Advanced Micro Devices is taking the wrapping off its first Fusion accelerate processing units for the embedded market.
AMD launched the G-Series Jan. 19, just weeks after unveiling its first Fusion processors for consumer PCs at the 2011 Consumer Electronics Show. Like those processors, the G-Series embedded APUs include the x86 CPU, graphics unit and memory controller on the same piece of silicon.
It’s that design that gives AMD’s offerings an advantage over those from rival Intel, according to John Fruehe, director of product marketing for server, embedded and FireStream products at AMD.
“This allows [systems makers] to create much more compact, much smaller, more energy-efficient designs,” Fruehe said in an interview with eWEEK.
AMD, which has been playing in the embedded space since 2003 with its Geode processor line, is fighting in a highly competitive market that includes not only Intel and its Atom platform, but other chip vendors such as Via, ARM Holdings and MIPS. Fruehe said that x86 processors from AMD and Intel are found in about one-third of what some AMD officials say is a $10 billion market.
With the G-Series, AMD not only expects to offer current customers upgrades, but also to expand its embedded business. AMD has seen a 350 percent increase in its embedded business over the past three years, and demand for embedded technologies is growing, fueling an expected 10 percent annual growth rate.
That demand is coming from a growing number of applications in a wide range of verticals, from thin clients, digital signage and POS (point-of-sale) and kiosk systems to medical imaging, telecommunications and networking. Fruehe pointed to gaming machines and low-power PCs as potential uses, and said technologies that demand high-performance graphics, such as set-top boxes and TVs, are a key target for AMD.
The G-Series comes integrated with AMD’s “Bobcat” CPU core and its DirectX 11-capable GPU on a single die. AMD is offering single-core chips that require 9 watts of power, and dual-core products that draw 18 watts. Each will have speeds of up to 1.6GHz.
Having such high-performance graphics capabilities integrated with the compute functionality gives AMD an edge when it comes to applications that require high-end graphics capabilities, Fruehe said. Pointing to Intel’s “Sandy Bridge” second-generation Core-i chips, which also were introduced at CES, he said the processors might have integrated graphics, but not DirectX 11-capable graphics.
“It’s going to be a very big differentiator for AMD because … while [Intel and other rivals] have products with low power, they don’t have products with low power and high-end graphics,” he said.
Intel this quarter will launch its E600C series Atom chips-codenamed Tunnel Creek-which integrate Altera FPGAs (field-programmable gate arrays), which let device makers program the chips for specific customers needs.
Fruehe said he expects rapid adoption of the G-Series among systems makers, including Wyse Technology, Advansus, Mitec and Starnet. Already some vendors are making their plans known. In a blog on AMD’s Website, Kontron said it will use the G-Series in three small form factor designs, and that the strong graphics capabilities in the embedded APUs will be attractive for such applications as multimedia content delivery, kiosks, POS and gaming.
“By implementing the new AMD G-Series APUs on the most common form factors for graphics-intensive applications, Kontron is making the benefits of this innovative architecture readily available for application development,” Norbert Hauser, vice president of marketing at Kontron, said in the blog.
In another blog, Fujitsu Technology Solutions said it will use the G-Series in two new Mini-ITX mainboards.