This document summarizes an agenda for an AWS event on using cloud computing for semiconductor and electronics applications. The agenda includes presentations on using AWS for design and simulation, global collaboration, and big data analytics in manufacturing. Key themes discussed are using the cloud for scalability to enable faster simulations with higher accuracy, global collaboration for improved IP security and efficiency, and agility to quickly adapt cloud resources to changing needs. The cloud is presented as a way to improve innovation through more frequent experimentation at lower cost compared to traditional on-premises infrastructure. Case studies are provided of customers using AWS for applications like electronic design automation and manufacturing yield analysis.
How AI, OpenAI, and ChatGPT impact business and software.
AWS for Semiconductor and Electronics Design | Hsinchu, April 10
1. AWS for Semiconductor
and Electronics
Using Cloud for Design, Engineering, Manufacturing
April 10, 2014
David Pellerin, Principal Business Development Manager, HPC
Amazon Web Services
2. Agenda
13:30
AWS Cloud forIT Enterprise – OpeningRemarks and Case Studies
James Tien,Sales & Business Development Manager,Amazon Web Services
14:00
15:00
AWS Cloud for Design and Simulation – Case Studies in CAE and EDA
David Pellerin,HPC Business Development Principal,Amazon Web Services
15:15
16:00
Using Cloud forGlobal Collaboration – Demonstrations in CAE and EDA
David Pellerin,HPC Business Development Principal,Amazon Web Services
16:00
16:45
MentorGraphics Design Collaboration
Julian Sun,Business Development Director,Mentor Graphics
David Pellerin,HPC Business Development Principal,Amazon Web Services
16:45
AWS Kinesis – Big Data Management and Analytics in Manufacturing
Ken Chan,Solutions Architect,Amazon Web Services
David Pellerin,HPC Business Development Principal,Amazon Web Services
17:15
ClosingRemarks and Q&A
James Tien,Sales & Business Development Manager,Amazon Web Services
8. 10 AWS Regions Worldwide
25 Availability Zones
Tokyo
Region
Sydney
Region
Singapore
Region
China
Region
9. Global Content Delivery Network
51 Edge Locations
Europe
Amsterdam (2)
Dublin
Frankfurt (3)
London (3)
Madrid
Marseille
Milan
Paris (2)
Stockholm
Warsaw
Asia
Chennai
Hong Kong (2)
Manila
Mumbai
Osaka
Seoul
Singapore (2)
Sydney
Taipei
Tokyo (2)
South America
Sao Paulo
Rio de Janeiro
North America
Ashburn, VA (3)
Atlanta, GA
Dallas, TX (2)
Hayward, CA
Jacksonville, FL
Los Angeles, CA (2)
Miami, FL
Newark, NJ
New York, NY (3)
Palo Alto, CA
Seattle, WA
San Jose, CA
South Bend, IN
St. Louis, MO
10. Compute Networking
Storage &
CDN
Database App Services Management
Amazon EC2
Amazon ELB
AutoScaling
Amazon WorkSpaces
Amazon Route 53
Amazon VPC
AWS Direct Connect
Amazon S3
Amazon Glacier
Amazon EBS
AWS Storage Gateway
AWS Import/Export
Amazon CloudFront
Amazon RDS
Amazon DynamoDB
Amazon Elasticache
Amazon RedShift
Amazon AppStream
Amazon CloudSearch
Amazon SWF
Amazon SQS
Amazon SNS
Amazon SES
Amazon Elastic Transcoder
Mobile Push
AWS IAM
Amazon CloudWatch
AWS Elastic Beanstalk
AWS CloudFormation
AWS OpsWorks
AWS CloudHSM
AWS CloudTrail
AWS Trusted Advisor
AWS Marketplace
AWS Premium
Support
AWS Professional
Services
AWS
Training
Over 40 Broad & Deep Services to Support Virtually Any
Cloud Workload
Analytics
AWS Data Pipeline
Amazon Kinesis
Amazon EMR
13. Media Sharing Explosive traffic
accommodation
Consumer
social app
Ticket pricing
optimization
SAP &
Sharepoint
Securities Trading
Data Archiving
Marketing
campaign
Marketing web
site
Interactive
TV apps
Fast development
and deployment
R&D data
analysis
Machine Learning
system development
Big data
analytics
Customized
movie
suggestion
Disaster
recovery
Media streaming
Web and mobile
apps
Streaming
webcasts
Facebook
app
Consumer social
app
Every Imaginable Use Case
Global
game
service
16. 2. Lower Total Cost of IT
