Cavium to Add Nvidia GPU Support to ARM-Based Server Chips
SAN JOSE, Calif.—Cavium is adding support for Nvidia's Tesla GPU accelerators to its ARM-based ThunderX server processors.
- Navigating Your IT Career
- Exploring the Private Cloud for Your Organization
- IT Manager's Guide to Social Networking
Cavium launched its 64-bit systems-on-a-chip (SoCs) in December 2014. It is one of a growing number of chip makers looking to leverage ARM's low-power architecture to move into a server market dominated by Intel.
Adding support for Nvidia's Tesla Accelerated Computing platform will help Cavium make ThunderX more attractive to organizations in such areas as high-performance computing (HPC), scientific computing, cloud computing and data analytics, according to the company. The GPU accelerators will improve the performance and efficiency of the ThunderX SoCs.
Cavium's March 16 announcement came a day before Nvidia kicked off its GPU Technology Conference 2015 (GTC) here.
"Our collaboration with Nvidia is yet another demonstration of the workload-optimized focus that Cavium is driving in the server market with ThunderX," Gopal Hegde, vice president and general manager of Cavium's Data Center Processor Group, said in a statement. "Nvidia's leadership in high-performance computing solutions for the HPC and data analytics markets is well recognized and complements Cavium's continued innovation in processors for next generation data center and cloud applications."
ARM and its chip-making partners are looking to leverage the low-power architecture found in most smartphones and tablets for server environments that require high performance and power efficiency. Several chip makers, including Cavium and Applied Micro, have launched SoCs based on the ARMv8 architecture, with others like Advanced Micro Devices and Qualcomm moving forward with plans.
Cavium's ThunderX chip family offers up to 48 cores, single- and dual-socket configurations, and integrated accelerators, high- memory bandwidth, large memory capacity and integrated bandwidth network and storage I/O. ThunderX also includes a low-latency Ethernet fabric.
HPC organizations over the past several years have begun to leverage GPU accelerators from Nvidia and AMD—as well as x86 co-processors from Intel—to boost the performance of their systems while keeping a lid on power consumption and operating costs. Nvidia in November 2014 rolled out the Tesla K80, the latest generation of its GPU accelerators.
Cavium's ThunderX chips already are finding their way into the HPC space. Cray announced in November that it will evaluate using ARM-based chips in its systems and is working with Cavium to create Cray compute clusters powered by Cavium's 48-core ThunderX SoCs.
"We see alternatives such as 64-bit ARM, custom ASICs and low-power Intel processors as enabling technologies for certain HPC [high-performance computing] and analytics workloads, and a natural fit for our strategy," Steve Scott, senior vice president and CTO for Cray, said in a statement at the time.
Cavium officials said they will have customer reference platforms with Nvidia's GPU accelerators on display at the GTC show. Support in ThunderX for the Tesla accelerators is scheduled to be available in the second quarter.
Read more on "Server Infrastructure Tools" »