The PGI CUDA C compiler for x86 platforms allows developers using CUDA to compile and optimize their CUDA applications to run on x86-based workstations, servers and clusters with or without an NVIDIA GPU accelerator. When run on x86-based systems without a GPU, PGI CUDA C applications uses multiple cores and the streaming SIMD (Single Instruction Multiple Data) capabilities of Intel and AMD CPUs for parallel execution.
With PGI CUDA C compilers, CUDA developers now can have a common code path for both NVIDIA GPU and x86 platform.
PGI CUDA C for multi-core X86
The PGI CUDA C compiler implements the current NVIDIA CUDA C language for GPUs, and it will closely track the evolution of CUDA C moving forward. PGI CUDA C for x86 implementation is proceeding in phases, the first release is available now with most CUDA C functionality. A performance release is scheduled for Q4 2011.
- Processes CUDA C/C++ as a native parallel programming language for multi-core x86
- Reconstitutes the chevron syntax into parallel/vector loops
- Uses multiple cores and SSE/AVX instructions to effect parallel execution
- Executes each CUDA thread block using a single host core
- All traditional PGI X86-64 optimizations are supported for Intel/AMD host code
For more information
- PGI CUDA -X86 home page
- Running CUDA Code Natively on x86 Processors by Rob Farber, DrDobbs.com, August 2011
- PGI CUDA-X86 frequently asked questions (FAQ)
- Developers can register for a free 15 day trial.
- PGI Accelerators are available for purchase from a network of world-wide resellers
CUDA X86 is part of the PGI Accelerator product and is supported on Linux, Mac OS X and Windows.