[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Octave-bug-tracker] [bug #49517] [mxe] failing tests in 64bit built (ac
From: |
Hartmut |
Subject: |
[Octave-bug-tracker] [bug #49517] [mxe] failing tests in 64bit built (accuracy related) |
Date: |
Wed, 2 Nov 2016 19:02:29 +0000 (UTC) |
User-agent: |
Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:49.0) Gecko/20100101 Firefox/49.0 |
URL:
<http://savannah.gnu.org/bugs/?49517>
Summary: [mxe] failing tests in 64bit built (accuracy
related)
Project: GNU Octave
Submitted by: hardy
Submitted on: Wed 02 Nov 2016 07:02:26 PM GMT
Category: Test Suite
Severity: 3 - Normal
Priority: 5 - Normal
Item Group: Inaccurate Result
Status: None
Assigned to: None
Originator Name:
Originator Email:
Open/Closed: Open
Discussion Lock: Any
Release: 4.2.0-rc3
Operating System: Microsoft Windows
_______________________________________________________
Details:
I have installed the "official" Octave-4.2.0-rc3 Windows installer from
alpha.gnu.org in its 64bit flavour on a Win7 system (I am not sure what kind
of "64bit" is inside this installer).
I see the following failing tests:
* publish.m as reported in bug #49501
* eig.cc-tst as reported in bug #49470
* Two more failing tests, which seem to be accuracy related (as bug #49470).
They happen in "axis.m" and in "assert.m".
These are the relevant parts from fntest.log for those two "new" test
failures:
>>>>> processing
C:\Octave\Octave-4.2.0-rc3-w64\share\octave\4.2.0-rc3\m\plot\appearance\axis.m
***** test
hf = figure ("visible", "off");
unwind_protect
a = logspace (-5, 1, 10);
loglog (a, -a);
axis tight;
assert (axis (), [1e-5, 10, -10, -1e-5]);
unwind_protect_cleanup
close (hf);
end_unwind_protect
!!!!! test failed
ASSERT errors for: assert (axis (),[1e-5, 10, -10, -1e-5])
Location | Observed | Expected | Reason
(1) 1e-005 1e-005 Abs err 5.0822e-021 exceeds tol 0
(4) -1e-005 -1e-005 Abs err 5.0822e-021 exceeds tol 0
>>>>> processing
C:\Octave\Octave-4.2.0-rc3-w64\share\octave\4.2.0-rc3\m\testfun\assert.m
***** test
x = [-40:0];
y1 = (10.^x).*(10.^x);
y2 = 10.^(2*x);
assert (y1, y2, eps (y1));
fail ("assert (y1, y2 + eps*1e-70, eps (y1))");
!!!!! test failed
ASSERT errors for: assert (y1,y2,eps (y1))
Location | Observed | Expected | Reason
(11) 1e-060 1e-060 Abs err 4.1454e-076 exceeds tol
1.3818e-076
(13) 1e-056 1e-056 Abs err 3.3959e-072 exceeds tol
1.132e-072
(14) 1e-054 1e-054 Abs err 4.3467e-070 exceeds tol
1.4489e-070
(19) 1e-044 1e-044 Abs err 3.7338e-060 exceeds tol
1.2446e-060
(22) 1e-038 1e-038 Abs err 2.6101e-054 exceeds tol
1.3051e-054
(23) 1e-036 1e-036 Abs err 3.341e-052 exceeds tol
1.6705e-052
(27) 1e-028 1e-028 Abs err 2.2421e-044 exceeds tol
1.121e-044
(29) 1e-024 1e-024 Abs err 3.6734e-040 exceeds tol
1.8367e-040
I don't know if those test failures would also happen in a "64bit" built under
linux.
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/bugs/?49517>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/
- [Octave-bug-tracker] [bug #49517] [mxe] failing tests in 64bit built (accuracy related),
Hartmut <=