Re: Parallel Seq Scan vs kernel read ahead

On Fri, 19 Jun 2020 at 14:10, David Rowley <dgrowleyml(at)gmail(dot)com> wrote:
> Here's a patch which caps the maximum chunk size to 131072. If
> someone doubles the page size then that'll be 2GB instead of 1GB. I'm
> not personally worried about that.
>
> I tested the performance on a Windows 10 laptop using the test case from [1]

I also tested this an AMD machine running Ubuntu 20.04 on kernel
version 5.4.0-37. I used the same 100GB table I mentioned in [1], but
with the query "select * from t where a < 0;", which saves having to
do any aggregate work.

There seems to be quite a big win with Linux too. See the attached
graphs. Both graphs are based on the same results, just the MB/sec
one takes the query time in milliseconds and converts that into MB/sec
for the 100 GB table. i.e. 100*1024/(<milliseconds> /1000)

The machine is a 64core / 128 thread AMD machine (3990x) with a 1TB
Samsung 970 Pro evo plus SSD, 64GB RAM

> [1] https://www.postgresql.org/message-id/CAApHDvrfJfYH51_WY-iQqPw8yGR4fDoTxAQKqn%2BSa7NTKEVWtg%40mail.gmail.com