Re: RFC: Getting rid of LTR in VMX
Date: Mon Feb 20 2017 - 17:24:02 EST
On February 20, 2017 2:02:53 PM PST, Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote:
>> > Yes. But 150-200 clock cycles are nothing compared to the cache
>> > you get from preemption, so I'd ignore that. Saving 300 clock
>> > userspace exits from TR+GSBASE would be about 5% on my Haswell.
>> That's still 5% :)
>Yes, 5% on userspace exits is good (though they're rare).
>OTOH, 300 clock cycles on preemption are nothing to write home about.
Oh, I would say it is... especially when it's basically free.
Sent from my Android device with K-9 Mail. Please excuse my brevity.