Close to Metal
From Wikipedia, the free encyclopedia
Close To Metal ("CTM" in short, originally called Close-to-the-Metal, or previously "Data Parallel Virtual Machine" and "DPVM" in research state) is an open thin hardware interface developed by the former ATI (now AMD Graphics Products Group), aimed to increase processing performance of emerging GPGPU technologies for stream processing applications by as much as eight times more than traditional 3D application programming interfaces (API) . The interface was implemented in FireStream line of stream processors (and was once renamed AMD Stream Processor).
Contents |
[edit] Overview
CTM gives developers unfettered access to the native instruction set and memory of the massively parallel computational elements in AMD Stream Processors, and its Radeon series of GPUs. Using CTM, stream processors effectively become powerful, programmable open architectures like today’s central processing units (CPUs). By opening up the architecture, CTM provides developers with the low-level, deterministic, and repeatable access to hardware that is necessary for API to develop essential high-level programming tools such as compilers, debuggers, math libraries, and application platforms . The CTM is also made to hide all graphics specific features, such as video output support.
All Radeon video cards (including R580 and later generations, R600 and RV670) and AMD FireStream stream processors (also known as AMD Stream Processors), are implemented with CTM interface. The AMD FireStream SDK was released under AMD EULA in December 2007 after the software stack was rewritten. [1] The inclusion of Compute Abstraction Layer (CAL) fine tunes the programming code via the CTM. AMD officials stated that the CAL is natural evolution to the CTM.
[edit] Open Source
The CTM interface was made open source via the Beta version of Close-to-the-Metal (CTM) project on SourceForge.net on February 2, 2007. The source code is designed to let CTM present a virtual machine abstraction for GPUs. CTM presents a thin interface to this hardware by hiding graphics-specific features of the device.
Some components in Compute Abstraction Layer, such as the Brook+, will also be made available open source at the time of official release.
[edit] See also
- CUDA, Nvidia's competing GPGPU technology for GeForce-based GPUs
- GPGPU (General-Purpose Computing on Graphics Processing Units)
- Stream programming
- Shader
- BrookGPU, a GPGPU library from Stanford University for the C programming language
- Brook+
- Lib Sh, a GPGPU library for C++
[edit] References
- ^ FireStream SDK download page and FireStream SDK EULA, retrieved December 29, 2007
- ^ AMD “Close to Metal” Technology Unleashes the Power of Stream Computing: AMD Press Release, November 14, 2006.
- ^ AnandTech report: ATI's Stream Processing & Folding@Home, September 30, 2006.
- ^ Universit¨at Dortmund, Fachbereich Mathematik research: Accelerating Double precision on GPUs (Proceedings of ASIM 2005), Dominik Goddeke, Robert Strzodka, and Stefan Turek. 18th Symposium on Simulation Technique, 2005.
- ^ TGDaily report: Nvidia activates a supercomputer in your PC, February 16, 2007.
[edit] External links
- ATI official site
- AMD official website
- ATI DPVM SIGGRAPH 2006 sketchPDF (134 KiB)
- ATI DVPM SIGGRAPH 2006 PresentationPDF (671 KiB)
- CTM Guide - CTI Technical Reference ManualPDF (866 KiB)
- AMD Close-to-the-Metal (CTM) open source project site
|
|||||||||||||||||||||||||||||||||||
