Opencl
Open Computing Language (OpenCL) is a framework for writing programs that execute across heterogeneous platforms consisting of key processing units (CPUs), graphics processing units (GPUs), digital dot processors (DSPs), or field-programmable gate arrays (FPGAs). Heterogeneous computing refers to systems that purpose to a greater extent than than i sort of processor or cores for high functioning or liberate energy efficiency.
OpenCL views a computing organization every bit consisting of a release of compute devices (GPUs, CPUs, FPGAs) attached to a host processor (a CPU). It defines a C-like linguistic communication for writing programs. Functions executed on an OpenCL device are called /kernels/. A unmarried compute device typically consists of several compute units, which inwards plough incorporate multiple processing elements (PEs). A unmarried meat execution tin flame run on all or many of the PEs inwards parallel.
In add-on to its C-like programming language, OpenCL defines an API that allows programs running on the host to launch kernels on the compute devices in addition to create out device memory. Programs inwards the OpenCL linguistic communication are intended to move compiled at run-time, in addition to hence that OpenCL-using applications are portable betwixt implementations for diverse host devices.
OpenCL is an opened upwards measure maintained past times the non-profit technology scientific discipline consortium Khronos Group. Conformant implementations are available from Altera, AMD, Apple, ARM, Creative, IBM, Imagination, Intel, NVIDIA, Qualcomm, Samsung, Vivante, Xilinx, in addition to ZiiLABS. Although OpenCL provides an choice to CUDA, it has only about back upwards from NVDIA.
OpenCL is supported past times Android, FreeBSD, Arch Linux, Linux, macOS, Windows operating systems.
A 2015 paper, Bridging OpenCL in addition to CUDA: a comparative analysis in addition to translation, too provides CUDA-OpenCL translation every bit good every bit OpenCL-CUDA translation.
These translators, however, create non render industry-grade in addition to ruggedized implementations.
Recently an OpenCL port of Caffee was made available. This Caffe port was shown/evaluated for AMD chipsets, but it should too utilise for ARM platforms that back upwards OpenCL.
OpenCL views a computing organization every bit consisting of a release of compute devices (GPUs, CPUs, FPGAs) attached to a host processor (a CPU). It defines a C-like linguistic communication for writing programs. Functions executed on an OpenCL device are called /kernels/. A unmarried compute device typically consists of several compute units, which inwards plough incorporate multiple processing elements (PEs). A unmarried meat execution tin flame run on all or many of the PEs inwards parallel.
In add-on to its C-like programming language, OpenCL defines an API that allows programs running on the host to launch kernels on the compute devices in addition to create out device memory. Programs inwards the OpenCL linguistic communication are intended to move compiled at run-time, in addition to hence that OpenCL-using applications are portable betwixt implementations for diverse host devices.
OpenCL is an opened upwards measure maintained past times the non-profit technology scientific discipline consortium Khronos Group. Conformant implementations are available from Altera, AMD, Apple, ARM, Creative, IBM, Imagination, Intel, NVIDIA, Qualcomm, Samsung, Vivante, Xilinx, in addition to ZiiLABS. Although OpenCL provides an choice to CUDA, it has only about back upwards from NVDIA.
OpenCL is supported past times Android, FreeBSD, Arch Linux, Linux, macOS, Windows operating systems.
CUDA to OpenCL translators
A paradigm implementation exists for CUDA to OpenCL translator.A 2015 paper, Bridging OpenCL in addition to CUDA: a comparative analysis in addition to translation, too provides CUDA-OpenCL translation every bit good every bit OpenCL-CUDA translation.
These translators, however, create non render industry-grade in addition to ruggedized implementations.
OpenCL for ML
OpenCL back upwards is soundless underwhelming for deep learning, but it is getting better.
Recently an OpenCL port of Caffee was made available. This Caffe port was shown/evaluated for AMD chipsets, but it should too utilise for ARM platforms that back upwards OpenCL.
0 Response to "Opencl"
Post a Comment