A GPU array is a computer array in which anniversary bulge is able with a Graphics Processing Unit (GPU). By harnessing the computational ability of avant-garde GPUs via General-Purpose Computing on Graphics Processing Units (GPGPU), actual fast calculations can be performed with a GPU cluster.
Saturday, 10 December 2011
Hardware (GPU)
The accouterments allocation of GPU clusters abatement into two categories: Amalgamate and Homogeneous.
Heterogeneous
Hardware from both of the aloft IHV's can be acclimated (ATi and nVidia). Even if altered models of the aforementioned GPU are acclimated (e.g. 8800GT alloyed with 8800GTX) the gpu array is advised heterogeneous.
Homogeneous
Every distinct GPU is of the aforementioned accouterments class, make, and model. (i.e. a constant array absolute 100 8800GTs, all with the aforementioned bulk of VRAM)
Classifying a GPU array according to the aloft semantics abundantly directscomputer application development on the cluster, as altered GPUs accept altered capabilities that can be utilized
Heterogeneous
Hardware from both of the aloft IHV's can be acclimated (ATi and nVidia). Even if altered models of the aforementioned GPU are acclimated (e.g. 8800GT alloyed with 8800GTX) the gpu array is advised heterogeneous.
Homogeneous
Every distinct GPU is of the aforementioned accouterments class, make, and model. (i.e. a constant array absolute 100 8800GTs, all with the aforementioned bulk of VRAM)
Classifying a GPU array according to the aloft semantics abundantly directscomputer application development on the cluster, as altered GPUs accept altered capabilities that can be utilized
Hardware (Other)
In accession to the computer nodes and their corresponding GPUs, a fast abundant interconnect is bare in adjustment to shuttle abstracts amidst the nodes. The blazon of interconnect abundantly depends on the cardinal of nodes present. Some examples of interconnects accommodate Gigabit Ethernet and InfiniBand.
Vendors
NVIDIA provides a account of committed Tesla Preferred Partners (TPP) with the adequacy of architecture and carrying a absolutely configured GPU array application the NVIDIA® Tesla™ 20-series GPGPU processors. AMAX Advice Technologies, Dell, Hewlett-Packard and Silicon Graphics are some of the few companies that provides a complete band of GPU clusters and systems. [1]
ATI currently does not accept a account of Preferred Partners. ATI Marketing has been contacted for acceptance as this advice is actuality consistently updated
Vendors
NVIDIA provides a account of committed Tesla Preferred Partners (TPP) with the adequacy of architecture and carrying a absolutely configured GPU array application the NVIDIA® Tesla™ 20-series GPGPU processors. AMAX Advice Technologies, Dell, Hewlett-Packard and Silicon Graphics are some of the few companies that provides a complete band of GPU clusters and systems. [1]
ATI currently does not accept a account of Preferred Partners. ATI Marketing has been contacted for acceptance as this advice is actuality consistently updated
Software
Thecomputer application apparatus that are appropriate to accomplish abounding GPU-equipped machines act as one include:
Operating System
GPU disciplinarian for the anniversary blazon of GPU present in anniversary array node.
Clustering API (such as the Message Passing Interface, MPI).
Virtual OpenCL (VCL) array belvedere 1
Operating System
GPU disciplinarian for the anniversary blazon of GPU present in anniversary array node.
Clustering API (such as the Message Passing Interface, MPI).
Virtual OpenCL (VCL) array belvedere 1
Algorithm mapping
Mapping an algorithm to run a GPU arrangement is somewhat agnate to mapping an algorithm to run on a acceptable computer cluster. Example: rather than distributing pieces of an arrangement from RAM, a arrangement is disconnected up amidst the nodes of the GPU cluster.
Subscribe to:
Posts (Atom)