Re: 'point' test suite failure on arm

Hi PostgreSQL porters!

I recently noticed that the PostgreSQL test suite fails on arm in the
'point' test (diff pasted below). However, the difference is really
small and is well within the rounding error range. It just seems that
the expected output has one digit less than the actual result.

Is this a mere bug in the test suite? Any idea why it just occurs on
the arm platform? It works fine on all other 11 platforms of Debian.

Please see [1] for the full build log which includes the test suite
output. This is the log for 8.1.1, but the same occurs with 7.4 and
8.0.

Thanks in advance for any hint,

Martin

*** ./expected/point.out Thu Apr 7 01:51:40 2005
--- ./results/point.out Mon Dec 12 20:51:59 2005
***************
*** 101,107 ****
| (-3,4) | 5
| (-10,0) | 10
| (-5,-12) | 13
! | (10,10) | 14.142135623731
| (5.1,34.5) | 34.8749193547455
(6 rows)

--- 101,107 ----
| (-3,4) | 5
| (-10,0) | 10
| (-5,-12) | 13
! | (10,10) | 14.1421356237309
| (5.1,34.5) | 34.8749193547455
(6 rows)

***************
*** 127,134 ****
| (-5,-12) | (-10,0) | 13
| (-5,-12) | (0,0) | 13
| (0,0) | (-5,-12) | 13
! | (0,0) | (10,10) | 14.142135623731
! | (10,10) | (0,0) | 14.142135623731
| (-3,4) | (10,10) | 14.3178210632764
| (10,10) | (-3,4) | 14.3178210632764
| (-5,-12) | (-3,4) | 16.1245154965971
--- 127,134 ----
| (-5,-12) | (-10,0) | 13
| (-5,-12) | (0,0) | 13
| (0,0) | (-5,-12) | 13
! | (0,0) | (10,10) | 14.1421356237309
! | (10,10) | (0,0) | 14.1421356237309
| (-3,4) | (10,10) | 14.3178210632764
| (10,10) | (-3,4) | 14.3178210632764
| (-5,-12) | (-3,4) | 16.1245154965971
***************
*** 198,204 ****
| (-10,0) | (0,0) | 10
| (-10,0) | (-5,-12) | 13
| (-5,-12) | (0,0) | 13
! | (0,0) | (10,10) | 14.142135623731
| (-3,4) | (10,10) | 14.3178210632764
| (-5,-12) | (-3,4) | 16.1245154965971
| (-10,0) | (10,10) | 22.3606797749979
--- 198,204 ----
| (-10,0) | (0,0) | 10
| (-10,0) | (-5,-12) | 13
| (-5,-12) | (0,0) | 13
! | (0,0) | (10,10) | 14.1421356237309
| (-3,4) | (10,10) | 14.3178210632764
| (-5,-12) | (-3,4) | 16.1245154965971
| (-10,0) | (10,10) | 22.3606797749979

[1] http://buildd.debian.org/fetch.php?&pkg=postgresql-8.1&ver=8.1.1-1&arch=arm&stamp=1134423129&file=log&as=raw
--
Martin Pitt http://www.piware.de
Ubuntu Developer http://www.ubuntu.com
Debian Developer http://www.debian.org

In a world without walls and fences, who needs Windows and Gates?

Martin Pitt <martin(at)piware(dot)de> writes:
> I recently noticed that the PostgreSQL test suite fails on arm in the
> 'point' test (diff pasted below). However, the difference is really
> small and is well within the rounding error range. It just seems that
> the expected output has one digit less than the actual result.

> Is this a mere bug in the test suite? Any idea why it just occurs on
> the arm platform? It works fine on all other 11 platforms of Debian.

Floating point math a bit different on the ARM? The quoted result is
basically sqrt(200::float8), and we've never before heard reports of
different platforms printing different answers, though I suppose it's
possible if the FP hardware is not IEEE-spec. I'd suggest asking
whoever maintains libm for Debian.

regards, tom lane

Hi Tom!

