Page 1 of 1
Error while running make for Octave on Fatdog7.02
Posted: Wed 05 Jul 2017, 21:55
by PappyPuppy
Fortran files such as genbet.f, genmul.f - generally, it seems that it's files that have real constants defined as eg. 87.1 rather than 87.0. I am perhaps guessing (based on some reading of posts in other threads on other forums), that perhaps it's the GMP library (multi arithmetic library). My version is circa 2013 - 5.1.3. I am thinking of upgrading this to about 6.1.0 or 6.1.2 or thereabouts. I am wondering if anyone has their GMP libs - a bunch of library files libgmpxxxx.a/la/so or whatever already upgraded, and had no problems with gcc builds that involve calls to gfortran (to build .f files or .for files).
I guess I've already downloaded the tar files for these later gmp's so I'll try it when I have time in a few days, just thought I'd throw this out there.
Not a big deal - I have Octave working on Windows right now. At least I can use it.
I'm using GNU gcc 4.8.3
Re: Error while running make for Octave on Fatdog7.02
Posted: Thu 06 Jul 2017, 14:10
by Moose On The Loose
PappyPuppy wrote:Fortran files such as genbet.f, genmul.f - generally, it seems that it's files that have real constants defined as eg. 87.1 rather than 87.0. I am perhaps guessing (based on some reading of posts in other threads on other forums), that perhaps it's the GMP library (multi arithmetic library). My version is circa 2013 - 5.1.3. I am thinking of upgrading this to about 6.1.0 or 6.1.2 or thereabouts. I am wondering if anyone has their GMP libs - a bunch of library files libgmpxxxx.a/la/so or whatever already upgraded, and had no problems with gcc builds that involve calls to gfortran (to build .f files or .for files).
I guess I've already downloaded the tar files for these later gmp's so I'll try it when I have time in a few days, just thought I'd throw this out there.
Not a big deal - I have Octave working on Windows right now. At least I can use it.
I'm using GNU gcc 4.8.3
I use Octave on 528 and it appears to be working correctly.
For those who don't know:
Octave is a really great scripting language for dealing with arrays of numbers as matrixes and vectors. It also works well for doing things with arrays as just arrays.
Code: Select all
octave:1>
octave:1> AA=1:5; # Make 1...5
octave:2> AA
AA =
1 2 3 4 5
octave:3> # see nice numbers
octave:3> AA .* 3 # .* means element by element
ans =
3 6 9 12 15
octave:4>
It is very like matlab but in a fair few ways better. Perhaps the most important is that because it is free (an in liberty) it works perfectly well without a connection to a license server.
to make your basic graph of X,Y you do
plot(X,Y)
It doesn't take long to pick it up and use it for stuff.
I believe it's the GMP library
Posted: Fri 07 Jul 2017, 12:35
by PappyPuppy
I am using an older GMP, perhaps the newer one will have the fix I need.
Yes, I do long simulations with Octave's tools and it works great! The plots are great - you can run many plots in one Figure using the subplot command:
For 3, use subplot(3, 1, 1), subplot (3, 1, 2), etc..
Re: Error while running make for Octave on Fatdog7.02
Posted: Fri 07 Jul 2017, 12:51
by PappyPuppy
So you are using GNU C++ 5.28? I wonder what GMP libraries you have. GMP goes from about 5.1 up to 6.0 and beyond. I didn't know that GNU C++ had a 5.28? I thought it went from 5.20 to 5.30. I guess it has versions in-between.
Moose On The Loose wrote:PappyPuppy wrote:Fortran files such as genbet.f, genmul.f - generally, it seems that it's files that have real constants defined as eg. 87.1 rather than 87.0. I am perhaps guessing (based on some reading of posts in other threads on other forums), that perhaps it's the GMP library (multi arithmetic library). My version is circa 2013 - 5.1.3. I am thinking of upgrading this to about 6.1.0 or 6.1.2 or thereabouts. I am wondering if anyone has their GMP libs - a bunch of library files libgmpxxxx.a/la/so or whatever already upgraded, and had no problems with gcc builds that involve calls to gfortran (to build .f files or .for files).
I guess I've already downloaded the tar files for these later gmp's so I'll try it when I have time in a few days, just thought I'd throw this out there.
Not a big deal - I have Octave working on Windows right now. At least I can use it.
I'm using GNU gcc 4.8.3
I use Octave on 528 and it appears to be working correctly.
For those who don't know:
Octave is a really great scripting language for dealing with arrays of numbers as matrixes and vectors. It also works well for doing things with arrays as just arrays.
Code: Select all
octave:1>
octave:1> AA=1:5; # Make 1...5
octave:2> AA
AA =
1 2 3 4 5
octave:3> # see nice numbers
octave:3> AA .* 3 # .* means element by element
ans =
3 6 9 12 15
octave:4>
It is very like matlab but in a fair few ways better. Perhaps the most important is that because it is free (an in liberty) it works perfectly well without a connection to a license server.
to make your basic graph of X,Y you do
plot(X,Y)
It doesn't take long to pick it up and use it for stuff.
Re: Error while running make for Octave on Fatdog7.02
Posted: Fri 07 Jul 2017, 13:50
by Moose On The Loose
PappyPuppy wrote:So you are using GNU C++ 5.28? I wonder what GMP libraries you have. GMP goes from about 5.1 up to 6.0 and beyond. I didn't know that GNU C++ had a 5.28? I thought it went from 5.20 to 5.30. I guess it has versions in-between.
Perhaps I should have been clearer.
Puppy version = 528
I've made 4 pets to make it work.
Which one should I investigate further to find what you are asking for?
gfortran-4.2.2.pet
glpk-4.40-i486.pet
gnuplot-4.2.5.pet
octave-3.2.3-i486.pet
[/code]
Sorry, I didn't notice your post
Posted: Sat 29 Jul 2017, 18:04
by PappyPuppy
And on RHEL Linux 7.2. But I still want to answer your post, ie. try what you said, I'll get back to it soon. In the meantime, I just need to use Octave to do some work.
I've installed configured, made, and installed gmp 6-1.0
Posted: Sat 29 Jul 2017, 18:53
by PappyPuppy
I want to go through the ./configure process for Fatdog and install some more stuff, before running make again. The config.log should help.