Re: Large inlines in include/linux/skbuff.h

[Eric Dumazet]

James Morris wrote:

> On Thu, 22 Apr 2004, Andi Kleen wrote:
>
>  
>
> > > How will these changes impact performance?  I asked this last time you 
> > > posted about inlines and didn't see any response.
> > >      
> > I don't think it will be an issue. The optimization guidelines
> > of AMD and Intel recommend to move functions that generate 
> > more than 30-40 instructions out of line. 100 instructions 
> > is certainly enough to amortize the call overhead, and you 
> > safe some icache too so it may be even faster.
> >    
>
> Of course, but it would be good to see some measurements.
>
>
> - James
>  
It depends a lot of the workload of the machine.

If this a specialized machine, with a small program that mostly uses 
recv() & send() syscalls, then, inlining functions is a gain, since 
icache may have a 100% hit ratio. Optimization guidelines are good for 
the common cases, not every cases.

Eric Dumazet

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/