Can we boost more HPC performance? Integrate IBM Power servers with GPUs to OpenStack Environment
224 views · Published 24 May 2018 · 28:13 · Indexed 4 November 2026
Channel: OpenInfra Foundation · 2018 · Science & Technology
Deep learning(DL), High-Performance Computing(HPC), Big Data Analysis etc. has achieved remarkable progress in the recent years. To manage the heavy workloads, we need high-speed communication between CPU-GPU and GPU-GPU. At the same time, rapid enhancement of GPGPU programming is required but its so complicated. We know utilizing OpenACC and CUDA Unified Memory will be helpful to reduce programming complexity but it requires more cost between CPU and GPU.So we choose IBM Power server with the f
More from this channel
-
30:59
San Diego 2012 Keynotes: Mark Shuttleworth
-
39:42
Building an OpenStack Security Group
-
32:06
War Stories Moving from Eucalytups to OpenStack
-
41:41
Making clouds faster for fun and profit
-
36:25
Service Resiliency Doesnt always mean HA or cluster
-
41:18
Cloudifying Virtual Desktops
-
28:05
Red Hats Cloud Strategy
-
41:36
OpenStack its impact on the Enterprise and startup community