Hi David,
On 06/22/2018 02:28 AM, David Rowley wrote:
> I've attached 2 patches:
>
> 0001: implements items 1-6
> 0002: Is not intended for commit. It just a demo of where we could get
> the performance if we were smarter about locking partitions. I've just
> included this to show 0001's worth.
>
I did some tests with a 64 hash partition setup, and see a speedup for
INSERT / UPDATE scenarios.
> I don't want to discuss the locking on this thread. That discussion
> will detract from discussing what I'm proposing here... Which is not
> to change anything relating to locks. I'm still working on that and
> will post elsewhere.
With 0002 INSERTs get close to the TPS of the non-partitioned case.
However, UPDATEs doesn't see the same speedup. But, as you said, a
discussion for another thread.
Best regards,
Jesper