Vendors Boost Open-Source Tools Support

At the OSCON show, Black Duck and Intel are offering better ways for open-source developers to accelerate software development and improve their parallel processing capabilities.

PORTLAND, Ore.-Black Duck Software and Intel are the latest vendors to make substantial improvements to their open-source development tools, the latest indicators of the open-source movement's growing presence in the application development field.

Black Duck, which offers products and services for accelerating software development through the managed use of open-source software, announced July 22 the expansion to the volume of open-source software managed in the code search engine.

At the O'Reilly Open Source Conventions here, Black Duck officials said now contains 15 percent more code through the addition of C and C++ code governed by open-source and other licenses. The addition will enhance Koders' position as the search engine of choice for developers looking for reusable open-source code, methods, examples and algorithms, they said.

Black Duck bought Koders in April and has worked to enhance it for the more than 30,000 individual developers and software-centric organizations that use the code search engine.

"When we acquired Koders, our promise to that community of users and to Black Duck customers was that we'd make searches even more productive and valuable, and this is a very tangible step in that direction," said Doug Levin, Black Duck's founder and CEO.

Canonical's Mark Shuttleworth puts out a call to make desktop Linux more attractive than Apple's Mac OS.

Meanwhile, Intel is significantly upgrading its parallel processing tool, Intel Threading Building Blocks (TBB).

The chip maker announced TBB 2.1 at the show here July 22. TBB is a C++ template library available on Windows, Mac OS X and Linux.

Phil De La Zerda, director of business development for Intel Software Development Products, said TBB is a key time-saving developer tool that helps programmers implement threading into applications and improves application performance on multicore processors. TBB 2.1 makes it easier for developers to apply TBB to new uses in their software, including GUIs, artificial intelligence and network input/output, he said. TBB 2.1 will be available July 23.

"To get the best performance out of multicore processors, you have to look at parallel programming, and TBB helps you there," De La Zerda said. "TBB is the most popular abstraction for adding parallelism to C++."

TBB has undergone some significant improvements over the last year since Intel announced the open-source release of the technology at OSCON 2007. These improvements include expanded availability, greater community involvement, increased portability, and improved functionality and performance, De La Zerda said.

In terms of expanded availability, he said that in the past year several vendors have begun distributing Intel's TBB technology, including Sun Microsystems in Sun OpenSolaris, Red Hat in Fedora Core 8 and 9, Novell in OpenSUSE, Canonical in Ubuntu, TurboLinux, Red Flag, Haansoft, Miracle Linux, and VietSoftware.

Intel TBB has been ported to support other operating systems and hardware as well, including Sun Solaris on Intel Architecture, Apple Mac OS X on PowerPC and Microsoft Xbox 360, De La Zerda said.

TBB 2.1 helps parallel programming developers do more, do it faster and do it easier, he added. The technology now supports waiting tasks using tbb_thread. It also enables developers to break down work using one-, two- or three-dimensional ranges.

TBB 2.1 features improved performance using its portable affinity mechanism. And there is better performance throughout algorithms and containers, and freedom to choose any memory allocator. The new version also provides easier-to-use, improved APIs; better support for expert users using black belt hooks; and the capability to do task cancellation and exceptions, De La Zerda said.

The commercial version of TBB 2.1 is available for download July 23 here. The open-source version is available here.