[Beowulf] Bernhard Kuhn's real time interrupt patch

Andrew Piskorski atp at piskorski.com
Fri Mar 25 15:52:38 PST 2005


On Fri, Mar 25, 2005 at 04:10:15PM -0500, Patrick Geoffray wrote:

> I don't think this patch improves the interrupt latency or the interrupt 
> overhead when there is no contention. Context switching is the metric 
> you want to improve.
>
> With OS-bypass NICs, interrupts are rare and there is usually no contention.

Contention for what, the NIC?

I was thinking that Kuhn looked at interrupt latency under heavy CPU
usage, but I guess I was confused.  What he actually measured was
greatly tightening up the distribtuion of interrupt latency under
varying network (ping), disk (looping on 'find /'), and "graphics"
(glxgears) loads.  I don't understand the ramifications of that.

So is it true that kernel interrupt latency only matters for
non-OS-bypass NICs?  (E.g., your typical ethernet card.)  And stuff
like Myrinet and SCI doesn't care?

-- 
Andrew Piskorski <atp at piskorski.com>
http://www.piskorski.com/



More information about the Beowulf mailing list