Commit eeba9de0 authored by Lee, Seyong's avatar Lee, Seyong
Browse files

Update scripts.

parent 7414a6f6

OpenARCSetup.bash

+1 −0
Original line number Diff line number Diff line
@@ -4,4 +4,5 @@ module load cuda 
module load pgi

module load cudampi

export openarc=/scratch/${USER}/local/openarc

export openarcexamples=scratch/${USER}/local/openarcexamples

README.md

+51 −9
Original line number Diff line number Diff line
@@ -19,14 +19,17 @@ ENVIRONMENT SETUP 
-------------------------------------------------------------------------------

* Module Setting

	- Load the following modules: java, cuda, pgi, and cudampi

	(Run the commands below.)

	//Run the commands below:

	$ module load java

	$ module load cuda

	$ module load pgi

	$ module cudampi



	//Or run the following command:

	$ source /home01/kedu01/shared/OpenARCSetup.bash



* OpenARC Environment Setting

	- Set the following environment variables:

	- Set the following environment variables (targeting NVIDIA CUDA GPUs):

	$ export OPENARC_ARCH=0

	//Optional setting to use CUDA Unified Memory

	$ export OPENARCRT_UNIFIEDMEM=1
@@ -38,9 +41,12 @@ ENVIRONMENT SETUP 
	$ /home01/kedu01/shared/installOpenARC.bash



	- Set environment variable, openarc, to the OpenARC install directory:

	$ export openarc=$HOME/local/openarc

	$ export openarc=/scratch/${USER}/local/openarc



* Additional Settings for example tests

	- Set environment variable, openarcexamples to the root directory containing the OpenARC examples 

	$ export openarcexamples=scratch/${USER}/local/openarcexamples



	- If an input OpenACC program uses OpenMP, set environment variable, OMP_NUM_THREADS.

	$ export OMP_NUM_THREADS=10
@@ -48,18 +54,54 @@ ENVIRONMENT SETUP 
	$ export mpiinclude=/apps/compiler/pgi/linux86-64/19.1/cudampi/10.0/openmpi/3.1.0/include



-------------------------------------------------------------------------------

EXAMPLE TESTING

COMPILE & RUN EXAMPLES

-------------------------------------------------------------------------------

* To compile and run the examples in the openarcexamples directory, visit each example directory and run the following steps.



	

	- Run the OpenARC compiler (e.g., matmul example):

	//Go to openarcexamples directory. 

	$ cd  ./matmal

	$ ./O2GBuild.bash

	$ cd  ${openarcexamples}/matmal

	$ ./O2GBuild.script

	

	- Compile OpenARC-generated output files:

	$ make



	- To submit the output program:

	- To submit a job to run the output program:

	$ cd bin; sbatch gpu.sh



-------------------------------------------------------------------------------

EXPERIMENTS

-------------------------------------------------------------------------------

* Experiment1: learn how to compile and run the matmul benchmark.

	- Task1: compile and run matmul

	$ cd ${openarcexamples}/matmul

	//Run the OpenARC compiler.

	$ ./O2GBuild.script

	//Check generated output files.

	$ vi ./cetus_output/openarc_kernel.cu

	$ vi ./cetus_output/matmul.cpp

	//Compile the generated output files.

	$ make

	//Submit a job to run the output program.

	$ cd bin; sbatch gpu.sh 



	- Task2: learn how to use an OpenARC commandline option: showInternalAnnotations

	$ cd ${openarcexamples}/matmul

	//Open the OpenARC configuration file and set showInternalAnnotations to 1 (e.g., showInternalAnnotations=1)

	$ vi ./matmul/openarcConf.txt

	//Run O2GBuild.script script again.

	$ ./O2GBuild.script

	//Check  generated output files

	$ vi ./cetus_output/openarc_kernel.cu

	$ vi ./cetus_output/matmul.cpp

	//Repeat the above steps by changing showInternalAnnotations up to 3.



	- Task3: learn how to use commandline options for caching optimizations: shrdArryCachingOnTM

	$ cd ${openarcexamples}/matmul

	//Open the OpenARC configuration file and set shrdArryCachingOnTM to 1 (e.g., shrdArryCachingOnTM=1)

	$ vi ./matmul/openarcConf.txt

	//Run O2GBuild.script script again.

	$ ./O2GBuild.script

	//Check  generated output files

	$ vi ./cetus_output/openarc_kernel.cu

	$ vi ./cetus_output/matmul.cpp