Live data from Hacker News

PG-Strom: SSD-To-GPU Direct SQL Execution

github.com

11–20 of 38 posts

Re: PG-Strom: SSD-To-GPU Direct SQL Execution

#11
post #4

It's great to see this type of evolution in database performance, maybe this will become the standard someday. The only thing missing is that there isn't a comparable project for MySQL.

Not trying to be a snoot here, but what is the point of using MySQL today? I can't think of something that I'm aware it does better than Postgres.

Re: PG-Strom: SSD-To-GPU Direct SQL Execution

#12
post #4

It's great to see this type of evolution in database performance, maybe this will become the standard someday. The only thing missing is that there isn't a comparable project for MySQL.

Not trying to be a snoot here, but what is the point of using MySQL today? I can't think of something that I'm aware it does better than Postgres.

Interoperating with code and data files that were made for/by MySQL.

Re: PG-Strom: SSD-To-GPU Direct SQL Execution

#13

Earlier quoted context omitted.

Not trying to be a snoot here, but what is the point of using MySQL today? I can't think of something that I'm aware it does better than Postgres.

Interoperating with code and data files that were made for/by MySQL.

Code I can understand, data a bit less. I'm just thinking that no MySQL program which has already been written is going to have any reliance on the performance characteristics of this flash-based database system, since it wasn't a thing when they were written.

If you could change the shape of your tables to suit this approach, then surely you could adopt Postgres.

Re: PG-Strom: SSD-To-GPU Direct SQL Execution

#14

Wondering how much the "moving data to the GPU" overhead won't help to make things significantly faster

Really depends on the shape of the data and query at hand. If you're doing aggregate computations over million/billions of scalar-sized rows, it could be quite fast. However joining two cold tables will still be more limited by schema and disk read than anything.

Re: PG-Strom: SSD-To-GPU Direct SQL Execution

#17
post #3
post #2

Slides : https://www.slideshare.net/kaigai/20170602ossummitanintellig...

Really loved the slides. I'm in no position to need PG-Strom at this point, but it's very interesting work and I hope it keeps progressing.

Interesting work indeed. It have the potential to beat Oracle Exadata.

Re: PG-Strom: SSD-To-GPU Direct SQL Execution

#18

Wondering how much the "moving data to the GPU" overhead won't help to make things significantly faster

This was posted on an old HN -http://kaigai.hatenablog.com/entry/2016/09/08/003556

The major performance gain is doing SSD-to-GPU direct memory access bypassing the need for data to needlessly transfer over PCI-Express twice.

Post reply on HN