Scale allows us to constantly
reduce our costs
We are comfortable running a high
volume, low margin business
We pass the savings along to
our customers in the form of
low prices
42 Price
Reductions
18. 4. Dramatically Increase Speed and Agility
Old World
Infrastructure in Weeks
Infrastructure in Minutes
Add New Dev Environment
Add New Production Environment
Add New Environment in Japan
Add 1,000 Servers
Remove 1,000 servers
Number of Instances 1,000
Instance Type M3 Extra Large
Availability Zone US-West-2b
Launch
aws.amazon.com/managementconsole
19
19. Experiment Often
Fail quickly at a low
cost
More Innovation
4. Increase Agility when Innovation is Fast and Low Risk
On-Premises
Experiment Infrequently
Failure is expensive
Less Innovation
20
Nearly $0
$ Millions
22. AWS Cloud for
Electronics and
Semiconductor
Introduction and Case Studies
April 10, 2014
David Pellerin, Principal Business Development Manager, HPC
Amazon Web Services
23. Cloud for Scalable EDA
• Technical capabilities
• Business realities
Cloud for Secure Global Collaboration
• New, more innovative solutions for EDA users
• New opportunities for EDA software vendors
Cloud for Big Data Analytics
• For manufacturing yield analytics
• For improved Design-for-Manufacturing
Themes: for Today and the Future
24. Scalability:
Go wide, go large for faster time-
to-results at higher accuracy
Global Collaboration:
For enhanced IP security, more
efficient operations
Agility:
React quickly to changing needs
with flexible cloud capacity
Motivators for the Cloud
26. Agenda
13:30
AWS Cloud forIT Enterprise – Overview and Case Studies
Tom O'Reilly, Head of Hong Kong & Taiwan,Amazon Web Services
14:00
15:00
AWS Cloud for Design and Simulation – Case Studies in CAE and EDA
David Pellerin,HPC Business Development Principal,Amazon Web Services
15:15
16:00
Using Cloud forGlobal Collaboration – Demonstrations in CAE and EDA
David Pellerin,HPC Business Development Principal,Amazon Web Services
16:00
16:45
MentorGraphics Design Collaboration
Julian Sun,Business Development Director,Mentor Graphics
David Pellerin,HPC Business Development Principal,Amazon Web Services
16:45
AWS Kinesis – Big Data Management and Analytics in Manufacturing
Ken Chan,Solutions Architect,Amazon Web Services
David Pellerin,HPC Business Development Principal,Amazon Web Services
17:15
ClosingRemarks and Q&A
James Tien,Sales & Business Development Manager,Amazon Web Services
27. AWS Cloud for Design
and Simulation
Why Scalability Matters for CAE and EDA
April 10, 2014
David Pellerin, Principal Business Development Manager, HPC
Amazon Web Services
28. Computer-Aided Design, Simulation, Analysis, Visualization
• Across industries, the trend is Simulation-Driven Design and Discovery
• Aerospace, semiconductor, automotive, civil engineering, energy exploration,
consumer products, finance, pharmaceuticals, many others
Examples in Design and Manufacturing
• Computer-Aided Design (CAD) including 3D models
• Finite Element Analysis (FEA) and Thermal Analysis
• Electronic Design Automation (EDA)
• Computational Fluid Dynamics
• Multi-physics simulations
• Molecular simulations for drug discovery
A Simulation-Driven World
29. A Collaborative World
Collaboration between functional groups
• Product Lifecycle Management
• Collaborative Design
• Concurrent Design
Collaboration for global teams
• Secure remote access to IP and applications
30. A Data-Intensive World
Managing big data for competitive advantage
• For design, engineering, production environments
• Quality and Yield Analysis
• Statistical Process Control
Processing
Input
Yield analysis
Manufacturing facilitymonitoring
In-field devicemonitoring
Logging
Log4J
Appender
push
to
Kinesis
ElasticMapReduce
Hive
Pig
Cascading
MapReduce
pull from
31. What are AWS Customers Telling Us?
“HGST is using AWS for a
higher performance, lower
cost, faster deployed
solution vs buying a huge
on-site cluster.”
