From: | "Blake, Geoff" <blakgeof(at)amazon(dot)com> |
---|---|
To: | Andres Freund <andres(at)anarazel(dot)de>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | "pgsql-hackers(at)lists(dot)postgresql(dot)org" <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
Subject: | Re: Add spin_delay() implementation for Arm in s_lock.h |
Date: | 2022-01-12 18:34:12 |
Message-ID: | 27332B1E-6504-479F-813F-C15B2EBBE718@amazon.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Tom, Andres,
I spun up a 64-core Graviton2 instance (where I reported seeing improvement with this patch) and ran the provided regression test with and without my proposed on top of mainline PG. I ran 4 runs each of 63 workers where we should see the most contention and most impact from the patch. I am reporting the average and standard deviation, the average with the patch is 10% lower latency, but there is overlap in the standard deviation. I'll gather additional data at lower worker counts and post later to see what the trend is.
Cmd: postgres=# SELECT test_shm_mq_pipelined(16384, 'xyzzy', 10000000, workers);
Avg +/- standard dev
63 workers w/o patch: 552443ms +/- 22841ms
63 workers w/ patch: 502727 +/- 45253ms
Best results
w/o patch: 521216ms
w/ patch: 436442ms
Thanks,
Geoff
From | Date | Subject | |
---|---|---|---|
Next Message | Robert Haas | 2022-01-12 18:42:25 | Re: Time to drop plpython2? |
Previous Message | Robert Haas | 2022-01-12 18:20:15 | Re: generic plans and "initial" pruning |