Re: [PATCH 0/6] Per-processor private data areas for i386

[Jeremy Fitzhardinge <jeremy@goop.org>]

Pavel Machek wrote:
> So we have 4% slowdown...
>   

Yes, that would be the worst-case slowdown in the hot-cache case.  
Rearranging the layout of the GDT would remove any theoretical 
cold-cache slowdown (I haven't measured if there's any impact in practice).

> ...and 0.2% smaller kernel. I guess you should demonstrate speedup at
> complex syscalls before wedecide it is worth it...

That would be nice, but this patch series isn't really intended to be a 
performance improvement.  That would be nice, but the main motivation is 
to make inline assembler patching for the paravirt work cleaner.

Rusty and I have also been investigating how to use the %gs-based memory 
to implement all percpu data, rather than the few special cases this 
patch series currently covers, which will help further amortize the 
entry/exit cost.

Rusty has also done more comprehensive benchmarks with his variant of 
this patch series, and found no statistically interesting performance 
difference.  Which is pretty much what I would expect, since it doesn't 
increase cache-misses at all.

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