I’ve been trying to find a math library capable of linear algebra matrix
operations.

Unfortunately the MATLAB C Compiler avenue came to a dead end. I have no
source code for their proprietary libraries and so I’m looking elsewhere for
math libraries with the functionality that I need.

I’ve recently come across the LAPACK libraries and this looks promising
since they provide the source code. The only problem is that the
installation looks pretty complicated.

Q1. What I’d like to know is if anyone out there is using them in QNX4?

Check out www.netlib.org, there’s a number of different algebra
packages I recall, and/or surf around. At least basic algebra is not
difficult to find packages for. TNT and MAT are two examples.

If QNX6 (RTP) is an alternative, you can use the Pentium-II optimized
BLAS and Lapack code libraries for Linux to get high performance.

Tom

Dan Carreira wrote:

I’ve been trying to find a math library capable of linear algebra matrix
operations.

Unfortunately the MATLAB C Compiler avenue came to a dead end. I have no
source code for their proprietary libraries and so I’m looking elsewhere for
math libraries with the functionality that I need.

I’ve recently come across the LAPACK libraries and this looks promising
since they provide the source code. The only problem is that the
installation looks pretty complicated.

Q1. What I’d like to know is if anyone out there is using them in QNX4?

When you say “Pentium-II optimized” do you mean using MMX or just using the FP
stack more efficiently than compilers usually do?

Tomas Högström wrote:

Check out > www.netlib.org> , there’s a number of different algebra
packages I recall, and/or surf around. At least basic algebra is not
difficult to find packages for. TNT and MAT are two examples.

If QNX6 (RTP) is an alternative, you can use the Pentium-II optimized
BLAS and Lapack code libraries for Linux to get high performance.

Tom

Dan Carreira wrote:

I’ve been trying to find a math library capable of linear algebra matrix
operations.

Unfortunately the MATLAB C Compiler avenue came to a dead end. I have no
source code for their proprietary libraries and so I’m looking elsewhere for
math libraries with the functionality that I need.

I’ve recently come across the LAPACK libraries and this looks promising
since they provide the source code. The only problem is that the
installation looks pretty complicated.

Q1. What I’d like to know is if anyone out there is using them in QNX4?

I’ve actually downloaded all the stuff I need from netlib, or at least I
think I have.

Unfortunately I’m stuck with QNX4 as my OS.

Does anyone know where I’d go about getting a BLAS library for a Pentium
133Mhz QNX4 machine?

Thanks again

Dan Carreira

Check out > www.netlib.org> , there’s a number of different algebra
packages I recall, and/or surf around. At least basic algebra is not
difficult to find packages for. TNT and MAT are two examples.

If QNX6 (RTP) is an alternative, you can use the Pentium-II optimized
BLAS and Lapack code libraries for Linux to get high performance.
Tom

Dan Carreira wrote:

I’ve been trying to find a math library capable of linear algebra matrix
operations.

Unfortunately the MATLAB C Compiler avenue came to a dead end. I have no
source code for their proprietary libraries and so I’m looking elsewhere
for
math libraries with the functionality that I need.

I’ve recently come across the LAPACK libraries and this looks promising
since they provide the source code. The only problem is that the
installation looks pretty complicated.

Q1. What I’d like to know is if anyone out there is using them in QNX4?

MMX (probably, the doc is not 100% clear). Part of it is a port of the Intel
Math Kernel Library, so it should be.
There is also a Pentium Pro optimized version. PIII optimized version under
development…
Dual processor SMP also supported, but I dont’t know if this works under RTP. http://www.cs.utk.edu/~ghenry/distrib/

When you say “Pentium-II optimized” do you mean using MMX or just using the FP
stack more efficiently than compilers usually do?

Tomas Högström wrote:

Check out > www.netlib.org> , there’s a number of different algebra
packages I recall, and/or surf around. At least basic algebra is not
difficult to find packages for. TNT and MAT are two examples.

If QNX6 (RTP) is an alternative, you can use the Pentium-II optimized
BLAS and Lapack code libraries for Linux to get high performance.

Tom

Dan Carreira wrote:

I’ve been trying to find a math library capable of linear algebra matrix
operations.

Unfortunately the MATLAB C Compiler avenue came to a dead end. I have no
source code for their proprietary libraries and so I’m looking elsewhere for
math libraries with the functionality that I need.

I’ve recently come across the LAPACK libraries and this looks promising
since they provide the source code. The only problem is that the
installation looks pretty complicated.

Q1. What I’d like to know is if anyone out there is using them in QNX4?

I’ve actually downloaded all the stuff I need from netlib, or at least I
think I have. >

Unfortunately I’m stuck with QNX4 as my OS.

Does anyone know where I’d go about getting a BLAS library for a Pentium
133Mhz QNX4 machine?

At one point in time I had all of rlab (http://rlab.sourceforge.net/)
compiled and working for QNX4 which included the netlib math libraries
(including libblas). If I remember right you just needed to get f2c
compiled (which should work out of the box) and then run it over the
fortran code.

Thomas

Check out > www.netlib.org> , there’s a number of different algebra
packages I recall, and/or surf around. At least basic algebra is not
difficult to find packages for. TNT and MAT are two examples.

If QNX6 (RTP) is an alternative, you can use the Pentium-II optimized
BLAS and Lapack code libraries for Linux to get high performance.
Tom

Dan Carreira wrote:

I’ve been trying to find a math library capable of linear algebra matrix
operations.

Unfortunately the MATLAB C Compiler avenue came to a dead end. I have no
source code for their proprietary libraries and so I’m looking elsewhere
for
math libraries with the functionality that I need.

I’ve recently come across the LAPACK libraries and this looks promising
since they provide the source code. The only problem is that the
installation looks pretty complicated.

Q1. What I’d like to know is if anyone out there is using them in QNX4?