Qualcomm Joins AMD, ARM, Samsung on HSA Foundation

The group is working toward a single architecture spec for chips that run in heterogeneous environments, which leverage both compute and graphics.

Qualcomm is the latest member of the Heterogeneous System Architecture Foundation, joining the likes of Advanced Micro Devices, ARM and Samsung Electronics in creating a single architecture spec for chips that leverage both compute and graphics capabilities.

HSA Foundation founders have said that such an open, single architecture specification and surrounding ecosystem will become the foundation for future computing devices and platforms, making it easier for software developers to program for such heterogeneous environments. Adding Qualcomm, whose ARM-based Snapdragon systems on a chip (SoCs) can be found in many mobile devices, including smartphones and tablets, enhances the stature and work of the organization, according to Phil Rogers, HSA Foundation president and AMD Corporate Fellow.

"It's great to see an innovative company like Qualcomm, which has revolutionized the wireless communications market, placing their support behind HSA," Rogers said in a statement. "With HSA, computing becomes much more power-efficient, enabling member companies like Qualcomm, to create unique and compelling experiences for the consumer."

HSA’s work also will benefit Qualcomm, according to Jim Thompson, senior vice president of engineering at the company.

"Future Snapdragon processors from Qualcomm will contain substantially more computing performance and integrated parallel processing technology in order to meet the high performance, low power needs of our mobile customers," Thompson said in a statement. "We believe that developers will be able to deliver faster and more innovative applications on future Snapdragon processors if certain aspects of heterogeneous computing are standardized."

Other members of the foundation, which was announced in June, include Texas Instruments, Imagination Technologies and MediaTek. Qualcomm joined the organization Oct. 3. Notably absent from the roster are Intel and Nvidia.

AMD and ARM were among the initial founders of the HSA Foundation, whose goal is to create the basis for general-purpose computing, which can leverage the compute power of CPUs and the parallel computing capabilities of graphics technologies. That includes rolling out the necessary tools for developers, from software development kits (SDKs) to libraries to documentation. In addition, the foundation is offering training, support and advice to programmers.

“The HSA vision for enabling heterogeneous computing across a wide range of devices is compelling,” Sasa Marinkovic, senior manager of technology marketing at AMD, said in an Oct. 3 post on AMD’s Fusion blog. “It requires the collaboration of device manufacturers, developers and semiconductor manufacturers working in concert to develop a robust HSA architecture that is open, spurring future innovations for years to come in the PCs, mobile, server, HPC and cloud computing markets.”

Qualcomm, Texas Instruments and Samsung are major licensees of ARM’s low-power SoC designs that are found in most mobile devices, including smartphones and tablets. AMD is trying to gain some traction in the booming mobile device space with its x86-based technology. However, AMD and ARM have an increasingly tight relationship, which could benefit not only AMD’s efforts in the mobile chip space, but also ARM’s ambitions to move its designs in to PCs and low-power servers.

In June, AMD officials announced that they will integrate ARM’s Cortex-A5 processor design with TrustZone security technology into future accelerated processing units (APUs), which integrate the CPU and graphics technology onto a single chip. TrustZone, which is designed to make online transactions and content streaming safer, is in all of ARM’s Cortex-A chips, which are found in mobile devices.

AMD also is a major driver of general-purpose heterogeneous computing with its APUs. Company officials also are driving a strategy they call “ambidextrous computing,” where AMD will use whatever platform—whether it’s their own x86 products or technology from other platforms, including ARM’s—to address customer workloads.

"You want to choose the right type of silicon for the right type of job," Margaret Lewis, director of software planning for AMD's server chip business, told eWEEK.