interrupt routing problem?

EM1897 at aol.com EM1897 at aol.com
Thu Feb 10 07:06:32 PST 2005


>:I do have a question about my top measurements. Running the same test on 
>:the same MB, In Dragonfly, I see about 8% system and 12% interrupt, but 
>:FreeBSD 4.11 only shows 12%-14% interrrupt and 0% for system. Is freeBSD 
>:not including cycles that Dfly is? Is this going to be like measuring a 
>:beachball
>:with a yardstick?
>
>    That's like comparing apples with oranges.  It sounds about right,
>    I would expect FreeBSD to handle packet routing entirely within the
>    interrupt.  Right now I think Dragonfly is routing through the ethernet
>    stack so it's getting messaged to a thread, and its going to stay that
>    way for a while.  I wouldn't expect a 14% vs 20% difference in load but
>    your earlier saturation tests seem to validate those numbers.
>
>    Optimizing network throughput is not a priority for us right now, I 
>    would be quite happy with a 14% vs 20% difference.  We are more focused
>    on getting the implementation correct first, and optimizing it later.
>

Its not exactly apples and oranges, because you're telling me that the
measurements are accurate, and I really am measuring the same thing.
I was hoping that FreeBSD wasn't counting something that you are.
 
This is particularly bad news, as my reason for looking at Dragonfly was
because FreeBSD 5.3 is 15% slower than 4.x, and 4.x likely won't
work on the "next" chipset/platform that comes out. But I guess I'm stuck
with 4.11 for the next year or so.
 
I like the direction you're going, but unfortunately you're apparently too
far away for me to justify spending more of my company's time on it.
I know its difficult to know, but how far away, in terms of months, do
you estimate you are from having Dfly to a point where it can be
considered as a performance rival to FreeBSD 4.11 for a commercial
networking product?





More information about the Users mailing list