Hi,
I'm still fighting with snapshot_too_old. The feature is just badly
undertested, underdocumented, and there's lots of other oddities. I've
now spent about as much time on that feature than on the whole rest of
the patchset.
As an example for under-documented, here's a definitely non-trivial
On Sun, Mar 29, 2020 at 9:50 PM Andres Freund wrote:
> On March 29, 2020 11:24:32 AM PDT, Alexander Korotkov
> wrote:
> > clearly a big win on majority
> >of workloads, I think we still need to investigate different workloads
> >on different hardware to ensure there is no regression.
>
>
On Sun, Mar 29, 2020 at 11:50:10AM -0700, Andres Freund wrote:
Hi,
On March 29, 2020 11:24:32 AM PDT, Alexander Korotkov
wrote:
clearly a big win on majority
of workloads, I think we still need to investigate different workloads
on different hardware to ensure there is no regression.
Hi,
On March 29, 2020 11:24:32 AM PDT, Alexander Korotkov
wrote:
> clearly a big win on majority
>of workloads, I think we still need to investigate different workloads
>on different hardware to ensure there is no regression.
Definitely. Which workloads are you thinking of? I can think of
On Sun, Mar 29, 2020 at 4:40 AM Peter Geoghegan wrote:
> On Sun, Mar 1, 2020 at 12:36 AM Andres Freund wrote:
> > The workload is a pgbench readonly, with pgbench -M prepared -c $conns
> > -j $conns -S -n for each client count. This is on a machine with 2
> > Intel(R) Xeon(R) Platinum 8168, but
Hi,
On 2020-03-28 18:39:32 -0700, Peter Geoghegan wrote:
> I have heard quite a few complaints about the scalability of snapshot
> acquisition in Postgres. Generally from very large users that are not
> well represented on the mailing lists, for a variety of reasons. The
> GetSnapshotData()
On Sat, Mar 28, 2020 at 06:39:32PM -0700, Peter Geoghegan wrote:
> On Sun, Mar 1, 2020 at 12:36 AM Andres Freund wrote:
> > The workload is a pgbench readonly, with pgbench -M prepared -c $conns
> > -j $conns -S -n for each client count. This is on a machine with 2
> > Intel(R) Xeon(R) Platinum
On Sun, Mar 1, 2020 at 12:36 AM Andres Freund wrote:
> The workload is a pgbench readonly, with pgbench -M prepared -c $conns
> -j $conns -S -n for each client count. This is on a machine with 2
> Intel(R) Xeon(R) Platinum 8168, but virtualized.
>
> conns tps master tps
On Sun, Mar 29, 2020 at 1:49 PM Andres Freund wrote:
> > 13. is -> are
> >
> > * accessed data is stored contiguously in memory in as few cache lines as
>
> Oh? 'data are stored' sounds wrong to me, somehow.
In computer contexts it seems pretty well established that we treat
"data" as an
Hi,
On 2020-03-17 23:59:14 +1300, David Rowley wrote:
> Nice performance gains.
Thanks.
> On Sun, 1 Mar 2020 at 21:36, Andres Freund wrote:
> 2. I don't quite understand your change in
> UpdateSubscriptionRelState(). snap seems unused. Drilling down into
> SearchSysCacheCopy2, in
Hi,
Thanks for looking!
On 2020-03-20 18:23:03 +1300, Thomas Munro wrote:
> On Sun, Mar 1, 2020 at 9:36 PM Andres Freund wrote:
> > I'm still working on cleaning that part of the patch up, I'll post it in
> > a bit.
>
> I looked at that part on your public pgxact-split branch. In that
>
On Sun, Mar 1, 2020 at 9:36 PM Andres Freund wrote:
> conns tps master tps pgxact-split
>
> 1 26842.49284526524.194821
> 10 246923.158682 249224.782661
> 50 695956.539704 709833.746374
> 100 1054727.043139 1903616.306028
Hi,
Nice performance gains.
On Sun, 1 Mar 2020 at 21:36, Andres Freund wrote:
> The series currently consists out of:
>
> 0001-0005: Fixes and assert improvements that are independent of the patch,
> but
>are hit by the new code (see also separate thread).
>
> 0006: Move delayChkpt
On 3/1/20 3:36 AM, Andres Freund wrote:
I think these are pretty nice results.
Indeed they are.
Is the target version PG13 or PG14? It seems like a pretty big patch to
go in the last commitfest for PG13.
Regards,
--
-David
da...@pgmasters.net
Hi,
On 2020-03-01 00:36:01 -0800, Andres Freund wrote:
> Here are some numbers for the submitted patch series. I'd to cull some
> further improvements to make it more manageable, but I think the numbers
> still are quite convincing.
>
> The workload is a pgbench readonly, with pgbench -M
On Sun, Mar 1, 2020 at 2:17 PM Andres Freund wrote:
>
> Hi,
>
> On 2020-03-01 00:36:01 -0800, Andres Freund wrote:
> > conns tps mastertps pgxact-split
> >
> > 1 26842.49284526524.194821
> > 10 246923.158682 249224.782661
> > 50 695956.539704
Hi,
On 2020-03-01 00:36:01 -0800, Andres Freund wrote:
> conns tps mastertps pgxact-split
>
> 1 26842.49284526524.194821
> 10 246923.158682 249224.782661
> 50 695956.539704 709833.746374
> 100 1054727.043139
101 - 117 of 117 matches
Mail list logo