Tom Lane [2005-12-20 11:29 -0500]:
> Martin Pitt <martin(at)piware(dot)de> writes:
> > I recently noticed that the PostgreSQL test suite fails on arm in the
> > 'point' test (diff pasted below). However, the difference is really
> > small and is well within the rounding error range. It just seems that
> > the expected output has one digit less than the actual result.
>
> > Is this a mere bug in the test suite? Any idea why it just occurs on
> > the arm platform? It works fine on all other 11 platforms of Debian.
>
> Floating point math a bit different on the ARM? The quoted result is
> basically sqrt(200::float8), and we've never before heard reports of
> different platforms printing different answers, though I suppose it's
> possible if the FP hardware is not IEEE-spec. I'd suggest asking
> whoever maintains libm for Debian.

I will do that. I just wondered why the expected and actual result are
actually the same number, but differ only in the presented precision.

Thanks,

Martin

--
Martin Pitt http://www.piware.de
Ubuntu Developer http://www.ubuntu.com
Debian Developer http://www.debian.org

In a world without walls and fences, who needs Windows and Gates?

On Tue, Dec 20, 2005 at 05:45:33PM +0100, Martin Pitt wrote:
> Hi Tom!
>
> Tom Lane [2005-12-20 11:29 -0500]:
> > Martin Pitt <martin(at)piware(dot)de> writes:
> > > I recently noticed that the PostgreSQL test suite fails on arm in the
<snip>
> I will do that. I just wondered why the expected and actual result are
> actually the same number, but differ only in the presented precision.

You should consider setting that machine up in the buildfarm as well (I
don't see any arm machines in there right now):
http://pgbuildfarm.org/cgi-bin/show_status.pl
--
Jim C. Nasby, Sr. Engineering Consultant jnasby(at)pervasive(dot)com
Pervasive Software http://pervasive.com work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf cell: 512-569-9461

Hi Jim!

Jim C. Nasby [2005-12-20 15:37 -0600]:
> On Tue, Dec 20, 2005 at 05:45:33PM +0100, Martin Pitt wrote:
> > Hi Tom!
> >
> > Tom Lane [2005-12-20 11:29 -0500]:
> > > Martin Pitt <martin(at)piware(dot)de> writes:
> > > > I recently noticed that the PostgreSQL test suite fails on arm in the
> <snip>
> > I will do that. I just wondered why the expected and actual result are
> > actually the same number, but differ only in the presented precision.
>
> You should consider setting that machine up in the buildfarm as well (I
> don't see any arm machines in there right now):
> http://pgbuildfarm.org/cgi-bin/show_status.pl

Unfortunately these are not my machines, these are the official Debian
package autobuilders. Access to them is very restricted and I doubt
that we can use them for PostgreSQL. I will ask, though.

BTW, that problem does only occur for 7.4, it seems to work fine for
8.0 and 8.1. So I will just ignore this particular test on arm for
7.4.

Thanks,

Martin
--
Martin Pitt http://www.piware.de
Ubuntu Developer http://www.ubuntu.com
Debian Developer http://www.debian.org

Martin Pitt wrote:
> Hi Jim!
>
> Jim C. Nasby [2005-12-20 15:37 -0600]:
> > On Tue, Dec 20, 2005 at 05:45:33PM +0100, Martin Pitt wrote:
> > > Hi Tom!
> > >
> > > Tom Lane [2005-12-20 11:29 -0500]:
> > > > Martin Pitt <martin(at)piware(dot)de> writes:
> > > > > I recently noticed that the PostgreSQL test suite fails on arm in the
> > <snip>
> > > I will do that. I just wondered why the expected and actual result are
> > > actually the same number, but differ only in the presented precision.
> >
> > You should consider setting that machine up in the buildfarm as well (I
> > don't see any arm machines in there right now):
> > http://pgbuildfarm.org/cgi-bin/show_status.pl
>
> Unfortunately these are not my machines, these are the official Debian
> package autobuilders. Access to them is very restricted and I doubt
> that we can use them for PostgreSQL. I will ask, though.

As another option, you can do a nightly cvs update and check and
email any compile/regression to yourself. See src/tools/pgtest for an
example.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073