[Asterisk-Users] zttest - 100% ?

DRi at b-w-computer.de DRi at b-w-computer.de
Fri Sep 30 00:02:01 MST 2005


just as an (bad) example:
        we are using an x206 and couldn't get the zttest above 99.975 
equal what we were doing
        single irq, w/o acpi, w/o apic, different kernels, w/o 
hyperthreading, different slots, a.s.o.
        for an Digium wildcard TE110P 

so if someone got such a board to zttest 100% maybe could give some 
information if where's something
special to run asterisk on such boards...
otherwise I think there are production differences on the ibm-mainboards 
or the used chipsets

we'll change hardware next...

asterisk-users-bounces at lists.digium.com wrote on 29.09.2005 18:35:03:

> This might seem a silly question but, what is the true meaning of the 
numbers zttest spits out?

> On 9/29/05, Marco Supino < marco at sd.com> wrote:
> Hi,
> 
> My TDM is on its own IRQ, and the x306 has only one full-size PCI slot.. 

> so no playing with it,
> 
> what results do you get from zttest ? what IRQ is the card on ?
> 
> Marco.
> 
> 
> Damian Funnell wrote:
> > Have you checked that the TDM400P isn't sharing an IRQ with anything 
> > else?  Don't trust /proc/interrupts - run lspci -v to confirm this.
> >
> > We have * running on an x206 and found that the only way to stop the
> > TDP400P sharing an IRQ with other devices was to juggle cards between 
> > slots.
> >
> > Hope this helps!
> > Damian.
> >
> >
> > Marco Supino wrote:
> >
> >> Hi,
> >>
> >> I would like to know what type of configuration could get me closer 
to 
> >> 100% hits in zttest, when testing a TDM400P with 4 FXO ports,
> >>
> >> I am currently running kernel 2.4.31, on a IBM Xseries 306, with 3gh
> >> CPU, HT is disabled, PCI latency was changed, i still cant get more 
> >> then 99.975% in the zttest testings,
> >>
> >> Thanks for any info.
> >>
> >> Marco.
> >>




More information about the asterisk-users mailing list