On Wed, 28 Jul 2021 at 14:25, David Fetter wrote:
> Intuitively, I'd expect us to get things in the neighborhood of 1 a
> lot more often than things in the neighborhood of 1 << (30 or 60). Do
> we have some idea of the distribution, or at least of the distribution
> family, that we should expect
On Wed, Jul 28, 2021 at 01:17:43PM +1200, David Rowley wrote:
> On Wed, 28 Jul 2021 at 01:44, Alvaro Herrera wrote:
> > So how much faster is it than the original?
>
> I only did some very quick tests. They're a bit noisey. The results
> indicate an average speedup of 1.7%, but the noise level i
On Wed, 28 Jul 2021 at 01:44, Alvaro Herrera wrote:
> So how much faster is it than the original?
I only did some very quick tests. They're a bit noisey. The results
indicate an average speedup of 1.7%, but the noise level is above
that, so unsure.
create table a (a int);
insert into a select a
On 2021-Jul-26, David Fetter wrote:
> Folks,
>
> Please find attached a patch to do $subject. It's down to a one table
> lookup and 3 instructions.
So how much faster is it than the original?
--
Álvaro Herrera 39°49'30"S 73°17'W — https://www.EnterpriseDB.com/
"La rebeldía es la v
On Tue, Jul 27, 2021 at 3:08 PM Andrew Gierth
wrote:
>
> In short, please don't include d2s_intrinsics.h from anywhere other than
> d2s.c
>
Thanks for the clarification.
The patch author will need to take this into account for the patch's
div1e8() usage.
Regards,
Greg Nancarrow
Fujitsu Australi
> "David" == David Rowley writes:
David> I think the mistake is that the header file is not in
David> src/include/common. For some reason, it's ended up with all the
David> .c files in src/common.
David> I imagine Andrew did this because he didn't ever expect anything
David> else to hav
On Tue, 27 Jul 2021 at 15:08, Greg Nancarrow wrote:
>
> On Tue, Jul 27, 2021 at 12:42 PM Michael Paquier wrote:
> >
> > #include "common/int.h"
> > +#include "d2s_intrinsics.h"
> > Er, are you sure about this part? The first version of the patch did
> > that in a different, also incorrect, way.
On Tue, Jul 27, 2021 at 12:42 PM Michael Paquier wrote:
>
> #include "common/int.h"
> +#include "d2s_intrinsics.h"
> Er, are you sure about this part? The first version of the patch did
> that in a different, also incorrect, way.
Er, I was just trying to help out, so at least the patch could be
On Tue, 27 Jul 2021 at 14:42, Michael Paquier wrote:
> When applying some micro-benchmarking to stress those APIs, how much
> does this change things? At the end of the day, this also comes down
> to an evaluation of pg_ulltoa_n() and pg_ultoa_n().
I'd suggest something like creating a table wit
On Tue, Jul 27, 2021 at 12:28:22PM +1000, Greg Nancarrow wrote:
> That patch didn't apply for me (on latest source) so I've attached an
> equivalent with those changes, that does apply, and also tweaks the
> Makefile include path to address that #include issue.
When applying some micro-benchmarkin
On Tue, Jul 27, 2021 at 9:51 AM David Fetter wrote:
>
> In covering the int64 versions, I swiped a light weight division from
> the Ryu stuff. I'm pretty sure that what I did is not how to do
> #includes, but it's a PoC. What would be a better way to do this?
>
That patch didn't apply for me (o
Folks,
Please find attached a patch to do $subject. It's down to a one table
lookup and 3 instructions.
In covering the int64 versions, I swiped a light weight division from
the Ryu stuff. I'm pretty sure that what I did is not how to do
#includes, but it's a PoC. What would be a better way to
12 matches
Mail list logo
