Azure Big Compute Cloud HPC

With UberCloud and Azure the benefits of Cloud HPC come with the ease of a workstation

https://www.theubercloud.com/cloud-hpc

Transcript

1. UberCloud Delivering performance! to compute intensive applications!
2. Today’s product development Tools in use for Product Development! ! Desktop: 94%! Server: 5%! Cloud : 1%!
3. Desktop User Interface Flexible Cluster Sizes and Software Licenses Support from CAE and HPC Experts Cloud HPC What’s a better way?
4. Right-Sized Infrastructure 1 End User in Control 3 Enables New Use Cases 4 Faster Time to Solution 2 Cloud HPC Benefits
5. Desktop like graphics with NVIDIA GRID 2.0
6. With Docker Containers Auto-scaling with Azure Batch 80% discount for low-priority jobs Ready to run
7. Azure Marketplace for HPC
8. Ask us for a case study 01 Aerodynamics 02 Automotive 03 Ship Building 04 Crash 05 Fluid Flow 06 Energy 01 02 03 04 05 06 http://blog.theubercloud.com/case-studies-and-resources!
9. Fully interactive GUI for quick post- processing that facilitates iterative development. Access UberCloud- ANSYS on Linux via browser. Offers high performance with desktop ease-of-use. ABAQUS etc. OpenFOAM etc. 1 4 3 2 UberCloud's production-ready containers
10. UberCloud's production-ready containers Welcome email from UberCloud – application install, MPI configuration, HW setup all complete Access UberCloud via browser. Offers performance with desktop ease-of-use. Containers store infrastructure usage statistics in a time-series database that can be viewed on a monitoring dashboard. Fully interactive GUI for quick post-processing that facilitates iterative development. Detailed system performance data within the container. 1 4 3 2
11. Example Project Mini-Cluster ü  32 Gbit/s InfiniBand network with remote direct memory access (RDMA) technology. ü  Ideal for Message Passing Interface (MPI) applications, high-performance clusters, modeling and simulations For small to medium projects ü  8 Node cluster with Intel® Xeon® processor E5-2667 @ 3.6 GHz ü  DDR 4 memory GUI Head Node 128 Core Compute Mini Cluster 382 GB Storage High-speed network for RDMA ü  GUI node for setting up models, visualization, post-processing. ü  Optional GPU capabilities for High-end graphics ü  2000 GB local SSD Storage for faster performance ü  Extra NFS node for shared storage
12. UberCloud CAE Services Cloud Readiness Performance Management Security & Compliance Management Change Management Remote Access ü Interactive GUI ü Workstation-like ü NICE DCV for high end visualization ü Traffic encryption with HTTPS & VPN ü User access ü Upgrades, patches CAE Application ü Library of CAE apps ü App Monitoring ü Industry-specific app benchmarks License compliance to ISV requirements ü Software upgrades ü New applications Data Management ü Data transfer options ü Shared file system ü Tiered storage ü Data transfer planning ü Data security planning ü Optional data encryption HPC Stack ü MPI Setup ü Cluster restarts ü MPI Tuning ü MPI job monitoring ü MPI version Cloud Hardware Cloud infrastructure set up guidelines ü Type & number of compute instances ü Type & size of data storage ü Networking configuration ü Bare-metal & virtualized HW support ü HW performance management & recommendations ü Cloud benchmarks ü Dedicated HW ü Authentication & authorization ü Audit ü Security compliance to customer’s requirements ü Pick your cloud ü Pick your geography ü Frequent HW upgrades ü OS patches & upgrades Thought Leadership - Case Studies - Customer References - ROI & TCO Studies - Performance Benchmarks - Video shorts - POCs
13. Delivering performance! to compute intensive applications! https://www.theubercloud.com/cloud-hpc!