Intel prepares to offer CT parallel programming support

Intel's CT (C for throughput) software technology, being developed for parallel programming products, will be deployed in a beta version later this year, according to Intel senior vice president Pat Gelsinger, speaking last week at the Intel Developer Forum in Beijing.

Software support products based on CT will deliver data parallel capabilities through standard C++ templates, the company said. In this way, the C++ developer community will be able to build applications that scale across hundreds of processor cores. The first product will enter beta by late 2009 and will complement other parallel programming tools available from Intel, Threading Building Blocks and Intel Parallel Studio.
In principal, CT works with any standard C++ compiler because it is a standards-compliant C++ library. When one initializes the CT library, one loads a runtime that includes the compiler, threading runtime, memory manager, essentially all the components one needs for threaded and/or vectorized code generation.

CT code is dynamically compiled, so the runtime tries to aggregate as many smaller tasks or data parallel work quanta so that it can minimize threading overhead and control the granularity according to run-time conditions. With the CT dynamic engine, users would get precise inter-procedural traces to compile, which is extremely useful in the highly modularized and indirect world of object-oriented programming.

"After 2 years of technology incubation as an Intel research project, the CT Technology team is delighted to be working very closely with Intel's data-parallel product development team," Anwar Ghuloum, chief architect of CT, said in a statement on the Intel Web site.

"Our development team is immensely talented and we are deeply engaged with this worldwide capability, which spans three continents. CT Technology is going to help transform programmer productivity and this will accelerate the arrival of a whole new wave of amazing multi-core and many-core applications," Ghuloum added.

"We believe CT represents an exciting and essential new development for the future of parallel programming and we are convinced that CT will be become one of the cornerstones of CERN's future programming environment. We are already adopting the existing software for the on-line event reconstruction of the ALICE experiment at CERN to use CT," said Professor Volker Lindenstruth, chair of computer engineering at the University of Heidelberg and head of the ALICE high-level trigger at the CERN nuclear research facility.

A white paper on CT could be found here when this story was first posted.

BY Peter Clarke
Source:EE Times Europe



Copyright 2008-2009 Daily IT News | Contact Us