Wolfram, meet Stefan and Boltzmann

[Curtis Osterhoudt]

Again, this happens on Linux, too. It's not limited to OS X

In[1]:= 1 + 1

Out[1]= 2

In[2]:= $Version

Out[2]= "7.0 for Linux x86 (32-bit) (November 11, 2008)"

In[3]:= AbsoluteTiming[Integrate[x^3/(Exp[x] - 1), {x, 0, Infinity}]]

Out[3]= {15.64583`7.645943600598422, \[Pi]^4/15}

AES, are you running Mathematica 7? If so, then maybe it's a problem with
Mathematica 7 for OS X (all the other people with the same issue as you who
posted in this thread seem to be running Mathematica 7 on OS X).
Heike.

Have to confess, that's what I'm running -- and I really thought I was
running 8, until I looked just now.

[Peter Pein]

Post by Curtis Osterhoudt
Again, this happens on Linux, too. It's not limited to OS X
In[1]:= 1 + 1
Out[1]= 2
In[2]:= $Version
Out[2]= "7.0 for Linux x86 (32-bit) (November 11, 2008)"
In[3]:= AbsoluteTiming[Integrate[x^3/(Exp[x] - 1), {x, 0, Infinity}]]
Out[3]= {15.64583`7.645943600598422, \[Pi]^4/15}

...
sorry, I can not reproduce your results:
In[1]:= 1+1
$Version
$ReleaseNumber
AbsoluteTiming[Integrate[x^3/(Exp[x]-1),{x,0,Infinity}]]


Out[1]= 2
Out[2]= 8.0 for Microsoft Windows (64-bit) (February 23, 2011)
Out[3]= 1
Out[4]= {1.6692030,\[Pi]^4/15}

and similar results under Linux

maybe a clean reinstall (uninstall Mathematica, reboot, install Mathematica) might help?

[Peter Pein]

Post by Curtis Osterhoudt
Again, this happens on Linux, too. It's not limited to OS X
In[1]:= 1 + 1
Out[1]= 2
In[2]:= $Version
Out[2]= "7.0 for Linux x86 (32-bit) (November 11, 2008)"
In[3]:= AbsoluteTiming[Integrate[x^3/(Exp[x] - 1), {x, 0, Infinity}]]
Out[3]= {15.64583`7.645943600598422, \[Pi]^4/15}

I do not know which problems arise when using OS X under Windows7 I get:
In[1]:= 1 + 1
$Version
$ReleaseNumber
AbsoluteTiming[Integrate[x^3/(Exp[x] - 1), {x, 0, Infinity}]]



Out[1]= 2

Out[2]= "8.0 for Microsoft Windows (64-bit) (February 23, 2011)"

Out[3]= 1

Out[4]= {1.6692030, \[Pi]^4/15}

and comparable results using ubuntu 10.10..

[George Woodrow III]

OK. This has gone on long enough that I want to add my experience:

In[1]:= 1+1
$Version
$ReleaseNumber
AbsoluteTiming[Integrate[x^3/(Exp[x]-1),{x,0,Infinity}]]

Out[1]= 2
Out[2]= 8.0 for Mac OS X x86 (64-bit) (February 23, 2011)
Out[3]= 1
Out[4]= {2.079332,\[Pi]^4/15}

This is on a 2010 iMac with 8 Gbytes RAM, but otherwise stock. Using the latest release of Snow Leopard (10.6.7).

On my MacBook Pro, I get a timing of 3.239342 seconds. This is a year older, but also has 8 Gbytes RAM running the OS X 10.6.7. It uses the faster GPU.

Since there seems to be no correlation between hardware and slow execution, there must be something else going on.

george

Post by Peter Pein

Post by Curtis Osterhoudt
Again, this happens on Linux, too. It's not limited to OS X
In[1]:= 1 + 1
Out[1]= 2
In[2]:= $Version
Out[2]= "7.0 for Linux x86 (32-bit) (November 11, 2008)"
In[3]:= AbsoluteTiming[Integrate[x^3/(Exp[x] - 1), {x, 0, Infinity}]]
Out[3]= {15.64583`7.645943600598422, \[Pi]^4/15}

...
In[1]:= 1+1
$Version
$ReleaseNumber
AbsoluteTiming[Integrate[x^3/(Exp[x]-1),{x,0,Infinity}]]
Out[1]= 2
Out[2]= 8.0 for Microsoft Windows (64-bit) (February 23, 2011)
Out[3]= 1
Out[4]= {1.6692030,\[Pi]^4/15}
and similar results under Linux
maybe a clean reinstall (uninstall Mathematica, reboot, install Mathematica) might help?

[Daniel Lichtblau]

Post by Peter Pein

Post by Curtis Osterhoudt
Again, this happens on Linux, too. It's not limited to OS X
In[1]:= 1 + 1
Out[1]= 2
In[2]:= $Version
Out[2]= "7.0 for Linux x86 (32-bit) (November 11, 2008)"
In[3]:= AbsoluteTiming[Integrate[x^3/(Exp[x] - 1), {x, 0, Infinity}]]
Out[3]= {15.64583`7.645943600598422, \[Pi]^4/15}

...
In[1]:= 1+1
$Version
$ReleaseNumber
AbsoluteTiming[Integrate[x^3/(Exp[x]-1),{x,0,Infinity}]]
Out[1]= 2
Out[2]= 8.0 for Microsoft Windows (64-bit) (February 23, 2011)
Out[3]= 1
Out[4]= {1.6692030,\[Pi]^4/15}
and similar results under Linux
maybe a clean reinstall (uninstall Mathematica, reboot, install Mathematica) might help?

Not needed. I can reproduce the result (the bad timing) in version 7. It
got faster in version 8.

Daniel Lichtblau
Wolfram Research

[AES]

Post by Daniel Lichtblau
Not needed. I can reproduce the result (the bad timing) in version 7. It
got faster in version 8.
Daniel Lichtblau
Wolfram Research

Again, my apologies, as the OP on this extended thread. When I first
got the 19 second timing on evaluating this integral I believed I was
running version 8, but I was actually still running version 7, for which
this is apparently the correct (if somewhat surprisingly long) timing.
I'll spare you the details of how my original error happened. --AES