Mathematica 5.2

[phoebus]

I've been trying Mathematica 5.2 - Student verision on osx 86/Rosetta. It shows a peculiar behaviour: the frontend runs, the kernel (MathKernel) runs from the Terminal, and actually computes calculations. But when I start a calculation from the frontend (doodle+enter) the kernel starts, and hangs, it grabs a lot of cpu time and never exits. Any hint ?

[maltloaf]

I've been trying Mathematica 5.2 - Student verision on osx 86/Rosetta. It shows a peculiar behaviour: the frontend runs, the kernel (MathKernel) runs from the Terminal, and actually computes calculations. But when I start a calculation from the frontend (doodle+enter) the kernel starts, and hangs, it grabs a lot of cpu time and never exits. Any hint ?

<{POST_SNAPBACK}>

 

i would have thought that this (if the latest version) was x86 native already as mathematica was the very first non apple app to be compiled as universal. They used it as an example in the keynote speech where jobs announced intel macs.

[phoebus]

i would have thought that this (if the latest version) was x86 native already as mathematica was the very first non apple app to be compiled as universal. They used it as an example in the keynote speech where jobs announced intel macs.

<{POST_SNAPBACK}>

 

Um, no.. It hast just some 64 bit enhancements but it is PPC only. The WWDC demo was based on an internal developement version with special features.

[bakel]

Same problem here with Mathematica 5.1 .

Any solutions yet?

[bakel]

Same problem here with Mathematica 5.1 .

Any solutions yet?

<{POST_SNAPBACK}>

 

I found a solution:

 

- Enable SSH and install an authorized_keys file in ~/.ssh/ for login from localhost

- put a script "math.sh" in /sbin/ that calls the mathematica kernel (as suggested by wolfram).

- Change the properties of the local kernel so it will connect to a remote kernel on localhost using ssh ( user@localhost ) and redirect stderr to /dev/null ( 2&>/dev/null )just before you call math.sh .

 

This worked fine for me on my IBM Thinkpand T42, Pentium M 1.7GHz 768MB RAM.

Result of the mathematica benchmark:

<<Utilities`

BenchmarkReport[];

 

is 0.51

 

Does anyone have numbers for their machine (also powerbooks etc.).

I think performance is good knowing rosetta is translating the powerpc application.

[phoebus]

This worked fine for me on my IBM Thinkpand T42, Pentium M 1.7GHz 768MB RAM.

Result of the mathematica benchmark:

<<Utilities`

BenchmarkReport[];

 

is 0.51

 

Does anyone have numbers for their machine (also powerbooks etc.).

I think performance is good knowing rosetta is translating the powerpc application.

<{POST_SNAPBACK}>

 

It works, thanks! Can't post the benchmakr results, tough, since Mathematica hangs on Test 3 - Discrete Fourier Transform. On my Windows System (a 3.0 P4 Prescott with 1mb of cache) I get 1.35