An Integrated FPGA orchestrator for seamless scalability and resource management

About

Recently major cloud and HPC providers like aws, Alibaba, Huawei and Nimbix have started deploying FPGAs in their data centers. However, currently the development tools and frameworks offered by FPGA vendors and cloud providers do not allow the utilization of FPGA clusters from multiple applications. For example, if one application wants to distribute the workload into several FPGAs in a server, users have to do a manual scheduling to perform the distributions. Similarly, when multiple applications want to share one or more FPGAs there is not any framework available to perform the sharing of the resources and the scheduling of the request to the FPGA cluster.

In this talk we will present Coral FPGA resource manager for Intel that allows the software community to instantiate and utilize a cluster of Intel FPGAs with the same easy as invoking typical software functions. Coral FPGA resource manager allows multiple applications to share and utilize a cluster of FPGAs in the same node (server) without worrying about the scheduling, load balancing and the resource management of each FPGA. The performance evaluation shows that the specific framework does not add any significant overhead in the servers and allows the instant utilization and sharing of the FPGA resources. We will demonstrate how the FPGA resource manager can be applied to scale Machine Learning and HPC applications on cloud or on-prem.

0:00:00 Slide 1

0:00:56 Slide 2

0:01:27 Slide 3

0:02:05 Slide 4

0:02:31 Slide 5

0:03:14 Slide 6

0:04:10 Slide 7

0:05:22 Slide 8

0:06:04 Slide 9

0:06:37 Slide 10