IRIX and Software

hey!!! its working now...mhhh
feeling like a dumb
thanks for help

Just revisiting this old thread...

squeen, are you still running octave? I just dowloaded it from neko's server and I've tried it out and got into trouble:

a=rand(50);
b=inv(a);

and Octave bombs out with:


Unable to start dynamic threads: Operation not permitted


I was wondering if this has something to do with the way it was compiled (threads)? So I tried to get the latest sources and follow the instructions in this thread to build it but, boy did I see a lot of errors...

My octave usage is limited since that is not the only error I've bumped into.
I tried a few weeks ago another port to nekoware, but it was not a determined effort and I gave up for now (may errors).

Basically, last year I got sick and tired of Mathworks dropping IRIX and Octave's C++ spagetti code so started writing a MATLAB compatable replacement in C using lex and yacc. It's native IRIX with threading. shared-memory, Motif GUI, a real-time/parallel processing Simulink replacement and an OpenGL Performer multi-pipe visualization back-end that kicks STK's (Satellite Tool Kit) ass. Unfortunately, since it is work realated (a front-end to our simulator) I cannot release the code unless I can get permission to GLP it. So far it's working out really nice. It's surprising how simple the whole parser is, certainly not worth $50K/yr for maintainace (what we pay for all of our licenses).

Also, I really find it absurd that Octave is such a non-portable pain-in-the-ass since it's a command line program---but that seems to be standard fare for C++ apps (e.g. Celestia, OpenOffice, AbiWord, etc.). Oh well...

That said, I'm gonna try for neko_octave again at some point, just 'cause nekoware wouldn't be complete without it.

hi,

I didn't see Octave in the nekoware directory. Is the current build still unstable on Irix?

thanks

mnorton wrote: hi,

I didn't see Octave in the nekoware directory. Is the current build still unstable on Irix?

thanks

We never had a Nekoware version of Octave - just a version I built a while back that doesn't work correctly.

Another revisit to this aging thread:

On an unrealed matter, I figured out what does the "Unable to start dynamic threads: Operation not permitted" error message mean and how to get rid of it:

It has to do with OpenMP created dynamic threads and the ability/willingness of IRIX to create more threads than the number of available CPUs...

So, by setting



the error massage goes away and Octave happily computes the values and works OK...so far.

PixelDust1 wrote: Just revisiting this old thread...

squeen, are you still running octave? I just dowloaded it from neko's server and I've tried it out and got into trouble:

a=rand(50);
b=inv(a);

and Octave bombs out with:


Unable to start dynamic threads: Operation not permitted


I was wondering if this has something to do with the way it was compiled (threads)? So I tried to get the latest sources and follow the instructions in this thread to build it but, boy did I see a lot of errors...

Interesting. I haven't really gotten into OpenMP much, but that good to know.

Hi!

Has anybody here already tried to compile the final Octave 3.0 ( http://www.gnu.org/software/octave/download.html ) under IRIX?
I would like to play with it, but i'm not very enthused to fight against the compiler for ages... I would give it a try this evening if not somebody else has it already done.

Michael