[Beowulf] What services do you run on your cluster nodes?

[Perry E. Metzger]

Patrick Geoffray <patrick at myri.com> writes:
> Perry E. Metzger wrote:
>> from processing interrupts, or prevent your OS from properly switching
>> to a high priority process following an interrupt, but SMM will and
>> you can't get rid of it.
>
> You can usually disable SMI, either through the BIOS or directly from
> the chipset.

I haven't noticed the ability to do that in the BIOS.  Clearly some
chipsets could allow you to disable the interrupt since it originates
outside to the processor itself, though it has been some time since
I've looked at doing that.

> However, you will lose most thermal management (including
> over-temperature protection) and some chipsets require SMM to work
> around hardware bugs.

Yes, and yes.

> The 10-20 us you will lose to SMM every couple of minutes is not very
> high on the list for HPC, compared to daemons burning two orders of
> magnitude more cycles.

It is a lot more frequent than that, and a lot longer than that, at
least on most modern systems I've worked with.

As for the daemons, remember that with a proper scheduler, you will
switch straight from an incoming network interrupt to a high priority
process that is expecting the incoming packet, and that even works
correctly on some (but not all) Linux kernels. A user process cannot
take priority over other tasks, at least not without someone being
quite deliberate about it.

-- 
Perry E. Metzger		perry at piermont.com