- Steve Philpott, CIO
HGST application roadmap:
Molecular dynamics
CAD, CFD, EDA
Collaboration tools for engineering
Big data for manufacturing yield analysis
Every application
presents unique
challenges… some
technical, some
business
32. Cloud Provides Agility
Wasted Resources
Project Delays
Actual demand
Predicted Demand
Rigid On-Premise Resources Elastic Cloud-Based Resources
Actual demand
Resources scaled to demand
3 to 5 year architecture commitment Little or no architecture commitment
33. Maintaining an EDA cluster is expensive
Is it worth your organization’s time and effort?
39. The Hidden Cost of Queues
Conflicting goals
• EDA users seek fastest possible time-to-results
• Simulations are not steady-state workloads
• IT support team seeks highest possible utilization
Result:
• The job queue becomes the capacity buffer
• Job completion times are hard to predict
• Users are frustrated and run fewer simulations
Fewer simulations = lost opportunity!
?
40. The Hidden Cost of Queues
This is what
100% utilization
looks like
41. On the cloud, clusters are created on-demand
and can be balanced dynamically for each job…
46. Use automation to manage cluster
sizing and monitor jobs and costs
AWS Auto Scaling works
with existing HPC
scheduling software
47. Who Uses Cloud Today?
global enterprises, global applications
48.
49. Worldwide Research and Development
“The Amazon Virtual Private Cloud was a unique option that offered an additional
level of security and an ability to integrate with other aspects of our infrastructure.”
“AWS enables Pfizer’s WRD to explore specific difficult or deep
scientific questions in a timely, scalable manner and helps
Pfizer make better decisions more quickly”
Dr. Michael Miller, Head of HPC for R&D, Pfizer
http://aws.amazon.com/solutions/case-studies/pfizer/
52. Courtesy of Cypress Semiconductor
Supporting Innovation in Electronic Product Design
53. EM Field Simulations
for TRUETOUCH® Touchscreen Controllers– Cypress Semiconductor
3D FEM simulations SPICE
OUTPUT: sensor speed, SNR, and
signal disparity
OUTPUT: unit cell parameters in
respect to sensor design
Finite-element mesh used for 3D simulations consists of over one million vertices
Lack of computational resources can limit the capability to model complex
geometries and/or increase simulation time Courtesy of Cypress Semiconductor
54. MASTER Node 01
Virtual Private Cloud
Job 01: parameter set 2
Job 02: parameter set 2
Job NN: parameter set NN
Job submission
Accumulated
simulation results
Simulations can be submitted as an array of jobs
that share the same executable and libraries,
different input parameters
Result: simulation time reduced
from weeks to just hours
Node 02
Node N
EM Field Simulations
for TRUETOUCH® Touchscreen Controllers– Cypress Semiconductor
Courtesy of Cypress Semiconductor
57. What Does Scale Mean in the Cloud?
18 hours
205,000 materials analyzed
156,314 AWS Spot cores at peak
2.3M core-hours
Total spending: $33K
(Under 1.5 cents per core-hour)
58. How do you Scale an EDA Cluster?
Actual demand
Predicted Demand
What size of cluster do you need?
• A different size of cluster is needed at different
points in the engineering process
• Pace of innovationwill depend on making the right
sizing decision
And what kind of cluster is it?
• Large memory?
• More and faster cores?
• Faster storage?
• Faster networks?
• What generation of processor?
• IT hardware is a long-term commitment – when is
the right time to buy?
