function call
overhead is only visible in extremely sort copies -- 64 bytes
or less.

> Perhaps we can develop se customizability to this. (switch to enable
> full inline and switch to disable alignment handling for instance)
> and default to something like rep movsl + maximally two branches.

That doesn't sound unreasonable.

> Note that our strlen unroller falls into same problem with the code size.

Yes, I was just thinking that.


More information about the Gcc-patches mailing list