Flemming Frandsen wrote:
> Sam Leffler wrote:
>> ath cards do use more cpu cycles in certain cases (e.g. beacon
>> processing) but you can easily get everything in the channel which this
>> person isn't even seeing. All I can say is that there are many people
>> running freebsd and netbsd that get better performance than has been
>> reported here.
>
> Good to hear, what are those numbers then, usuing what iperf parameters?
I've already cited numbers. I rarely use iperf because it frequently
makes linux-specific assumptions. For example, one version I looked at
assumed write would block when the send q filled so udp blast tests
would terminate prematurely on overflowing the q.
>
>
>> So stop with this FUD. If there are issues it's with the linux code and
>> this is not the proper forum.
>
> Three things dude:
> 1) Calling a valid technical discussion FUD is impolite and wrong.
When someone makes comments like "the Atheros wireless driver relies
heavily on the CPU where the Ethernet does not" w/o an ounce of
justification then it's FUD. I'm beginning to think responding to you
is similarly productive.
> 2) Yelling FUD does not make an issue go away.
> 3) If there are issues with the Linux version of madwifi then help
> illuminate the problem by posting BSD benchmarks and details, rather
> than try to cover up any perceived controversy that threatens your world
> view.
>
I wrote the original madwifi code. I wrote the freebsd ath code. I
wrote the hal code that underlies both code bases. I know how the
hardware and software works. I've profiled all this code and done
extensive performance tuning for currently shipping products that use
this code. You clearly haven't a clue who you're talking to.
As to the linux code base I've already told you that your best bet is to
talk to the madwifi folks because they are presently working on reducing
the cpu consumption.
Sam
_______________________________________________
Soekris-tech mailing list
[email protected]
http://lists.soekris.com/mailman/listinfo/soekris-tech
