Re: 9.4 regression

Thom Brown <thom@linux.com> wrote:

> pgbench -j 80 -c 80 -T 3600
>
> 269e78: 606.268013
> 8800d8: 779.583129

As another data point I duplicated Thom's original tests:

max_connections = 500
shared_buffers = 4GB
effective_cache_size = 12GB
random_page_cost = 2.0
cpu_tuple_cost = 0.03
wal_buffers = 32MB
work_mem = 100MB
maintenance_work_mem = 512MB
checkpoint_segments = 32
checkpoint_timeout = 15min
checkpoint_completion_target = 0.8
commit_delay = 50
commit_siblings = 15

pgbench -i -s 20 pgbench
pgbench -j 80 -c 80 -T 1800 pgbench
(note the high contention pgbench setup)

On this system:

Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz
  (hyperthreading enabled)
2 x 15k RPS drives in software RAID 1, ext4, write cache disabled
3.5.0-37-generic #58-Ubuntu SMP Mon Jul 8 22:07:55 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux

commit 8800d8061dd151d6556f5f8d58f8211fd830169f
number of transactions actually processed: 1298864
tps = 721.384612 (including connections establishing)
tps = 721.388938 (excluding connections establishing)

commit 269e780822abb2e44189afaccd6b0ee7aefa7ddd
number of transactions actually processed: 875900
tps = 486.524741 (including connections establishing)
tps = 486.527891 (excluding connections establishing)

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company