Re: Add spin_delay() implementation for Arm in s_lock.h

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: 2022-01-13 15:35:12
Message-ID: 5842706F-9E7C-480B-BDD4-EF79BF578399@amazon.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

As promised, here is the remaining data:

1 worker, w/o patch: 5236 ms +/- 252ms
1 worker, w/ patch: 5529 ms +/- 168ms

2 worker, w/o patch: 4917 ms +/- 180ms
2 worker, w/ patch: 4745 ms +/- 169ms

4 worker, w/o patch: 6564 ms +/- 336ms
4 worker, w/ patch: 6105 ms +/- 177ms

8 worker, w/o patch: 9575 ms +/- 2375ms
8 worker, w/ patch: 8115 ms +/- 391ms

16 worker, w/o patch: 19367 ms +/- 3543ms
16 worker, w/ patch: 18004 ms +/- 3701ms

32 worker, w/o patch: 101509 ms +/- 22651ms
32 worker, w/ patch: 104234 ms +/- 26821ms

48 worker, w/o patch: 243329 ms +/- 70037ms
48 worker, w/ patch: 189965 ms +/- 79459ms

64 worker, w/o patch: 552443 ms +/- 22841ms
64 worker, w/ patch: 502727 ms +/- 45253ms

From this data, on average the patch is beneficial at high worker (CPU) counts tested: 48, 63. At 32 and below the performance is relatively close to each other.

Thanks,
Geoff

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2022-01-13 15:39:20 Re: tab completion of enum values is broken
Previous Message Floris Van Nee 2022-01-13 15:27:08 RE: MDAM techniques and Index Skip Scan patch