Re: [PATCH 13/15] x86: add cmpxchg_flag() variant

[H. Peter Anvin]

On 08/24/2011 01:38 PM, Jeremy Fitzhardinge wrote:
> On 08/24/2011 01:31 PM, H. Peter Anvin wrote:
>> The sete is actually more expensive than the compare for the single-word
>> case.  The double-word case is a different matter.
> 
> Hm, setcc has a longer latency but a higher throughput, is that
> correct?  Still, its recommended where you can use it to avoid a branch.

Yes, but you don't know when that is (the compiler does, of course.)

> I was also thinking that using it reduces register pressure, since you
> don't need to keep the "old" value around, so it dies sooner.

... but you burn a register for the intermediate flag value, so you're
just as bad off.

	-hpa
--
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/