Re: testing ProcArrayLock patches

On Fri, Nov 18, 2011 at 11:26 AM, Kevin Grittner
<Kevin(dot)Grittner(at)wicourts(dot)gov> wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>> Nate Boley's AMD 6128 box (which has 32 cores) and an HP Integrity
>> server (also with 32 cores).
>
>> [clear improvement with flexlock patch]
>
> Hmm.  We have a 32-core Intel box (4 x X7560 @ 2.27GHz) with 256 GB
> RAM.  It's about a week from going into production, at which point
> it will be extremely hard to schedule such tests, but for a few days
> more I've got shots at it.  The flexlock patch doesn't appear to be
> such a clear win here.
>
> I started from Robert's tests, but used these settings so that I
> could go to higher client counts and better test serializable
> transactions.  Everything is fully cached.
>
> max_connections = 200
> max_pred_locks_per_transaction = 256
> shared_buffers = 8GB
> maintenance_work_mem = 1GB
> checkpoint_segments = 30
> checkpoint_timeout = 15min
> checkpoint_completion_target = 0.9
> seq_page_cost = 0.1
> random_page_cost = 0.1
> cpu_tuple_cost = 0.05
> effective_cache_size = 40GB
> default_transaction_isolation = '$iso'

I had a dismaying benchmarking experience recently that involved
settings very similar to the ones you've got there - in particular, I
also had checkpoint_segments set to 30. When I raised it to 300,
performance improved dramatically at 8 clients and above.

Then again, is this a regular pgbench test or is this SELECT-only?
Because the absolute numbers you're posting are vastly higher than
anything I've ever seen on a write test.

Can you by any chance check top or vmstat during the 32-client test
and see what percentage you have of user time/system time/idle time?

What OS are you running?

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