Can anyone explain this pgbench results? - Mailing list pgsql-performance

From Joost Kraaijeveld
Subject Can anyone explain this pgbench results?
Date
Msg-id A3D1526C98B7C1409A687E0943EAC410605FC1@obelix.askesis.nl
Whole thread Raw
Responses Re: Can anyone explain this pgbench results?  (Michael Fuhr <mike@fuhr.org>)
Re: Can anyone explain this pgbench results?  ("Jim C. Nasby" <jnasby@pervasive.com>)
List pgsql-performance
Hi,

Below are some results of running pgbench, run on a machine that is doing nothing else than running PostgreSQL woth
pgbench.The strange thing is that the results are *constantly alternating* hight (750-850 transactions)and low (50-80
transactions),no matter how many test I run. If I wait a long time (> 5 minutes) after running the test, I always get a
hightscore, followed by a low one, followed by a high one, low one etc.  

I was expecting a low(ish) score the first run (because the tables are not loaded in the cache yet), followed by
continueshigh(ish) scores, but not an alternating pattern. I also did not expect so much difference, given the hardware
Ihave (Dual Opteron, 4GB memory , 3Ware SATA RAID5 with 5 disks, seerate swap and pg_log disks). 

Anyone any idea?

Results of pgbench:

postgres@panoramix:/usr/lib/postgresql/8.1/bin$ ./pgbench  -c 10 -t 150 test
starting vacuum...end.
transaction type: TPC-B (sort of)
scaling factor: 100
number of clients: 10
number of transactions per client: 150
number of transactions actually processed: 1500/1500
tps = 50.651705 (including connections establishing)
tps = 50.736338 (excluding connections establishing)
postgres@panoramix:/usr/lib/postgresql/8.1/bin$ ./pgbench  -c 10 -t 150 test
starting vacuum...end.
transaction type: TPC-B (sort of)
scaling factor: 100
number of clients: 10
number of transactions per client: 150
number of transactions actually processed: 1500/1500
tps = 816.972995 (including connections establishing)
tps = 836.951755 (excluding connections establishing)
postgres@panoramix:/usr/lib/postgresql/8.1/bin$ ./pgbench  -c 10 -t 150 test
starting vacuum...end.
transaction type: TPC-B (sort of)
scaling factor: 100
number of clients: 10
number of transactions per client: 150
number of transactions actually processed: 1500/1500
tps = 42.924294 (including connections establishing)
tps = 42.986747 (excluding connections establishing)
postgres@panoramix:/usr/lib/postgresql/8.1/bin$ ./pgbench  -c 10 -t 150 test
starting vacuum...end.
transaction type: TPC-B (sort of)
scaling factor: 100
number of clients: 10
number of transactions per client: 150
number of transactions actually processed: 1500/1500
tps = 730.651970 (including connections establishing)
tps = 748.538852 (excluding connections establishing)


TIA

Groeten,

Joost Kraaijeveld
Askesis B.V.
Molukkenstraat 14
6524NB Nijmegen
tel: 024-3888063 / 06-51855277
fax: 024-3608416
e-mail: J.Kraaijeveld@Askesis.nl
web: www.askesis.nl

pgsql-performance by date:

Previous
From: Kevin
Date:
Subject: Re: Postgres on VPS - how much is enough?
Next
From: "Nagita Karunaratne"
Date:
Subject: Re: Postgres on VPS - how much is enough?