MathGroup Archive 2006

[Date Index] [Thread Index] [Author Index]

Search the Archive

Kernel hang during FindFit under 5.1 for Linux

  • To: mathgroup at smc.vnet.net
  • Subject: [mg63556] Kernel hang during FindFit under 5.1 for Linux
  • From: "Xerxes" <Xerxes314 at gmail.com>
  • Date: Thu, 5 Jan 2006 03:12:24 -0500 (EST)
  • Sender: owner-wri-mathgroup at wolfram.com

Hi all,

I was running some of my analysis code that I originally produced under
version 5.0 on Linux, and to my surprise, the notebook would not run
under 5.1. I fiddled with the part where it was hanging and reduced it
to the appended test case in which increasingly many data points
are added to a trivial FindFit to a constant. No result is produced
when 17 data points are taken; at that point, the Kernel hangs and
does not quit when Quit is selected from the menu. Its process
cannot be killed normally and must be killed using kill -9. The bug
is not reproduced under 5.1 for Windows.

I fished around here but did not see any similar cases. So my questions
would be:

Is there a workaround for this bug?
Should I bother the sysadmin to upgrade me to 5.2?
Should I report this to Wolfram?

Thanks,
Xerxes

In[1]:=
$Version
data = Table[Random[], {30}];
Do[
  Print[i, ": ",
    FindFit[Take[data, i], a, {a}, x]
  ],
{i, Length[data]}]

Out[1]=
5.1 for Linux (October 25, 2004)

1: {a->0.481796}
...
clipped 2 through 15
...
16: {a->0.576249}


  • Prev by Date: Re: Re: Re: Questions regarding MatrixExp, and its usage
  • Next by Date: Re: Best linear Fit to slope data with Fixed starting point/value.
  • Previous by thread: Re: Re: Re: Questions regarding MatrixExp, and its usage
  • Next by thread: Re: Kernel hang during FindFit under 5.1 for Linux