eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Advanced Micro Devices is rolling out the second generation of its Embedded R-series chips, taking another step in its efforts to expand its reach in the increasingly important embedded systems market.
AMD officials unveiled the new chip portfolio—code-named “Bald Eagle”—May 20. The high-end embedded family includes CPUs and accelerated processing units (APUs) that feature both the CPU and graphics on the same piece of silicon. The chips offer a number of firsts, including the first to incorporate Heterogeneous System Architecture (HSA) features and the first to support 4K high-definition applications.
The Bald Eagle chips also support enhanced power management features and AMD’s Graphics core Next (GCN) architecture, which is designed to give embedded processors greater visual and parallel-processing capabilities.
The new embedded chips are a key part of AMD’s larger strategy to reduce its reliance on the stagnant PC market by extending its reach into several growth areas, including dense servers, ultraportable devices, semi-custom chips, professional graphics as well as the embedded market. Two years ago, about 95 percent of AMD’s revenues came from its PC chip business. AMD executives have said that by the end of 2015, these new growth markets will account for more than 50 percent of the company’s sales.
AMD’s capabilities—including its strong graphics technology, APU architecture and support for the HSA architecture—give the vendor advantages in an embedded space that is about $10 billion now and growing and is in transition, according to Scott Aylor, corporate vice president and general manager of embedded solutions at AMD.
“The embedded market is diverse and it’s fragmented,” Aylor told eWEEK.
Systems like HD displays are increasing the demand for strong graphics technology, and embedded applications are increasingly parallel, which also plays to AMD’s strength in graphics. The HSA architecture is designed to make it easier to move workloads between the CPU and GPU. Before, the workloads were sent to the CPU, which then decided whether to move them to the GPU. With HSA, the workloads immediately go to whichever element will work best.
The burgeoning Internet of things also will increase the demand for embedded components.
AMD is aiming Bald Eagle chips at a number of industries, including medical imaging systems, digital signage, gaming machines, industrial control and automation, communications, and networking infrastructure. With the rise of such trends as software-defined networking (SDN) and network-functions virtualization (NFV), networking systems are increasingly able to take advantage of the parallel processing capabilities of the APU, Aylor said.
In addition, by offloading such tasks as deep-packet inspection and encryption and decryption onto the GPU, organizations can take advantage of the increased headroom in the CPU, he said.
The new chips, based on the vendor’s Steamroller CPU architecture, offer CPU frequencies that range from 2.2GHz to 3.6GHz. With the GCN graphics architecture, the GPU frequencies range from 533HMz to 686MHz. Aylor said the new products offer compute and graphics performance advantages over both AMD’s previous generation released last year and competing chips from Intel. In GPU performance, the Bald Eagle chips offer 44 percent improvements over comparable Intel products, and a 55 percent boost over their predecessors. In compute performance, the improvements are 44 percent over Intel and 66 percent over AMD’s earlier chips.
The chips also offer access to customized embedded Linux development and commercial support through AMD’s work with Mentor Graphics and the Yocto Project, an open-source collaboration effort though the Linux Foundation. The new Embedded R-series chips also include features aimed at the embedded space, including 10-year longevity, dual-channel memory with error-correcting code (ECC), DDR3 support and the ability to optimize the processor at a lower thermal design point (TDP) for improved energy efficiency.
The chips will begin appearing in embedded systems later this year.