This directory contains a set of examples and excersizes for 
RCS Tutorial "Introduction to GPU programming"
In particular, this directory explores implementation of GEMM algorithm from BLASS library using CUDA and openACC

Directory Structure
-------------------

   cuGEMM.cpp - running GEMM program using CUDA libraries
   accGEMM.c  - running GEMM program using openACC

Building the program
--------------------

   To build both programs execute make command: make
   To build cuGEMM, execute: make cuGEMM
   To build accGEMM, execute: make accGEMM
   To clean directory, : make clean

Running the program
-------------------

   To run this programs you have to request a node with GPU.
   Such node can be requested on SCC cluster, running the folloing command:
      qsh -V -l gpus=1

   To submit batch job:
      qsub -l gpus=1 cuGEMM
      qsub -l gpus=1 accGEMM

Contact Information
-------------------

   Katia Oleinik
   koleinik@bu.edu

Operating System Requirements
-----------------------------
 
   The examples presented in this directory were written for 
   SCC cluster with Linux OS. Makefile contains system-specific 
   paths, which could be edited to use on other machines.
 
   NVIDIA's GPU(s) with compute capability 2.0 or higher/
   NVIDIA CUDA Toolkit
   Portland Group compiler version 12.6 or higher

Updates
-------

   2013-06-01: First version of these examples created for
               Introduction to GPU programming tutorial