LegUp Computing Inc. provides a platform that enables software developers to program, deploy, scale, and manage FPGA devices for accelerating high-performance applications. Originally developed at the labs of Professors Jason Anderson and Stephen Brown at the University of Toronto, the LegUp platform offers a novel programming model that allows threaded C/C++ software to target FPGA devices connected to high-bandwidth networks for efficient processing of low-latency database, real-time analytics, and machine learning workloads. The company was founded by Dr. Andrew Canis, Dr. Jongsok Choi, Ruolong Lian, and Professor Jason Anderson in 2015 to commercialize the award-winning open-source LegUp high-level synthesis tool.
LegUp's core technology offers a programming model for software developers to achieve 10X better throughput and latency using FPGAs available on leading cloud service providers. We manage the infrastructure required to deploy and scale the FPGA servers and we provide a run-time platform that includes a high-performance network stack and abstracts away the hardware. LegUp allows users to quickly accelerate cloud applications like NoSQL databases to achieve sub-millisecond latencies that can handle millions of requests per second. LegUp offers compression, encryption and machine learning libraries for processing high-velocity data in the cloud. Other applications include real-time decision-making systems, data analytics, video processing, genomics, deep packet inspection and financial trading.
LegUp offers ready-made FPGA acceleration solutions for the datacenter, such as a memcached database running on Amazon's F1 instance achieving 10X better throughput and latency than an AWS ElastiCache memcached instance running on a CPU.