AMD x86 Open64 Compiler Suite

The AMD x86 Open64 Compiler Suite is a high performance, production quality code generation tool designed for high performance parallel computing workloads. The x86 Open64 environment provides the developer the essential choices when building and optimizing C, C++, and Fortran applications targeting 32-bit and 64-bit Linux platforms.

The x86 Open64 compiler system offers a high level of advanced optimizations, multi-threading, and processor support that includes global optimization, vectorization, interprocedural analysis, feedback directed optimizations, loop transformations, and code generation which extracts the optimal performance from each x86 processor core.

The x86 Open64 Compiler Suite simplifies and accelerates development and tuning for x86, AMD64 (AMD® x86-64 Architecture), and Intel64 (Intel® x86-64 Architecture) applications. The x86 Open64 compilers create a strong foundation for building robust, high performance parallel code through MPI, OpenMP®, and auto-parallelization along with highly optimized libraries*.

Features:

• Language Standards:
• ANSI C99, ISO C++98
• Conforms to ISO/IEC 9899: 1999, Programming Languages - C standard
• Conforms to ISO/IEC 14882:1998(E), Programming Languages - C++ standard
• Fortran 77, 90, 95
• Conforms to ISO/IEC 1539-1: 1997 Programming Languages - Fortran
• Inter-language calling
• IEEE 754 floating point support
• Pragmas to control optimizations
• Platform Highlights:
• x86 32-bit and x86 64-bit code generation
• Large File support on 32-bit systems
• Byte-swapping I/O
• Vector and scalar SSE/SSE2/SSE3 code generation
• OpenMP 2.5 for shared memory models
• MPICH2 for distributed and shared memory models
• Optimized AMD Core Math Library (ACML)
• Optimization Highlights
• Global optimizations, e.g.
• Partial redundancy elimination
• Constant propagation and code motion
• Strength reduction and expression simplification
• Dead code elimination and common sub-expression elimination
• Loop-nest optimizations (enabled with '-O3'), e.g.
• Loopfusion and distribution
• Loop interchange and cache locality optimization
• Vectorization for SSE* code generation
• Software data prefetching
• Code generation and optimizations, e.g.
• Advanced register allocation
• Loop unrolling
• Instruction selection and scheduling
• Peephole optimizations
• Feedback-directed optimizations (enabled with ‘-fb_create/-fb_opt'), e.g.
• Code layout
• Feedback-directed function inlining and de-virtualization
• Feedback-guided register spilling
• Value specialization
• Inter-procedural analysis and optimization (enabled with '-ipa'), e.g.
• Function inlining and cloning
• Inter-procedural alias analysis
• Data re-layout optimizations for structure members
• Inter-procedural constant propagation and dead code elimination

What is new in this release:

• This version includes new performance optimizations, support for the AMD family 15h, better building from source, and many bugfixes.

What is new in version 4.2.4:

• Improved function inlining heuristics and enhanced inline expansion of library functions
• Enhanced framework for multi-versioning
• Improved induction expression simplification and if-merging optimization
• Improved code generation for the % operator
• Improved interprocedural analysis for indirect function calls, virtual functions, and functions with "noreturn" attribute.
• Optimized exception handling
• Optimizing processing of Fortran 90 temporary arrays
• Improved processor affinity mapping in the OpenMP and parallel runtime library
• Added support for 1GB huge pages
• Simplified floating point options
• long double implemented as 80-bit IEEE double precision
• intrinsic added to -ffastmath
• Fortran ISO_C_BINDING module
• Bug fixes in many areas including:
• GCC compatibility
• Huge page support
• Fortran support
• OpenMP
• Building from source
• Extended platform support, including RHEL 5.5 and SLES11 SP1

What is new in version 4.2.4:

• Improved function inlining heuristics and enhanced inline expansion of library functions
• Enhanced framework for multi-versioning
• Improved induction expression simplification and if-merging optimization
• Improved code generation for the % operator
• Improved interprocedural analysis for indirect function calls, virtual functions, and functions with "noreturn" attribute.
• Optimized exception handling
• Optimizing processing of Fortran 90 temporary arrays
• Improved processor affinity mapping in the OpenMP and parallel runtime library
• Added support for 1GB huge pages
• Simplified floating point options
• long double implemented as 80-bit IEEE double precision
• intrinsic added to -ffastmath
• Fortran ISO_C_BINDING module
• Bug fixes in many areas including:
• GCC compatibility
• Huge page support
• Fortran support
• OpenMP
• Building from source
• Extended platform support, including RHEL 5.5 and SLES11 SP1

What is new in version 4.2.3:

• Improved interprocedural analysis to include structure array copy optimization and array remapping optimization.
• Improved loop optimizations: loop unrolling, loop unroll and jam, triangular loops, proactive loop interchange, loop distribution, loop peeling.
• Improved redundancy elimination optimizations for stores and memory initialization; better integration of re-association and common sub-expression elimination; enhanced expression factorization.
• Improved instruction selection and addressing code generation.
• Improved vectorization.
• Extended prefetching to include arrays with inductive base addresses.
• Enhanced loop multi-versioning.
• Improved OpenMP and auto-parallelization code generation.
• Improved tuning of OpenMP and parallel runtime library functions.
• Introduced aggressive optimizations to improve scalability/bandwidth utilization of multi-core processors. (Invoked by specifying the new "-mso" flag.)
• Improved gcc compatibility, including support for more attributes, command-line options, and built-in functions.
• Expanded platform support, including SLES 11 and RHEL 5.4.
• Many bug fixes.