|
AMD Core Math Library for Graphic Processors (ACML-GPU) provides an ATI Stream-accelerated version of ACML. ACML-GPU accelerates certain routines in ACML, such as SGEMM and DGEMM, by off-loading the computation to the compatible GPUs in the system. The library dynamically decides, based on the parameters passed to the routines, whether to run the computation on the CPU or GPU, depending on which processor will yield the best performance.
ACML-GPU automatically scales its computation across multiple GPUs, if available and can take advantage of the double precision floating point hardware in the GPU on products that contain hardware DPFP support.
» Features
» System Requirements
» Downloads
ACML-GPU includes the following features:
- Automatic selection of GPU or CPU algorithms based on problem size
- Seamless migration of applications from CPU to GPU
- Selection of GPU or CPU algorithms using an environment variable
ATI Stream-accelerated routines:
The following software and hardware components are required to use ACML-GPU in the Microsoft® Windows® operating system:
- Windows XP Professional, 64-bit edition or Windows Vista®, 64-bit edition
- Supported AMD graphics card
- 2 GB of system DRAM (4 GB of DRAM is recommended)
- Appropriate 64-bit ATI Catalyst™ drivers for the supported AMD graphics card (Catalyst 9.2 or later)
- Microsoft Visual Studio 2005 SP1 C compiler
- PGI Visual Fortran, version 7.1-6 (required to build the Fortran examples)
The following software and hardware components are required to use ACML-GPU in the Linux® operating system:
- SuSE Linux® SLES 10.3 or Red Hat Linux RHEL 5.1, 64-bit installation
- Supported AMD graphics card
- 4 GB of system DRAM
- Appropriate 64-bit ATI Catalyst drivers for the supported AMD graphics card (Catalyst 9.2 or later)
- GCC/GFORTRAN, version 4.1.2
ACML-GPU runs on graphics cards supported by the ATI Stream SDK. For more information on supported graphics cards, see the ATI Stream Computing System Requirements.
|