On 09/ 3/12 11:20 PM, Dr. David Kirkby wrote:
I get a doctest failure here on Solaris 10 SPARC with
sage/numerical/optimize.py. I can't find this is trac, but is anyone
aware of it being listed as a bug?
No answers, so it is now
http://trac.sagemath.org/sage_trac/ticket/13427
Does anyone know an accurate answer to this? I tried to do it in
Mathematica, using the FindMaximum[] command, but I can't see to get
that to return anything other than machine precision, despite adding
options I would have thought would have increased the accuracy
Again no answers, but I managed to work out how to do it in Mathematica.
In[3]:= FindMaximum[8 E^(-x) Sin[x] -1,{x,0,8},WorkingPrecision->50]
Out[3]= {1.5791755355586755940189354764810356787023224274029,
> {x -> 0.78539816339744830961566077712689095303902649201882}}
In[4]:=
(This result was computed by Mathematica 7 running on OpenSolaris on a quad core
Xeon x86 CPU, though the working precision would have meant Mathematica did not
use the FPU)
OK, so now I have what probably is a fairly accurate answer.
Anyway, here is the failure, on Solaris 10 SPARC, with gcc 4.5.0 (not
the gcc in Sage).
sage -t --long -force_lib
devel/sage/sage/numerical/backends/coin_backend.pyx
[11.0 s]
sage -t --long -force_lib devel/sage/sage/numerical/optimize.py
**********************************************************************
File
"/export/home/drkirkby/sage-5.3.rc0/devel/sage-main/sage/numerical/optimize.py",
line 135:
sage: find_local_maximum(fast_float(8*e^(-x)*sin(x) - 1, x), 0, 8)
Expected:
(1.579175535558..., 0.78539817...)
Got:
(1.5791755355586758, 0.78539816111340355)
Note the expected value of the Sage test "(1.579175535558..., 0.78539817...)" is
actually incorrect in the second instance as it should be 0.78539816 and not
0.78539817. The result returned on the SPARC processor is more accurate in this
particular instance, where the absolute error is -2.84x10^-9. So the test needs
changing to be
Expected:
(1.579175535558..., 0.7853981...)
to take account of the various errors when this test is run on different CPUs.
This is
Dave
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To post to this group, send email to sage-devel@googlegroups.com.
To unsubscribe from this group, send email to
sage-devel+unsubscr...@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel?hl=en.
