Accelerate Your Workloads at Massive Scale
HPC Clustering for Every Data Center
Scale Out Architecture on RoCE v2 and DCBx with existing data center server infrastructure
Shared workload and shared memory across multiple cards
MPI enables hyper-parallelism of Xilinx Adaptive Compute across nodes
Software developers and data scientists can unlock the benefits of Alveo and adaptive computing through high-level programmability of both the application and cluster utilizing the Vitis™ platform.
Xilinx’s Vitis™ Unified Software Platform includes an extensive set of open-source, performance-optimized libraries that offer out-of-the-box acceleration with minimal to zero-code changes to your existing applications.
Vitis HPC Clustering Libraries
Alveo U55C Card Acceleration in Action
Graph Database Analytics Acceleration
The Alveo U55C card is able to accelerate query times and predictions for recommendation engines from minutes down to milliseconds. The U55C’s superior computational power and memory bandwidth enable graph query speeds up to 96x faster compared to CPU-based clusters, while increasing the quality of scores by up to 35%.
This improvement results in greater confidence, dramatically lowering false positives to low single digits.
TigerGraph, the world’s most scalable graph database, turned to Xilinx to accelerate their software-based graph analytics solution.
Utilizing multiple Alveo U55C cards, TigerGraph tapped into Xilinx’s massive clustering ability to accelerate the two most prolific algorithms that drive graph-based recommendation engines: Cosine Similarity and Louvain Modularity.
In this test, the Alveo U55C card performed 96x faster than a CPU running the TigerGraph Cosine Similarity recommendation engine across millions of patient records to deliver results in milliseconds.
With TigerGraph’s Graph analytics database running on the Alveo U55C card, the Louvain clustering algorithm detected fraudulent attempts across hundreds of bank accounts 45x faster.
Finite Element Method Acceleration
Dimensions of matrix - 12 million
Time in sec: JPCG solver equation runtime
CPU: Intel Xeon E5-2665
LS-DYNA from ANSYS is a general-purpose finite element program capable of simulating complex real-world problems. LS-DYNA allows designers and engineers the ability to create simulations with an infinite amount of complexity.
In this this test, LS-DYNA tapped into the Alveo U55C cards’ scale-out fabric and non-hierarchical memory structure to obtain answers 5x faster than alternate solutions.
Real-Time Signal Processing Acceleration
The Commonwealth Scientific and Industrial Research Organization’s (CSIRO) Square Kilometer Array (SKA) - the world’s largest radio astronomy antenna array – depends on Alveo U55C cards to perform end-to-end signal packet processing on incoming data at 15Tbps for real-time data flow.
The Alveo card-based cluster allows CSIRO to tackle the massive compute task of aggregating, filtering, preparing and processing data from 131,000 antennae in real time. The 460GBs of HBM2 bandwidth across the signal processing cluster is served by 420 Alveo U55C cards fully networked together across P4-enabled 100Gbs switches.
The explosion of sensors at the Edge has accelerated the need to process data in real-time.
Commodity GPUs running streaming sensor data clusters simply can’t keep up with the price, performance and power that Alveo U55C cards deliver.
Easy Evaluation on the Cloud or On-Premises