HPC in the Cloud - CADFEM UK and Ireland - Simulation is ... · ANALYTICS APP SERVICES MOBILE SERVICES DEVELOPMENT & OPERATIONS ... IoT Rules Engine Device Shadows Device SDKs ...
Post on 28-Apr-2018
225 Views
Preview:
Transcript
HPC in the CloudRonan Guilfoyle
Solutions Architect, AWS Ireland
Why the Cloud?
Startup Customers
http://aws.amazon.com/solutions/case-studies/
Enterprise Customers
http://aws.amazon.com/solutions/case-studies/
Vast Infrastructure Technology Platform
ENTERPRISE
APPSDEVELOPMENT & OPERATIONSMOBILE SERVICESAPP SERVICESANALYTICS
Data
Warehousing
Hadoop/S
park
Streaming Data
Collection
Machine
Learning
Elastic
Search
Virtual
Desktops
Sharing &
Collaboration
Corporate
Backup
Queuing &
Notifications
Workflow
Search
Transcoding
One-click App
Deployment
Identity
Sync
Single Integrated
Console
Push
Notifications
DevOps Resource
Management
Application Lifecycle
Management
Containers
Triggers
Resource
Templates
TECHNICAL &
BUSINESS
SUPPORT
Account
Management
Support
Professional
Services
Training &
Certification
Security &
Pricing
Reports
Partner
Ecosystem
Solutions
Architects
MARKETPLACE
Business AppsBusiness
IntelligenceDatabasesDevOps Tools NetworkingSecurity Storage
Access
Control
Identity
Management
Key Management
& Storage
Monitoring
& Logs
Assessment &
Reporting
Resource & Usage
Auditing
SECURITY & COMPLIANCE
Configuration
Compliance
Web Application
Firewall
HYBRID
ARCHITECTURE
Data Backups
Integrated App
Deployments
Direct
Connect
Identity
Federation
Integrated
Resource
Management
Integrated
Networking
API
Gateway
IoT
Rules
Engine
Device Shadows
Device SDKs
Registry
Device
Gateway
Streaming Data
Analysis
Business
Intelligence
Mobile
Analytics
RegionsAvailability
Zones
Points of
Presence
INFRASTRUCTURE
CORE SERVICES
ComputeVMs, Auto-scaling, &
Load Balancing
StorageObject, Blocks, Archivals,
Import/Export
DatabasesRelational, NoSQL,
Caching, Migration
NetworkingVPC, DX, DNS
CDN
How do you use the
AWS Cloud for HPC?
RELEASE RESOURCES
WHEN NO LONGER
NEEDED
ACQUIRE RESOURCES
ON DEMAND
TURN FIXED COST
INTO VARIABLE
PAY FOR
WHAT YOU USE
LEVERAGE OTHER’S
CORE COMPETENCIES
Scale drives faster
innovation at Honda
Scalable Materials Simulations
Before: 80 in-house HPC nodes, 1 year to
complete all required simulations
After: Scalable, on-demand HPC cluster in AWS
Up to 1000 nodes, 16,000 cores at peak
4 months to complete all simulations
“Cloud offers us
an opportunity,
as we can
innovate faster
than before.”
- Ayumi Tada, IT
System Administrator,
Honda R&D
Customers running HPC workloads on AWS
Virtual screening of 10 million compounds against a cancer target
• Approximately 87,000 compute cores at peak
• Estimated 39 years of computational chemistry performed in 9 hours
• Three candidate compounds successfully identified
Cloud for Better Medicine
Steve
Lister,
PhD
Utilities Biopharma Materials Design
Manufacturing Academic Research Auto & Aerospace
Across a variety of industry verticals
HPC Deep Dive AWS July 2016
Partner Success Story : Ansys
• Extensive testing on AWS for running
CFD Simulation (Fluent)
• Airflow around F1 Car, 140M Cell test
scaling well beyond 1K cores
• 50K cells per CPU Core most
beneficial
http://www.ansys-blog.com/simulation-on-the-cloud/
Amazon EC2
C4, M4 & X1
Instance Types
Amazon EC2
P2 Instance
Type
• Up to 16 NVIDIA K80 GPUs
• 64 vCPUs
• 732 GiB of host memory
• 192 GB of GPU memory
• 40 thousand parallel processing cores
Amazon EC2
Spot Instances
On
Reserved Instances
Time
100%
Optimize Utilization on AWS with RI, On-Demand, Spot
Scale up Scale down
Spot
On-Demand
Placement Groups &
Enhanced Networking
Elastic Filesystem &
Object Storage
Container
Schedulers &
AWS Lambda
Experiment, Build, Have Fun
Ronan Guilfoyle
AWS Ireland
rguilfoy@amazon.com
top related