Intel has launched a multi-chip module integrating a Xeon Skylake processor with a field-programmable gate array (FPGA).
Currently, the product is only available to select customers; Japanese tech giant Fujitsu is planning to become one of the first vendors to deliver a system based on the new chipset.
In one package
The Xeon Scalable 6138P includes an Arria 10 GX 1150 FPGA, Intel’s most powerful FPGA, linked to the CPU cores using Intel’s Ultra Path Interconnect (UPI).
The company claims that UPI allows for coherent and direct access to data stored in the processor or FPGA caches, or in main memory.
The Arria 10 GX 1150 includes 1.5 million logic elements and DSP blocks capable of delivering 1.366 single-precision teraflops.
“This solution is also compatible with the Open Virtual Switch (OVS) framework and delivers a dramatic 3.2x throughput improvement at half the latency and 2x more VMs as compared to OVS running on an equivalent processor without FPGA acceleration,” Jennifer Huffstetler, VP of Intel’s data center group, said in a blog post.
“Additionally, code compatibility with Intel’s OVS-DPDK software makes data center retrofits simple and scalable to optimize operational agility.”
Microsoft has emerged as one of the largest users of Intel’s FPGAs to date - earlier this month, the company previewed ‘Project Brainwave,’ an Azure service that relies on Intel’s Stratix 10 FPGAs.
But Intel, which acquired FPGA-maker Altera for $16.7 billion in 2015, faces renewed competition from Xilinx, the original creator of this class of devices. The company’s new CEO, Victor Peng, hopes to make data centers his company’s main focus, telling DCD: “The data center will be our largest area for growth, and we will make it out highest priority.”