Re: AIX Large Pages and SMT4

From: Neil Chandler <neil_chandler_at_hotmail.com>
Date: Fri, 1 Nov 2019 16:32:10 +0000
Message-ID: <DB7PR10MB209038C0D8AF24B04A36386285620_at_DB7PR10MB2090.EURPRD10.PROD.OUTLOOK.COM>



You shouldn't really get any swapping unless you're overcommitting memory.

parallel_threads_per_cpu to 1 is always a good idea - but especially on SMT4 if you're using parallel. Nothing quite like watching an 8 core server spin up 8 x 4 x 2 x 2 parallel slaves and grind itself into the dust. Had to show a client not so long ago that parallel(16) was way faster then parallel(64) on their small LPAR. I usually recommend using Resource Manager if a DB is doing any parallel processing to stop runaway parallel queries taking the whole server.

Neil.



From: Shane Borden <sborden76_at_yahoo.com> Sent: 24 October 2019 13:30
To: neil_chandler_at_hotmail.com <neil_chandler_at_hotmail.com> Cc: Oracle-L_at_freelists.org <Oracle-L_at_freelists.org> Subject: Re: AIX Large Pages and SMT4

We utilized 64MB pages on a P7 and saw a lot less swapping over not having the pages configured. In addition with SMT4, we changed the parallel_threads_per_cpu to 1 per IBM recommendations and our cpu waits were reduced as well.

I am about to run a SLOB test on a P9 to see what happens in that architecture.

Thanks,

Shane Borden

On Oct 24, 2019, at 4:48 AM, Neil Chandler <neil_chandler_at_hotmail.com<mailto:neil_chandler_at_hotmail.com>> wrote:

Hi,

A couple of Oracle (11.2.0.4 and 19) on AIX (7.2) questions:

Does anyone have any input on whether to use medium pages or large pages on AIX in relation to performance? I'm primarily concerned about performance - especially when the server is stressed.

Does anyone have any input about using SMT2/4/8 usage - my slob testing has seen slightly better throughput with SMT4 over SMT2 (SMT4 vs SMT8 was the same) but my testing has been limited to this point and is far from conclusive.

Neil Chandler.

--
http://www.freelists.org/webpage/oracle-l
Received on Fri Nov 01 2019 - 17:32:10 CET

Original text of this message