Difference between revisions of "Applications/Openuh"

From HPC
Jump to: navigation , search
m (Pysdlb moved page Openuh to Applications/Openuh without leaving a redirect)
m
Line 9: Line 9:
 
==Usage==
 
==Usage==
  
'''Important''': This is a module installed to allow testing of the GPU Nodes, if you need this as a production module contact the Viper Team at [mailto:help@hull.ac.uk?Subject=Openuh/3.1.0 help@hull.ac.uk].
+
'''Important''': This is a module installed to allow testing of the GPU Nodes, if you need this as a production module contact the Viper Team at [https://support.hull.ac.uk/ Topdesk] .
  
  
Line 21: Line 21:
 
===Compiler Optimizations===
 
===Compiler Optimizations===
  
At the -O2 and -O3 optimization levels the compilers will use the IA-64 software pipelining instructions.  At the -O3 level a more powerful data dependence analysis will  
+
At the -O2 and -O3 optimization levels the compilers will use the IA-64 software pipe-lining instructions.  At the -O3 level a more powerful data dependence analysis will  
 
be performed.
 
be performed.
  
Line 33: Line 33:
 
* Dead code elimination  
 
* Dead code elimination  
 
* Expression simplification  
 
* Expression simplification  
* Common subexpression elimination
+
* Common sub expression elimination
 
* Strength reduction
 
* Strength reduction
  

Revision as of 15:29, 22 July 2020

Application Details

  • Description: OpenUH is an Open64 based open source OpenACC compiler supporting C and FORTRAN, developed by HPCTools group from University of Houston.
  • Version: 3.1.0
  • Module: openuh/3.1.0
  • Licence: Github, open source

Usage

Important: This is a module installed to allow testing of the GPU Nodes, if you need this as a production module contact the Viper Team at Topdesk .


Language highlights

  • C compatible with gcc
  • C++ compatible g++
  • Fortran90 with Fortran95 extensions


Compiler Optimizations

At the -O2 and -O3 optimization levels the compilers will use the IA-64 software pipe-lining instructions. At the -O3 level a more powerful data dependence analysis will be performed.


Scalar Optimizations

A partial list of optimizations performed include:

  • Code motion
  • Constant propagation
  • Dead code elimination
  • Expression simplification
  • Common sub expression elimination
  • Strength reduction


Module

[username@login01 ~]$ module add openuh/3.1.0



Further Information

Icon home.png