Re: Clock sweep not caching enough B-Tree leaf pages?

On Fri, Apr 25, 2014 at 10:45 AM, Peter Geoghegan <pg(at)heroku(dot)com> wrote:
> I've now done a non-limited comparative benchmark of master against
> the patch (once again, with usage_count starting at 6, and
> BM_MAX_USAGE_COUNT at 30) with a Gaussian distribution. Once again,
> the distribution threshold used was consistently 5.0, causing the
> patched pgbench to report for each test:
>
> transaction type: Custom query
> scaling factor: 5000
> standard deviation threshold: 5.00000
> access probability of top 20%, 10% and 5% records: 0.68269 0.38293 0.19741
>
> Results are available from:
>
> http://postgres-benchmarks.s3-website-us-east-1.amazonaws.com/3-sec-delay-gauss/

I updated this with various changes in bgwriter configuration. Perhaps
unsurprisingly, disabling the background writer entirely helps for
both master and patched. It is perhaps notable that the largest
difference between two comparable patch + master test runs is seen
when the background writer is disabled entirely, and 32 clients.

--
Peter Geoghegan