<div dir="ltr">The expensive part is often having to save and restore all the state in registers and other bits on context switch.<div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Jan 5, 2016 at 4:01 PM, Steinar H. Gunderson <span dir="ltr"><<a href="mailto:sgunderson@bigfoot.com" target="_blank">sgunderson@bigfoot.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">On Tue, Jan 05, 2016 at 03:36:10PM -0600, Benjamin Cronce wrote:<br>
> You can't have different virtual memory space and not take some large<br>
> switching overhead without devoting a lot of transistors to massive caches.<br>
> And the larger the caches, the higher the latency.<br>
<br>
</span>I'm sure you already know this, but just to add to what you're saying:<br>
Modern CPUs actually have cache-line tagging tricks so that they don't have<br>
to blow the entire L1 just because you do a context switch. It would be too<br>
expensive.<br>
<span class="im HOEnZb"><br>
/* Steinar */<br>
--<br>
Homepage: <a href="https://www.sesse.net/" rel="noreferrer" target="_blank">https://www.sesse.net/</a><br>
</span><div class="HOEnZb"><div class="h5">_______________________________________________<br>
Bloat mailing list<br>
<a href="mailto:Bloat@lists.bufferbloat.net">Bloat@lists.bufferbloat.net</a><br>
<a href="https://lists.bufferbloat.net/listinfo/bloat" rel="noreferrer" target="_blank">https://lists.bufferbloat.net/listinfo/bloat</a><br>
</div></div></blockquote></div><br></div>