On 21/04/2023 23:33, Nathan Bossart wrote:
On Fri, Apr 21, 2023 at 01:50:34PM +0700, John Naylor wrote:
On Wed, Mar 8, 2023 at 7:25 AM Nathan Bossart
wrote:
was mostly a fun weekend project, and I don't presently have any concrete
examples of workloads where this might help.
It seems like th
On Fri, Apr 21, 2023 at 01:50:34PM +0700, John Naylor wrote:
> On Wed, Mar 8, 2023 at 7:25 AM Nathan Bossart
> wrote:
>> was mostly a fun weekend project, and I don't presently have any concrete
>> examples of workloads where this might help.
>
> It seems like that should be demonstrated before s
On Wed, Mar 8, 2023 at 7:25 AM Nathan Bossart
wrote:
>
> was mostly a fun weekend project, and I don't presently have any concrete
> examples of workloads where this might help.
It seems like that should be demonstrated before seriously considering
this, like a profile where the relevant list fun
Here is a new patch set. I've split it into two patches: one for the
64-bit functions, and one for the 32-bit functions. I've also added tests
for pg_lfind64/pg_lfind64_idx and deduplicated the code a bit.
--
Nathan Bossart
Amazon Web Services: https://aws.amazon.com
>From 7dbbf61e8a2546a73413c
On 15/03/23 21:53, Nathan Bossart wrote:
Did you try building without SIMD support? This is what I see:
list.c: In function ‘list_member_ptr’:
list.c:697:2: warning: ISO C90 forbids mixed declarations and code
[-Wdeclaration-after-statement]
697 | const ListCell
On Wed, Mar 15, 2023 at 07:31:46PM +0530, Ankit Kumar Pandey wrote:
>> On 14/03/23 03:10, Nathan Bossart wrote:
>> On Sat, Mar 11, 2023 at 09:41:18AM +, Ankit Kumar Pandey wrote:
>> > 1. In list_member_ptr, will it be okay to bring `const ListCell
>> > *cell` from #ifdef USE_NO_SIMD
>> >con
Agree with your points Nathan. Just a headup.
On 14/03/23 03:10, Nathan Bossart wrote:
On Sat, Mar 11, 2023 at 09:41:18AM +, Ankit Kumar Pandey wrote:
1. In list_member_ptr, will it be okay to bring `const ListCell *cell` from
#ifdef USE_NO_SIMD
const ListCell *cell;
#endif
to #els
Thanks for taking a look.
On Sat, Mar 11, 2023 at 09:41:18AM +, Ankit Kumar Pandey wrote:
> 1. In list_member_ptr, will it be okay to bring `const ListCell *cell` from
> #ifdef USE_NO_SIMD
> const ListCell *cell;
> #endif
> to #else like as mentioned below? This will make visual separat
> 7. Is it possible to add a benchmark which shows improvement against
HEAD ?
Please ignore this from my earlier mail, I just saw stats now.
Thanks,
Ankit
The following review has been posted through the commitfest application:
make installcheck-world: tested, passed
Implements feature: not tested
Spec compliant: not tested
Documentation:not tested
Hello,
Adding some review comments:
1. In list_member_ptr, will it be
cfbot's Windows build wasn't happy with a couple of casts. I applied a
fix similar to c6a43c2 in v2. The patch is still very much a work in
progress.
--
Nathan Bossart
Amazon Web Services: https://aws.amazon.com
>From 055717233c47518ae48119938ebd203cc55f7f3c Mon Sep 17 00:00:00 2001
From: Natha
On Wed, Mar 08, 2023 at 01:54:15PM +1300, David Rowley wrote:
> Interesting and quite impressive performance numbers.
Thanks for taking a look.
> From having a quick glance at the patch, it looks like you'll need to
> take some extra time to make it work on 32-bit builds.
At the moment, the supp
On Wed, 8 Mar 2023 at 13:25, Nathan Bossart wrote:
> I've attached a work-in-progress patch that implements these optimizations
> for both x86 and arm, and I will register this in the July commitfest. I'm
> posting this a little early in order to gauge interest.
Interesting and quite impressive
I noticed that several of the List functions do simple linear searches that
can be optimized with SIMD intrinsics (as was done for XidInMVCCSnapshot in
37a6e5d). The following table shows the time spent iterating over a list
of n elements (via list_member_int) one billion times on my x86 laptop.
14 matches
Mail list logo
