Re: [Codel] [PATCH v5] pkt_sched: codel: Controlled Delay AQM

[dave taht <dave.taht@gmail.com>]

On 05/05/2012 03:09 PM, Eric Dumazet wrote:
> On Sat, 2012-05-05 at 15:03 -0700, dave taht wrote:
>
>> Maybe on your arch, but highly doubtful on a 680Mhz mips that isn't even
>> superscalar.
>>
> CPU are fast, memory is slow.
>
>> I'd prefer to leave it in and be able to compile it out, and actually
>> measure the difference.
> You optimize the case where there is no need to optimize (small queue)
>
> I can see count bigger than 100000 with 20 concurrent netperf
At what speeds?

Are you testing 10GigE? I'm dying to know what happens there...

If I could encourage you to ratchet down to 2 or 100Mbit
it would be easier to get comparable results.

I like using the ethtool trick for 10 and 100Mbit as that
leaves just bql and the qdisc in the mix.

We've discussed elsewhere some of the issues with htb.

Also I'd had an 'interesting' result at 2Mbit that I haven't had time
to duplicate, which I'm going to do now on mainstream hardware.

> This makes no sense to have a cache so big.
concur, which is why I'd also asked how big it got, and how fast you 
were running

>  Or there is a bug in codel


hmm.

>
>
>