J?rgen Keil wrote:
>
> Inlining these functions has the negative side effect that you
> can't interpose on these function calls any more.
>
> And the highly optimized platform specific implementations from
> libc_psr.so.1 are bypassed...
>
>
> So, does avoiding the function call and inlining these functions have
> a significant performance advantage?
>
Depends on the size and alignment of the source/destintation.
Lots of calls moving small chunks will not get VIS acceleration,
so you pay the overhead of making the call and end up falling
into a crude byte copy anyway.  The VIS optimization on SPARC
requires large chunks (384 bytes after alignment last time I checked).

I have a private build of Nevada handy if there is a specific test
you want me to run.


Reply via email to