Re: RC2 test failures

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: RC2 test failures

Rik-4
On 02/06/2019 09:00 AM, [hidden email] wrote:
Subject:
Re: RC2 / 5.0.91
From:
chechu [hidden email]
Date:
02/06/2019 07:33 AM
To:
[hidden email]
List-Post:
[hidden email]
Content-Transfer-Encoding:
7bit
Precedence:
list
MIME-Version:
1.0
References:
[hidden email] [hidden email]
In-Reply-To:
[hidden email]
Message-ID:
[hidden email]
Content-Type:
text/plain; charset=us-ascii
Message:
4

Good evening to everyne.

It compiles and checks with some Fails on Debian Stretch 9.0

- Jit Disabled
- 64bit indexing and BLAS
- Openblas 0.3.5 and Atlas 5.4.0


  PASS                            15156
  FAIL                                8
  XFAIL (reported bug)               23
  SKIP (missing feature)            391
  SKIP (run-time condition)          12


Fails on amd.cc-tst
Fails on which.m
Fails on gammainc.m
Could you locate the 8 failures in the file test/fntests.log and post them to a new bug report at bugs.octave.org?

I suspect the gammainc ones are just a tolerance issue.

--Rik
Reply | Threaded
Open this post in threaded view
|

Re: RC2 test failures

chechu
Good morning to everyone.

Thankyou so much Rik for your considerations.

On my +5 years Debian I had few ancient suitesparse libraries that seemed
to, somehow, make appear the other errors, rightnow they are cleaned and
after successfully clean, reconfig/rebuild only the gammainc errors remain.
They where not indeed bugs to report.

As You suspected it is because of tolerances, the actual hugh error is as
follows:

ASSERT errors for:  assert (gammainc (200, 200,
"scaledupper"),17.3984438553791505135122900,-2 * eps)

  Location  |  Observed  |  Expected  |  Reason
     ()        17.3984      17.3984      Rel err 6.1259e-16 exceeds tol
4.4409e-16 by 2e-16



I don't know if it may matter, but Debian is starting to use gcc 8
toolchains.


Summary:

  PASS                            15514
  FAIL                                3
  XFAIL (reported bug)               23
  SKIP (missing feature)             38        --LLVM
  SKIP (run-time condition)          12




Thankyou all for your efforts.
Kind regards







--
Sent from: http://octave.1599824.n4.nabble.com/Octave-Maintainers-f1638794.html

Reply | Threaded
Open this post in threaded view
|

Re: RC2 test failures

Mike Miller-4
Thanks for testing and following up on your observed test failures.

On Thu, Feb 07, 2019 at 02:10:29 -0600, chechu wrote:
> As You suspected it is because of tolerances, the actual hugh error is as
> follows:
>
> ASSERT errors for:  assert (gammainc (200, 200,
> "scaledupper"),17.3984438553791505135122900,-2 * eps)
>
>   Location  |  Observed  |  Expected  |  Reason
>      ()        17.3984      17.3984      Rel err 6.1259e-16 exceeds tol
> 4.4409e-16 by 2e-16

Fortunately this was already reported and fixed on the stable branch, so
this won't fail for you with the final Octave 5.1 release.

  https://hg.savannah.gnu.org/hgweb/octave/rev/eb1d4a619260

> I don't know if it may matter, but Debian is starting to use gcc 8
> toolchains.

I'm not sure about other Octave developers, but I've been working with
gcc 8 for at least 6 months now, and with gcc-trunk occasionally.

--
mike

signature.asc (849 bytes) Download Attachment