Intel has begun shipping samples of its new Stratix 10 programmable chips, targeting customers looking to boost performance in data centres and cloud and web-scale environments dealing in fast processing and heavy data workloads.
In an official blog post, Dan McNamara, Intel’s corporate VP and general manager of the Programmable Solutions Group, noted how his unit designed the Stratix 10 field-programmable gate array (FPGA) chips by pairing Intel’s 14-nanometre tri-gate process technology with new HyperFlex fabric architecture.
McNamara explained that this combination allows the FPGA accelerators to free up processing loads from the CPU and reduce power consumption, meaning they are able to meet increasingly demanding performance requirements in data centres, network infrastructure, cloud services, and radar and imaging systems.
‘We live in a smart and connected world where billions of devices are creating massive amounts of data that must be collected, rapidly processed and analyzed, and available from anywhere,’ wrote McNamara. ‘With Stratix 10 FPGAs, Intel is enabling service providers… to satisfy their insatiable demand for higher computational capabilities, lower latency, greater system flexibility and increased power efficiencies,’ he added.
The FPGAs are powered by an embedded quad-core 64-bit ARM Cortex-A53 processor, offer up to 10 teraflops of single-precision floating point DSP performance, and up to 1 TBps of memory bandwidth with integrated High-Bandwidth Memory (HBM2).
With the rise in cloud computing and emerging areas such as the Internet of Things (IoT), accelerators are becoming a popular tool for use in data centre servers. Intel is betting on this trend so much so that it expects to see FPGA technology deployed in 30% of data centres worldwide by 2020.
Intel’s push into flexible accelerators follows its acquisition of FPGA specialist Altera for $16.7 million (approx. £13 million) last year. Intel announced at the time that the technology would fit in with its analytics and IoT portfolio, as well as in autonomous vehicles development.
The company is also looking at integrating its FPGA technology with its 3D camera RealSense, to help develop object and facial recognition solutions.