installation problem/crash

I am trying to help a student get started with
Python/Scipy/Numpy/Matplotlib in windows. On one of his machines,
everything seems to install correctly, we can call figure(1) without a
problem, and plotting is fine until we try the show() command. Then
python crashes without much in the way of useful information. His
laptop is completely fine.

We have downloaded a current rc file and set the backend to TkAgg.

Any thoughts?

How do we get more info to track down the problem?

Thanks,

Ryan

I am trying to help a student get started with
Python/Scipy/Numpy/Matplotlib in windows. On one of his machines,
everything seems to install correctly, we can call figure(1) without a
problem, and plotting is fine until we try the show() command. Then
python crashes without much in the way of useful information. His
laptop is completely fine.

We have downloaded a current rc file and set the backend to TkAgg.

Any thoughts?

First try

  >>> import numpy
  >>> numpy.test()

there is a numpy problem on windows that affects older machines.

If that works, create a simple test file that generates a plot and
calls savefig and run it with

  > python myscript.py --verbose-debug-annoying

try different backends by adding -dPS or -dAgg to the verbose flag
from the command line and see which backends crash and report back
with the verbose output

JDH

···

On Dec 11, 2007 1:01 PM, Ryan Krauss <ryanlists@...287...> wrote:

Go to the windows information screens and fetch out some CPU details.
If it's a Pentium III, chances are the SSE2 instructions in the latest
numpy binary are the culprit. If it's a newer chip, we'll need to dig
deeper.

Cheers,

f

···

On Dec 11, 2007 12:01 PM, Ryan Krauss <ryanlists@...287...> wrote:

I am trying to help a student get started with
Python/Scipy/Numpy/Matplotlib in windows. On one of his machines,
everything seems to install correctly, we can call figure(1) without a
problem, and plotting is fine until we try the show() command. Then
python crashes without much in the way of useful information. His
laptop is completely fine.

We have downloaded a current rc file and set the backend to TkAgg.

Any thoughts?

How do we get more info to track down the problem?

Pretty sure it's a newer chip, but I will find out.

···

On Dec 11, 2007 2:06 PM, Fernando Perez <fperez.net@...287...> wrote:

On Dec 11, 2007 12:01 PM, Ryan Krauss <ryanlists@...287...> wrote:
> I am trying to help a student get started with
> Python/Scipy/Numpy/Matplotlib in windows. On one of his machines,
> everything seems to install correctly, we can call figure(1) without a
> problem, and plotting is fine until we try the show() command. Then
> python crashes without much in the way of useful information. His
> laptop is completely fine.
>
> We have downloaded a current rc file and set the backend to TkAgg.
>
> Any thoughts?
>
> How do we get more info to track down the problem?

Go to the windows information screens and fetch out some CPU details.
If it's a Pentium III, chances are the SSE2 instructions in the latest
numpy binary are the culprit. If it's a newer chip, we'll need to dig
deeper.

Cheers,

f

Fernando was right on. Here is his response to me:

Laptop - Ok
Windows XP Professional, Service Pack 2
AMD Athlon 64 3400+ (ClawHammer)
1.67 GHz, 768 MB of RAM
Chipset: SiS 755/755FX
Southbridge: SiS LPC Bridge
Instructions: MMX (+), 3DNow! (+), SSE, SSE2, x86-64

Machine 1 - Crashes
Windows XP Professional, Service Pack 2
AMD Athlon XP 2000+ (Thoroughbred)
1.67 GHz, 768 MB of RAM
ASUS A7V8X-X motherboard
Chipset: VIA KT400 (VT8377)
Southbridge: VIA VT8235
Instructions: MMX (+), 3DNow! (+), SSE

Machine 2 - Crashes
Windows XP Professional, Service Pack 2
AMD Athlon XP 2600+ (Barton)
1.92 GHz, 2.0 GB of RAM
ASUS A7V880 motherboard
Chipset: VIA KT880
Southbridge: VIA VT8237
Instructions: MMX (+), 3DNow! (+), SSE

I ran the following statements on both machines which caused it to crash:

import numpy
numpy.test()

Here is the output:

    Numpy is installed in C:\Python25\lib\site-packages\numpy
    Numpy version 1.0.4
    Python version 2.5.1 (r251:54863, Apr 18 2007, 08:51:08) [MSC
v.1310 32 bit (Int
    el)]
      Found 10/10 tests for numpy.core.defmatrix
      Found 36/36 tests for numpy.core.ma
      Found 223/223 tests for numpy.core.multiarray
      Found 65/65 tests for numpy.core.numeric
      Found 31/31 tests for numpy.core.numerictypes
      Found 12/12 tests for numpy.core.records
      Found 6/6 tests for numpy.core.scalarmath
      Found 14/14 tests for numpy.core.umath
      Found 4/4 tests for numpy.ctypeslib
      Found 5/5 tests for numpy.distutils.misc_util
      Found 1/1 tests for numpy.fft.fftpack
      Found 3/3 tests for numpy.fft.helper
      Found 9/9 tests for numpy.lib.arraysetops
      Found 46/46 tests for numpy.lib.function_base
      Found 5/5 tests for numpy.lib.getlimits
      Found 4/4 tests for numpy.lib.index_tricks
      Found 3/3 tests for numpy.lib.polynomial
      Found 49/49 tests for numpy.lib.shape_base
      Found 15/15 tests for numpy.lib.twodim_base
      Found 43/43 tests for numpy.lib.type_check
      Found 1/1 tests for numpy.lib.ufunclike
      Found 40/40 tests for numpy.linalg
      Found 2/2 tests for numpy.random
      Found 0/0 tests for __main__
    ................................................................................
    ................................................................................
    ................................................................................
    ................................................................................
    .....................................

Sounds like the problem is the fact that my desktop computers do not
support SSE2 instructions which are in the latest numpy binaries.
This also explains why it works fine on the laptop which does support
SSE2.

I piggy-backed onto an existing thread on the numpy list (is that bad
listserve etiquette? - probably: I now have the same question on two
lists and I tried to high jack a thread.). Unless someone has a
better idea, I will try and build from source for him. But my windows
building skills are not what they should be.....

Ryan

···

On Dec 11, 2007 2:06 PM, Fernando Perez <fperez.net@...287...> wrote:

On Dec 11, 2007 12:01 PM, Ryan Krauss <ryanlists@...287...> wrote:
> I am trying to help a student get started with
> Python/Scipy/Numpy/Matplotlib in windows. On one of his machines,
> everything seems to install correctly, we can call figure(1) without a
> problem, and plotting is fine until we try the show() command. Then
> python crashes without much in the way of useful information. His
> laptop is completely fine.
>
> We have downloaded a current rc file and set the backend to TkAgg.
>
> Any thoughts?
>
> How do we get more info to track down the problem?

Go to the windows information screens and fetch out some CPU details.
If it's a Pentium III, chances are the SSE2 instructions in the latest
numpy binary are the culprit. If it's a newer chip, we'll need to dig
deeper.

Cheers,

f