A summary of new features in each version of the PGI 2016 release.

PGI releases version update, as often as monthly, which may include both new features and bug fixes. The new features added with each version are listed below. Bug fixes are listed on the PGI Release 2016 Technical Problem Reports summary page.

Complete details on the new features listed here are included in the PGI Release Notes and the PGI Visual Fortran® Release Notes

PGI 2016 version 16.10 (released November 14, 2016)

  • First production release of PGI Compilers & Tools for OpenPOWER architecture CPUs
  • First PGI Community Edition release
  • PGI Accelerator™ Features and Enhancements
    • Integrated support for the production release of CUDA Toolkit 8
    • Improved support for the OpenACC cache directive
    • Improved support for structs, unions, derived types in accelerator regions
    • Extended data types supported by atomic operations
  • PGI Fortran, C, C++ compilers
    • Support for macOS Xcode 8.0
  • Plus 11 user-requested enhancements and fixes.

PGI 2016 version 16.9 (released September 8, 2016)

  • PGI Accelerator™ Features and Enhancements
    • Enabled the -ta=tesla:managed compiler option for Pascal GPUs
    • Added support for native atomic instructions on NVIDIA Pascal GPUs
  • Plus 17 user-requested enhancements and fixes

PGI 2016 version 16.7 (released August 2, 2016)

  • PGI Accelerator™ Features and Enhancements
    • Support for NVIDIA Pascal GPUs
    • Open Beta release for OpenPOWER CPUs and Tesla GPUs
    • Beta support for CUDA Toolkit 8 RC
    • Support for num_gangs, num_workers and vector_length clauses on the OpenACC Kernels construct
    • Behavior of automatic arrays and arrays included in a private loop clause inside an acc routine directive changes to be in compliance with the OpenACC specification
  • PGDBG® Debugger
    • Variable rollover added in graphical interface
  • Plus 23 user-requested enhancements and fixes

PGI 2016 version 16.5 (released May 24, 2016)

  • PGI Accelerator Features — Using the cache directive to put arrays of variable size in shared memory is now allowed
  • PGPROF® now supports CUDA 8.0 drivers in addition to CUDA 7.5
  • Open MPI updated to version 1.10.2; added installation option to disable CUDA-aware features
  • Plus 25 user-requested enhancements and fixes

PGI 2016 version 16.4 (released April 15, 2016)

  • Licensing
  • Plus 15 user-requested enhancements and fixes

PGI 2016 version 16.3 (released March 10, 2016)

  • First Windows release for 2016;
    • PGI Visual Fortran® is now supported from within Visual Studio 2015
    • Dropped support for Microsoft Visual Studio versions 2008, 2010 and 2012 by PVF®
    • Dropped support for Windows C++ in PGI Workstation™ and PGI Server
    • Dropped support for Windows version of the IMSL Fortran numerical library
  • Plus 14 user-requested enhancements and fixes

PGI 2016 version 16.1 (released February 2, 2016)

  • Comprehensive support for C++14; GCC 5.1 interoperability
  • Further optimized Fortran intrinsic functions
  • OpenACC 2.5 support
  • All new PGRPOF profiler supports both CPU and CPU+GPU profiling
  • New OpenBLAS based pre-build BLAS and LaPACK library
  • Updated pre-compiled MPICH, MVAPICH and Open MPI libraries; Open MPI now bundled with all Linux packages
  • Prebuilt netCDF and ESMF libraries
  • New license key format allows combining multiple floating licenses of varying types and seat counts into a single license file; users must retrieve new keys
  • Updated Free PGI for OS X package
  • Expanded operating system support
  • Plus 32 user-requested enhancements and fixes
Click me
Cookie Consent

This site uses cookies to store information on your computer. See our cookie policy for further details on how to block cookies.