AMD's Alveo UL3524

The Alveo UL3524 fintech accelerator card from AMD allows electronic traders, market makers, hedge funds, brokerages, and exchanges to engage in rapid electronic trading with nanosecond speeds using a custom 16nm Virtex UltraScale+ FPGA platform. This platform boasts hardened, optimized network connectivity cores, 64 transceivers, 780K LUTs of FPGA fabric, and 1,680 DSP slices of compute, enabling the acceleration of custom trading algorithms in hardware with low latency networking.

The Alveo UL3524 also supports traditional FPGA flows using Vivado Design Suite and comes with a suite of reference designs and performance benchmarks. Furthermore, AMD is providing the open-sourced FINN development framework to simplify AI adoption in the algorithmic trading market. This framework allows developers to reduce the size of AI models while maintaining accuracy, compiling to hardware IP, and integrating the network model into the algorithm's datapath for low latency performance. As an open-source initiative, FINN provides developers with flexibility and accessibility to the latest advancements as projects evolve, backed by AMD's global support from their domain experts.