* incoming packet latency in 2.4.[18-20]
@ 2003-10-15 15:45 Chris Friesen
2003-10-15 16:15 ` Matt H.
2003-10-15 16:16 ` Stephan von Krawczynski
0 siblings, 2 replies; 3+ messages in thread
From: Chris Friesen @ 2003-10-15 15:45 UTC (permalink / raw)
To: Linux Kernel
There is an issue with incoming packet latency in the kernels mentioned.
It seems that if you send in a burst of messages, the amount of time it
takes to wake the listening process is dependent on the size of the
message burst. 2.4.18-2.4.20 all show this behaviour, 2.6 doesn't.
Some numbers for a udp message size of 2 bytes:
1 packet, average latency 12 usecs
10 packets, average latency 66 usecs
100 packets, average latency 477 usecs
Is this a known issue? Is there an easy way to fix this, or is it
something inherent in the 2.4 architecture?
Thanks,
Chris
--
Chris Friesen | MailStop: 043/33/F10
Nortel Networks | work: (613) 765-0557
3500 Carling Avenue | fax: (613) 765-2986
Nepean, ON K2H 8E9 Canada | email: cfriesen@nortelnetworks.com
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: incoming packet latency in 2.4.[18-20]
2003-10-15 15:45 incoming packet latency in 2.4.[18-20] Chris Friesen
@ 2003-10-15 16:15 ` Matt H.
2003-10-15 16:16 ` Stephan von Krawczynski
1 sibling, 0 replies; 3+ messages in thread
From: Matt H. @ 2003-10-15 16:15 UTC (permalink / raw)
To: Linux Kernel
Have you tried a more recent kernel?
Matt H.
On Wednesday 15 October 2003 08:45 am, Chris Friesen wrote:
> There is an issue with incoming packet latency in the kernels mentioned.
>
> It seems that if you send in a burst of messages, the amount of time it
> takes to wake the listening process is dependent on the size of the
> message burst. 2.4.18-2.4.20 all show this behaviour, 2.6 doesn't.
>
> Some numbers for a udp message size of 2 bytes:
>
> 1 packet, average latency 12 usecs
> 10 packets, average latency 66 usecs
> 100 packets, average latency 477 usecs
>
> Is this a known issue? Is there an easy way to fix this, or is it
> something inherent in the 2.4 architecture?
>
> Thanks,
>
> Chris
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: incoming packet latency in 2.4.[18-20]
2003-10-15 15:45 incoming packet latency in 2.4.[18-20] Chris Friesen
2003-10-15 16:15 ` Matt H.
@ 2003-10-15 16:16 ` Stephan von Krawczynski
1 sibling, 0 replies; 3+ messages in thread
From: Stephan von Krawczynski @ 2003-10-15 16:16 UTC (permalink / raw)
To: Chris Friesen; +Cc: linux-kernel
On Wed, 15 Oct 2003 11:45:52 -0400
Chris Friesen <cfriesen@nortelnetworks.com> wrote:
>
> There is an issue with incoming packet latency in the kernels mentioned.
>
> It seems that if you send in a burst of messages, the amount of time it
> takes to wake the listening process is dependent on the size of the
> message burst. 2.4.18-2.4.20 all show this behaviour, 2.6 doesn't.
>
> Some numbers for a udp message size of 2 bytes:
>
> 1 packet, average latency 12 usecs
> 10 packets, average latency 66 usecs
> 100 packets, average latency 477 usecs
>
> Is this a known issue? Is there an easy way to fix this, or is it
> something inherent in the 2.4 architecture?
Can you verify these numbers with 2.4.22 and 2.4.23-pre7 ?
Regards,
Stephan
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2003-10-15 16:16 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2003-10-15 15:45 incoming packet latency in 2.4.[18-20] Chris Friesen
2003-10-15 16:15 ` Matt H.
2003-10-15 16:16 ` Stephan von Krawczynski
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.