***SPAM*** Re: [ofa-general] SRP/mlx4 interrupts throttling performance
worleys at gmail.com
Wed Oct 1 09:17:08 PDT 2008
On Wed, Oct 1, 2008 at 8:39 AM, Cameron Harr <cameron at harr.org> wrote:
> Cameron Harr wrote:
>> My target server (with DAS) contains 8 2.8 GHz CPU cores and can sustain
>> over 200K IOPs locally, but only around 73K IOPs over SRP. Looking at
>> /proc/interrupts, I see that the mlx_core (comp) device is pushing about
>> 135K Int/s on 1 of 2 CPUs. All CPUs are enabled for that PCI-E slot, but it
>> only ever uses 2 of the CPUs, and only 1 at a time. None of the other CPUs
>> has an interrupt rate more than about 40-50K/s.
I don't know iSCSI, but in MPI-land we just poll... assuming you don't
care if you loose a CPU to polling.
Is there a way to get SRP off it's dependence on Interrupts, and just poll?
More information about the general