Re: performance results on IBM POWER7

On Thu, Mar 1, 2012 at 11:38 AM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> On Thu, Mar 1, 2012 at 11:23 AM, Ants Aasma <ants(dot)aasma(at)eesti(dot)ee> wrote:
>> On Thu, Mar 1, 2012 at 4:54 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>>> ... After that I think maybe some testing of the
>>> remaining CommitFest patches might be in order (though personally I'd
>>> like to wrap this CommitFest up fairly soon) to see if any of those
>>> improve things.
>>
>> Besides performance testing, could you check how clocksources behave
>> on this kind of machine?
>> You can find pg_test_timing tool attached here:
>> http://archives.postgresql.org/pgsql-hackers/2012-01/msg00937.php
>>
>> To see which clocksources are available, you can do:
>> # cat /sys/devices/system/clocksource/clocksource0/available_clocksource
>> To switch the clocksource, just write the desired clocksource like this:
>> # echo hpet > /sys/devices/system/clocksource/clocksource0/current_clocksource
>
> Sure, I'll check that as soon as it's back up.

It seems that "timebase" is the only available clock source.
pg_test_timing says:

Testing timing overhead for 3 seconds.
Per timing duration including loop overhead: 38.47 ns
Histogram of timing durations:
< usec: count percent
32: 6 0.00001%
16: 4 0.00001%
8: 8 0.00001%
4: 282 0.00036%
2: 2999189 3.84628%
1: 74976816 96.15333%

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company