Latest question:
How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!







By Keith Reid
"Mercury is always on the lookout for processors used in high-volume markets that can be repurposed for high-performance specialized applications" said Joel Radford, Mercury vice president, strategic marketing and alliances. "The economics of the gaming industry are driving powerful technologies that we are now leveraging for innovations in medical imaging, video processing, seismic computation, semiconductor design and manufacturing and signal processing."
On the graphics side of visualization, NVIDIA (Santa Clara, Calif.) develops programmable graphics processing units that are able to do much of the graphics processing work that was once handled by the CPU. This is particularly the case with matrix math operations. A powerful driver behind this technology is the consumer 3D gaming market.
"NVIDIA is going to produce roughly 100 million GPUs in 2006, and while only several hundred thousand or a million may be used in image processing they will benefit from the economy of scale that goes with the consumer market," said Jeff Brown, general manager, NVIDIA Professional Services Group.
One significant NVIDIA technology, SLI, today represents the synergy between gaming and high end visualization development.
SLI, which initially stood for "Scan Line Interleave" mode, was first developed by graphics card manufacturer 3dfx in the late 1990s for use in gaming arcade machines and PCs. It basically allowed the linking of two 3D accelerator cards to a standard PCI graphics card, with each accelerator processing alternating horizontal lines of a single frame in a single full screen environment. NVIDIA acquired 3dfx in 2000, and began developing an improved version of SLI technology using 3dfx technology, and engineering talent it acquired from both 3dfx and SGI. The result is today's "Scalable Link Interface" version of SLI, which still allows the linking of multiple GPUs to scale processing power, but is optimized to handle the demands of multiple windows, multiple applications and multiple monitor environments.