PGI Professional Edition is a perpetual license to current and all previous releases of the PGI Fortran, C and C compilers and tools for multicore CPUs and NVIDIA Tesla GPUs, including all OpenACC, OpenMP and CUDA Fortran features. Enables development of performance-portable HPC applications with uniform source code across the most widely.
... | Last updated: 20 May, 2021 Copy Copied As an alternative to using Intel compilers for Fortran, C, and C++ programs, you can use the compilers and program development tools from PGI Compilers and Tools. By default, the PGI compilers generate code that is optimized for the compilation host. PGI CompilingSeveral versions of PGI compilers and tools are installed as modules under the /nasa system directory on Pleiades and Endeavour. To view available versions, run: The following command line will load version 18.10: Note: Starting with the PGI compiler 2018, two different code generators (PGI-proprietary vs LLVM-based) are included in each package. The default code generator for 2018 versions is the PGI-proprietary code generator and for 2019 and later versions is the LLVM-based generator. The PGI 2019 compilers using LLVM are not link-compatible with PGI 2018 and prior releases. Using the PGI CompilersPGI provides multiple commands for different languages or functions. In 18.10 and earlier versions, the following are available.
Note: By default, the PGI compilers generate code that is optimized for the type of on which compilation is performed (the compilation host). Be aware that the processors are forward-compatible, but not backward-compatible. Thus a code compiled and optimized on a newer-generation processor will not necessarily execute correctly on previous-generation processors. If you want to build an executable that targets a specific processor type, use the -tp flag:
TIP: Using the -tp=sandybridge-64,haswell-64,skylake-64 option will generate a single executable in which the code is optimized for the Intel Sandy Bridge, Haswell, and Skylake architectures. The choice of which optimized copy to execute is made at run time and depends on the machine executing the code. You must load comp-pgi/18.4 or later versions in order to use -tp= skylake-64. PGI recommends that for best performance on processors which support SSE instructions (including all NAS processor types), use pgfortran, even for FORTRAN 77 code, use the -fastsse option. For more information about the PGI compilers, see man pgfortran, man pgcc, and man pgc++ or use the pgfortran -help, pgcc -help, or pgc++ -help commands. Information about the PGI debugger and performance analysis tool can be found using the commands pgdbg -help and pgprof -help. |
PGI Visual Fortran – Parallel Fortran Compilersand Tools for Microsoft Windows. PGI Visual Fortran® (PVF®) bringsthe PGI suite of high-performance 64-bit and 32-bit parallelFortran compilers to Microsoft* Windows* developers using MicrosoftVisual Studio*.
- PGI® Fortran Language Reference For PGI Visual Fortran Release 2012. While every precaution has been taken in the preparation of this document, The Portland Group® (PGI®), a wholly-owned subsidiary of STMicroelectronics, Inc., makes no. FORTRAN § @. The PGI Visual Fortran User's Guide describes the general features.
- Would you like to react to this message? Create an account in a few clicks or log in to continue.
High-performance PGI Compilers
PVF features the latest version of PGI’s native OpenMP andauto-parallel Fortran compilers. PGI compilers are supported on andcan generate fully optimized code for a broad range of popularhigh-performance computing platforms including 64-bit x64 and32-bit x86 processor-based systems. PGI Fortran compilers offerworld-class performance and features including both automatic andOpenMP 3.0 parallelization support for optimizing performance onmulti-core processors.
Visual Studio Integration
PVF includes features to enable rapid code development. PVF addssyntax coloring, Fortran intrinsics tips and keyword completion tothe Visual Studio editor. PVF includes a complete Fortran buildenvironment for Visual Studio with Fortran specific features suchas automatic build dependency management. PVF is interoperable withMicrosoft Visual C++ so Fortran sub-programs can call C functionsand vice versa. PVF includes a number of DVF/CVF compatibilityfeatures including full support for the Win32 API. Sample projectsinclude building Fortran dialog boxes, console and Windowsapplications, DLLs and static libraries. PVF augments the VisualStudio debugger with a custom debug engine that provides Fortranlanguage specific debugging capability.
Here are some key Features of “PGI Visual Fortran”:
Visual Studio Integration:
– Fortran syntax coloring
– Fortran intrinsics tips
– Keyword completion
– Fortran debug engine
– Automatic dependency analysis
– Interoperable with Microsoft Visual C++
– Sample projects
– Bundled Visual Studio 2010 IDE
– Bundled ACML high-performance math library
– Supports multi-threaded execution with Intel MKL 10.1 andlater
– Optional IMSL Fortran numerical library available
– Build, launch and debug Microsoft MPI (MSMPI) applications
Fortran Compilers:
– Native OpenMP and auto-parallel Fortran 95/03 compiler withoptional support for PGI Accelerator directives and CUDAextensions
– Full 64-bit support on AMD64, Intel 64 and NVIDIA CUDA-enabledGPUs
– PGI Unified Binary technology—combine code optimized for multipleAMD64 processors, Intel 64 processors or NVIDIA GPUs into a singleexecutable file
– Full 32-bit support and optimization for x86/x87 targets
– Auto-parallelization
– OpenMP 3.0 parallelization
– Inter-language calling
– IEEE arithmetic options
– Vectorization/Optimization directives
– Integrated cpp pre-processor
– Fast compile times
– Compile-time optimization listings
– Annotated assembly code listings
– Comprehensive support for Fortran 2003
Pgfortran
Optimizations:
– AMD64 and Intel 64 cross-target optimization
– Interprocedural Analysis (IPA)
– Scalar SSE code generation
– Function inlining
– Loop fusion
– Profile feedback optimization
– Memory hierarchy and memory allocation optimization
– Huge pages support
– Instruction scheduling
– Invariant code and conditional hoisting
– Common sub-expression elimination
– Global constant propagation
– Global dependence analysis
– Global flow analysis
– Integrated local, global and vector register allocation
– Induction variable analysis
– Load/store analysis
– Tail recursion elimination
– Scalar expansion, scalar replacement
– Value propagation, strength reduction
– Dead code elimination
Vector/Parallel Transformations:
– State-of-the-art dependence analysis
– Nested loop auto-parallelization
– Nested loop vectorization
– Loop interchange
– Loop splitting, unroll & jam
– Iteration peeling
– Aligned access optimizations
– Scalar promotion
– Temporary vector creation
– Stripmining, Cache Tiling
– Concurrent call support
– Inline concurrent code segments
– Vector SSE code generation
– Software prefetching
Parallel Debug Engine:
– 64-bit and 32-bit targets
– Debug Fortran and Visual C++
– Debug single and multi-thread, parallel OpenMP, multi-threadMSMPI and hybrid MSMPI+OpenMP programs
– Variable roll-over
– One touch breakpoint setting
– Step into, over, out of functions
– Attach to running processes
– Traceback, Log files, Help
– One touch symbolic display
– Multiple format display of values or strings
– Track register states
Setup+Crack Size: 470 + 602 MB