Liqid has added FPGAs to the list of compute resources its customers can use to compose workload-specific computing systems.

An FPGA (field-programmable gate array) is a flexible compute chip composed of logic blocks, designed to be programmed or configured “in the field” using a hardware description language. They are more versatile than ASICs, which have a fixed function.

Liqid’s GRID product involves PCIe-based fabric hardware and Command Center orchestration software. It lets customers dynamically set up or compose CPU-GPU-NVMe storage-NIC combinations for specific workloads. Now FPGAs on Xilinx Alveo cards can be added to that mix, extending the range of workloads covered.

Xilinx Alveo FPGA

A suggested application area is artificial intelligence. Alveo cards contain a Deep Neural Network (xDNN) engine, which is good for machine learning inference. xDNN has been developed to generically execute convolutional neural networks like ResNet50, GoogleNet v1, and Inception v4. Grab a white paper here. ®

Source: https://go.theregister.com/feed/www.theregister.com/2018/11/13/fpgas_become_liqid/