The reconfigurable acceleration stack speeds things up at 6x the efficiency of other FPGAs, claims Xilinx.
Xilinx has introduced its FPGA-supported Reconfigurable Acceleration Stack designed to aid cloud platform deployment by service providers.
The Reconfigurable Acceleration Stack includes libraries, framework integrations, developer boards, and OpenStack support. It provides 40x compute efficiency with Xilinx FPGAs compared to x86 server CPUs, and up to six times compute efficiency over other FPGAs.
With the new offering, Xilinx enables silicon optimisation for heavy workloads such as machine learning, data analytics, and video transcoding. Workload optimisations can be done by swapping in optimal design bitstream.
Xilinx's FPGAs enable hyperscale data centres to operate at two to six times the compute efficiency in machine learning inference, derived from DSP architectural advantages for limited precision data types and on-chip memory resources.
The new FPGA Acceleration Stack includes math libraries designed for cloud computing workloads, application libraries integrated with major frameworks, such as Caffe for machine learning, a PCIe-based development board and reference design for high density servers, and an OpenStack support package.
The Xilinx Reconfigurable Acceleration Stack is already available to major cloud service providers.