[Openswan dev] ocf-bench bug & performance :)

[Harald Jenny]

On Mon, Dec 20, 2010 at 09:10:58PM +1000, David McCullough wrote:
> 
> Jivin Paul Wouters lays it down ...
> > 
> > I was benchmarking OCF and my machine was too fast for ocf-bench:
> > 
> > [  170.239129] OCF: 1040 requests of 1488 bytes in 0 jiffies (0.000 Mbps)
> > 
> > The bench mark is wrong for anything happening in under a jiffy :)
> 
> Now I just need you to send me a system so that I can reproduce that :-) :-)

*ggg*

> 
> > This was on an 8 core:
> > 
> > processor	: 7
> > vendor_id	: GenuineIntel
> > cpu family	: 6
> > model		: 23
> > model name	: Intel(R) Xeon(R) CPU           E5410  @ 2.33GHz
> > stepping	: 6
> > cpu MHz		: 2333.379
> > 
> > When I used 10000 requests instead of 1000 it worked better:
> > 
> > [  229.023024] OCF: 10037 requests of 1488 bytes in 5 jiffies (2389.608 Mbps)
> > 
> > Perhaps make 10000 the new default if request_num= is not specified? Though
> > even that "5" seems dangerously low. Running with request_num=100000 I got:
> > 
> > [  460.738971] OCF: 100035 requests of 1488 bytes in 49 jiffies (2430.238 Mbps)
> 
> Neat,  thats 4 times what I see on my core2 duo, just what the doctor
> ordered.

*smile* very impressive indeed.

> 
> Might be better to rework that code to run for a minimum number of jiffies ;-)

Well with these numbers maybe...

> 
> Cheers,
> Davidm

Kind regards
Harald

> 
> -- 
> David McCullough,      david_mccullough at mcafee.com,  Ph:+61 734352815
> McAfee - SnapGear      http://www.mcafee.com         http://www.uCdot.org
> _______________________________________________
> Dev mailing list
> Dev at openswan.org
> http://lists.openswan.org/mailman/listinfo/dev