OpenACC


OpenACC for open accelerators is a programming standard for parallel computing developed by Cray, CAPS, Nvidia and PGI The standard is designed to simplify parallel programming of heterogeneous CPU/GPU systems1

Like in OpenMP, the programmer can annotate C, C++ and Fortran source code to identify the areas that should be accelerated using compiler directives and additional functions2 Like OpenMP 40 and newer, code can be started on both the CPU and GPU

OpenACC members have worked as members of the OpenMP standard group to merge into OpenMP specification to create a common specification which extends OpenMP to support accelerators in a future release of OpenMP34 These efforts resulted in a technical report5 for comment and discussion timed to include the annual Supercomputing Conference November 2012, Salt Lake City and to address non-Nvidia accelerator support with input from hardware vendors who participate in OpenMP6

At ISC’12 OpenACC was demonstrated to work on Nvidia, AMD and Intel accelerators, without performance data7

In November 12, 2012, at the SC12 conference, a draft of the OpenACC version 20 specification was presented8 New suggested capabilities include new controls over data movement such as better handling of unstructured data and improvements in support for non-contiguous memory, and support for explicit function calls and separate compilation allowing the creation and reuse of libraries of accelerated code OpenACC 20 was officially released in June 20139

Version 25 of the specification was released in October 201510

Contents

  • 1 Compiler support
  • 2 Usage
    • 21 Directives
    • 22 Runtime API
  • 3 See also
  • 4 References
  • 5 External links

Compiler supportedit

Support of OpenACC is available in commercial compilers from PGI from version 126, and for Cray hardware only Cray 711

OpenUH12 is an Open64 based open source OpenACC compiler supporting C and FORTRAN, developed by HPCTools group from University of Houston

OpenARC13 is an open source C compiler developed at Oak Ridge National Laboratory to support all features in the OpenACC 10 specification An experimental14 open source compiler, accULL, is developed by the University of La Laguna C language only15

GCC support for OpenACC was slow in coming16 A GPU-targeting implementation from Samsung was announced in September 2013; this translated OpenACC 11-annotated code to OpenCL14 The announcement of a "real" implementation followed two months later, this time from NVIDIA and based on OpenACC 2017 This sparked some controversy, as the implementation would only target NVIDIA's own PTX assembly language, for which no open source assembler or runtime was available1819 Experimental support for OpenACC/PTX did end up in GCC as of version 512021

Usageedit

In a way similar to OpenMP 3x on homogeneous system or the earlier OpenHMPP, the primary mode of programming in OpenACC is directives22 The specifications also include a runtime library defining several support functions To exploit them, user should include "openacch" in C or "openacc_libh" in Fortran;23 and then call acc_init function

Directivesedit

OpenACC defines an extensive list of pragmas directives,2 for example:

#pragma acc parallel #pragma acc kernels

Both are used to define parallel computation kernels to be executed on the accelerator, using distinct semantics2425

#pragma acc data

Is the main directive to define and copy data to and from the accelerator

#pragma acc loop

Is used to define the type of parallelism in a parallel or kernels region

#pragma acc cache #pragma acc update #pragma acc declare #pragma acc wait

Runtime APIedit

There are some runtime API functions defined too: acc_get_num_devices, acc_set_device_type, acc_get_device_type, acc_set_device_num, acc_get_device_num, acc_async_test, acc_async_test_all, acc_async_wait, acc_async_wait_all, acc_init, acc_shutdown, acc_on_device, acc_malloc, acc_free

OpenACC generally takes care of work organisation for the target device however this can be overridden through the use of gangs and workers A gang consists of workers and operates over a number of processing elements as with a workgroup in OpenCL

See alsoedit

  • C++ AMP
  • OpenCL
  • OpenHMPP
  • OpenMP

Referencesedit

  1. ^ "Nvidia, Cray, PGI, and CAPS launch 'OpenACC' programming standard for parallel computing" The Inquirer 4 November 2011 
  2. ^ a b "OpenACC standard version 20" PDF OpenACCorg Retrieved 14 January 2014 
  3. ^ "How does the OpenACC API relate to the OpenMP API" OpenACCorg Retrieved 14 January 2014 
  4. ^ "How did the OpenACC specifications originate" OpenACCorg Retrieved 14 January 2014 
  5. ^ "The OpenMP Consortium Releases First Technical Report" OpenMPorg 5 November 2012 Retrieved 14 January 2014 
  6. ^ "OpenMP at SC12" OpenMPorg 29 August 2012 Retrieved 14 January 2014 
  7. ^ a b "OpenACC Group Reports Expanding Support for Accelerator Programming Standard" HPCwire 20 June 2012 Retrieved 14 January 2014 
  8. ^ "OpenACC Version 20 Posted for Comment" OpenACCorg 12 November 2012 Retrieved 14 January 2014 
  9. ^ "OpenACC 20 Spec | wwwopenaccorg" wwwopenaccorg Retrieved 2016-03-23 
  10. ^ "OpenACC Standards Group Announces Release of the 25 Specification; Member Vendors Add Support for ARM & x86 as Parallel Devices | wwwopenaccorg" wwwopenaccorg Retrieved 2016-03-22 
  11. ^ "OpenACC Standard to Help Developers to Take Advantage of GPU Compute Accelerators" Xbit laboratories 16 November 2011 Retrieved 14 January 2014 
  12. ^ "OpenUH Compiler" Retrieved 4 March 2014 
  13. ^ "OpenARC Compiler" Retrieved 4 November 2014 
  14. ^ a b Larabel, Michael 30 September 2013 "GCC Support Published For OpenACC On The GPU" Phoronix 
  15. ^ "accULL The OpenACC research implementation" Retrieved 14 January 2014 
  16. ^ Larabel, Michael 4 December 2012 "OpenACC Still Not Loved By Open Compilers" Phoronix 
  17. ^ Larabel, Michael 14 November 2013 "OpenACC 20 With GPU Support Coming To GCC" Phoronix 
  18. ^ Larabel, Michael 15 November 2013 "NVIDIA, Mentor Graphics May Harm GCC" Phoronix 
  19. ^ Larabel, Michael 21 November 2013 "In-Fighting Continues Over OpenACC In GCC" Phoronix 
  20. ^ https://gccgnuorg/wiki/OpenACC
  21. ^ Schwinge, Thomas 15 January 2015 "Merge current set of OpenACC changes from gomp-4_0-branch" gcc Mailing list gccgnuorg Retrieved 15 January 2015 
  22. ^ "Easy GPU Parallelism with OpenACC" DrDobb's 11 June 2012 Retrieved 14 January 2014 
  23. ^ "OpenACC API QuickReference Card, version 10" PDF NVidia November 2011 Retrieved 14 January 2014 
  24. ^ "OpenACC Kernels and Parallel Constructs" PGI insider August 2012 Retrieved 14 January 2014 
  25. ^ "OpenACC parallel section VS kernels" CAPS entreprise Knowledge Base 3 January 2013 Retrieved 14 January 2014 

External linksedit

  • http://wwwopenaccorg/
  • Usage example from NVIDIA: part1, part2


OpenACC Information about

OpenACC

OpenACC
OpenACC

OpenACC Information Video


OpenACC viewing the topic.
OpenACC what, OpenACC who, OpenACC explanation

There are excerpts from wikipedia on this article and video



Random Posts

Social Accounts

Facebook Twitter VK
Copyright © 2014. Search Engine