Search This Blog

Saturday, 20 June 2015

CudaFortran 15.5 on Ubuntu 12.10

First run

sudo apt-get update


Then check that you have Linux Standard Base with

lsb_release

if not run

sudo apt-get update 
sudo apt-get install lsb-core

Also make sure you have a new version of a gcc compiler, check it with

gcc -v

If you do not have it, get it with

sudo apt-get install build-essential

Create a PGI account at


Log in at


If you haven't acquire the license go to


If you already have a license then just log in and go to

and download the software that you need.

Unpack the PGI software. In the instructions that follow, replace <tarfile> with the name of the file that you downloaded.

Use the following command sequence to unpack the tar file into a temporary directory before installation.

mkdir /tmp/pgi % mv
mv <tarfile>.tar.gz /tmp/pgi
cd /tmp/pgi
tar xpfz <tarfile>.tar.gz

Then change the unpack file to another carpet and open it

Get in to the carpet and run

sudo ./install

Accept and say yes to everything (recommended).

You can also check in the manual the steps for the installation of the Cluster Development Kit if you need it.

Then open your .bash file and add
For 64bit

export PGI=/opt/pgi
export PATH=/opt/pgi/linux86-64/15.5/bin:$PATH
export MANPATH=$MANPATH:/opt/pgi/linux86-64/15.5/man
export LM_LICENSE_FILE=$LM_LICENSE_FILE:/opt/pgi/license.dat

For 32bit

export PGI=/opt/pgi;
export PATH=/opt/pgi/linux86/15.5/bin:$PATH;
export MANPATH=$MANPATH:/opt/pgi/linux86/15.5/man;
export LM_LICENSE_FILE=$LM_LICENSE_FILE:/opt/pgi/license.dat;

Verify the release number of the installed software with:

pgfortran -V

To check your ID run

lmutil lmhostid

Choose one of the next two options:

For permanent license:

Generate a Permanent License Key

1. Click Create permanent keys.

2. Click the PIN associated with the product for which you wish to generate license keys. If you don't see any PINs listed, you first need to tie one to your account: a. Obtain your PIN code from your original PGI order confirmation email, in the section labeled: "SHARING ADMINISTRATION" or contact license@pgroup.com. b. Click the link: Tie a PIN to this account. c. Enter the 16-digit PIN code and follow the instructions provided.
3. Click License keys to generate the keys.

Trial license:

Generate a Trial License Key
1. Click Create trial keys.
2. Accept the terms of the agreement.
3. Enter any one of the FlexNet hostids detected, exactly as it appears in the message that is displayed during installation or when you issue the command: lmutil lmhostid 4. Click the Generate License keys button.”

Install License Keys Once you have generated your trial or permanent license keys, copy and paste them into the file: /opt/pgi/license.dat, substituting the appropriate installation directory path if you have not installed in the default /opt/pgi directory. The license.dat depends on the product you are installing. However, each license.dat file looks somewhat similar to the following sample file from a PGI Workstation installation: SERVER <hostname> 0123456789ab 27000 DAEMON pgroupd PACKAGE PGI2015-<PGI_PIN>pgroupd<suscription end date>

In your license file:

<hostid> should match the hostid you submitted above when you generated your license keys. You should see it on the SERVER line.

If necessary, you can enter or edit the <hostname> entry manually, but you cannot edit the <hostid>
entry or you will invalidate the license keys.

Go to the file

/opt/pgi/linux86-64/15.5/bin

And copy the file lmgrd.rc to the carpet init.d


sudo cp lmgrd.rc ~/../../etc/init.d/lmgrd

Create a carpet /usr/tmp. In the root directory

sudo mkdir tmp

Then run the license daemon

cd $PGI/linux86-64/15.5/bin/sudo ./lmgrd.rc start

Finally, try to run a code with

pgfortran program.cuf

You can check the documentation at


Some of the paragraph are taken from:


Where you can find more detail about the installation.



Saturday, 6 June 2015

Dear all,

I like to post some of the work my colleagues and I have developed in my class of Software Engineering.

It is a calculator in C++, It uses a DLL, a wrapper and a CLI form as a user interface. It also include a Unit Test.

We worked under agile system and I can say that it works very fine.

This is the source code

https://github.com/ivonneleonor/Calculator-C-

Thanks to Afolabi, Antreas and Norah for their patient and contribution.

Thursday, 11 September 2014

Parallel solvers on CUDA

Hello everybody:

A year ago I coded some parallel algorithms called solvers, they use them for solving matrixes with a huge amount of elements up to 1000000.  These kind of systems frequently come from Partial Differential Equation problems.

I want to share these codes with you, some of them are part of the examples which you can find directly in the CUDA toolkit but I have added my own information for testing them.  I hope they could be useful for somebody and if you find a bug please give me the feedback.

https://github.com/ivonneleonor/Parallel-solver-in-Cuda

Thank you.

Wednesday, 3 September 2014

Install Cuda 6.5 under Ubuntu 12.04

You can check first you version of Ubuntu with

lsb_release -a

and your nvidia card with

lspci |grep NVIDIA

After this you can download the driver from

http://www.nvidia.com.mx/Download/index.aspx?

To install it, follow the next steps

cd Download
sudo chmod 777 -R NVIDIA-Linux-x86_64-340.32.run

if you have 64 linux version like me, or you have to choose your flavor.

Install all these libraries

sudo apt-get install freeglut3-dev build-essential libx11-dev libxmu-dev libxi-dev libgl1-mesa-glx libglu1-mesa libglu1-mesa-dev libglu1-mesa glew-utils mesa-utils

Modify the next file

sudo gedit /etc/modprobe.d/blacklist.conf

write down at the end of the file the next list

blacklist amd76x_edac
blacklist vga16fb
blacklist nouveau
blacklist rivafb
blacklist nvidiafb
blacklist rivatv

save and close the file and then type:

sudo apt-get remove --purge nvidia*
sudo apt-get remove --purge xserver-xorg-video-nouveau
sudo reboot

After that

CTRL ALT F5

Enter your login and pass and then type

sudo lightdm stop

enter your pass again.
After this the driver has to start, choose yes for everything except for the last question about the Xorg file.

When this finishes, get in to the graphic mode with

sudo lightdm start

If everything goes right you can go for CUDA.

Download 6.5 version at

https://developer.nvidia.com/cuda-downloads?sid=589535

I got the file
cuda_6.5.14_linux_64.run

in the Downloads archive.

Again type

CTRL ALT F5

Enter your login and pass and then type

sudo lightdm stop

chmod 777 -R cuda_6.5.14_linux_64.run

and sudo sh chmod 777 -R cuda_6.5.14_linux_64.run

Accept all the options

sudo lightdm start

Run a terminal and type

sudo gedit .bashrc

for 32 bits include these lines:
export PATH=/usr/local/cuda-6.5/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-6.5/lib:$LD_LIBRARY_PATH

and for 64 bits:
export PATH=/usr/local/cuda-6.5/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-6.5/lib:/usr/local/cuda-6.5/lib64:$LD_LIBRARY_PATH

and save the changes.

There you go.

Sunday, 18 May 2014

Tutorial para instalar cuda 5.0 en thinkpad W520

Hola, les dejo el link del tutorial.

http://www.ubuntu-es.org/node/176006#.U3izwHVdW0w