As core counts and vector widths increase,
programmers are challenged to take advantage
of all the resulting processing power. This
CodeBook explores some of the optimization
techniques and models that can help developers
create code that utilizes these capabilities
for better performance. With Intel® software
development tools, applications can scale from
few cores to many cores with consistent models,
languages, tools, and techniques—whether
developing for Intel® Core™ or Intel® Xeon®
processors, or the Intel® Xeon Phi™ coprocessor.
"For James Reinders, the
path to successful parallel
programming is simple
Understanding how to restructure code to expose more parallelism is critically important to enable the best performance
on processors, GPUs, or coprocessors. For James Reinders, Intel's director of parallel programming evangelism, the path
to successful parallel programming is simple and achievable: "program with lots of threads that use vectors, while
using preferred programming languages and parallelism models." This restructuring itself will generally yield benefits.
Reinders emphasizes the value of common programming languages, models, and tools that work across processors and
coprocessors—so developers do not have to rewrite code in order to scale their applications.
Here, we'll take a brief look at some of the possibilities when using Intel processors and coprocessors. Coprocessors
extend the scaling capabilities of Intel Xeon processor-based systems by exploiting additional processor vector
capabilities or memory bandwidth.
Intel Xeon Phi coprocessor peak
Discover the potential for your
applications with in-depth
presentations on parallel
Coding the Future (video) »
Intel Xeon coprocessor peak
Free Webinar Series »
Parallelism expert James'
Reinders' blog »
Intel® Xeon Phi™ Coprocessor
For more information regarding performance and optimization choices in Intel® software products, visit: http://software.intel.com/en-us/articles/optimization-notice