61. Performance Factors: CPU
• Intel Xeon E5-26XX v2 (Ivy Bridge) CPUs
• Available in AWS C3, R3, I2 instance types
• 2.8 GHz, Turbo enabled up to 3.6 GHz
• Intel® Advanced Vector Extensions (Intel® AVX):
• 256 bit instruction set extension
• Designed for applications that are floating-point (FP) intensive
• The “Ivy Bridge” microarchitecture enhances this with the
addition of float 16 format conversion instructions
62. C3: CPU-Optimized Instance Type
• 2.8 GHz Intel Xeon E5-2680v2 (Ivy Bridge) CPU
• Turbo enabled to 3.6 GHz
• Various instance sizes with 2, 4, 8, 16, 32 vCPUs
• From 3.75GiB to 60GiB RAM
• From 32GB to 640GB SSD
• High PPS, low-latency Enhanced Networking: over 1M PPS
• Supporting Cluster Placement Groups for all sizes
63. R3: Memory-Optimized Instance Type
• 2.5 GHz Intel Xeon E5-2680v2 (Ivy Bridge) CPU
• Multiple instances sizes with 2, 4, 8, 16, 32 vCPUs
• Up to 244 GiB RAM (~ 8GiB/vCPU)
• SSD Based Instance Storage
• High PPS, low-latency Enhanced Networking
64. I2: High-IOPS Instance Type
• 2.5 GHz Intel Xeon E5-2680v2 (Ivy Bridge) CPU
• Various instances sizes with 4, 8, 16, 32 vCPUs
• 30.5, 61, 122, 244 GiB RAM
• 16 vCPU: 3.2 TB SSD; 32 vCPU: 6.4 TB SSD
• 365K random read IOPS for 32 vCPU instance
• High PPS, low-latency Enhanced Networking
65. Performance Factors: Networks
• AWS proprietary, 10Gb networking
• Highest performance in .8xlargeinstance sizes
• Full bi-section bandwidth in placement groups
• No network oversubscription
• Enhanced Networking
• Availableon C3, R3, I2 (in VPC with HVM)
• Over 1M PPS performance, reduced instance-to-instance
latencies, more consistentperformance than earlier
generation AWS networks
66. Performance Factors: Accelerators
NVIDIA GPUs!
• For computing and for remote graphics
• CG1 and G2 instances
• GPU accelerators augment CPU-based
computing by offloading specialized
processing
• Performance gains depend on application-
level support
67. Today?
Testing and development,patch testing, user training
EDA vendor sales enablement via “test drives”
Customer POCs, using real production examples
Customer-managed production EDA
• With or without EDA vendor involvement
Tomorrow…
• Vendor-approved and documented cloud architectures for EDA
• Customer-approved security and compliance best-practices
• New EDA license models supporting extreme scalability
• New software architectures allowing faster time-to-results,
higher quality at reduced infrastructure cost
Cloud for EDA: Today and Tomorrow
68. 1) Customer Managed Application Hosting
• Customer has account with cloud provider and manages virtual infrastructure
• Cloud used for batch jobs via cluster management software
• Customer can also remote login and globally collaborate using GPU instances
• Customer maintains traditional software vendor relationships
• Software vendor optionally offers license flexibility for scalable computing
2) Software Vendor Managed Application Hosting
• SaaS or hybrid model for acceleration of batch tasks, for example rendering
• Customer pays software vendor for cloud-hosted services
• Customer does not need to manage virtual infrastructure
Options for Software Licensing
72. Cost Benefits of HPC in the Cloud
On-Premise
HPC
Metered, Pay As You Go Model
Use only what you need,
using on-demand, reserved, or spot
Flexible
Capital Expense Model
High upfront capital cost,
high cost of ongoing support
Inflexible
Cloud-Based
HPC
73. Optimize Costs
by Combining Reserved, Spot, and On-Demand Instances
0
2
4
6
8
10
12
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Heavy Utilization Reserved Instances
Light RI Light RILight RILight RI
On-DemandSpot and
On-Demand
100%
80%
60%
40%
20%
Percentage of Peak Requirements Over Time
75. Agenda
13:30
AWS Cloud forIT Enterprise – Overview and Case Studies
Tom O'Reilly, Head of Hong Kong & Taiwan,Amazon Web Services
14:00
15:00
AWS Cloud for Design and Simulation – Case Studies in CAE and EDA
David Pellerin,HPC Business Development Principal,Amazon Web Services
15:15
16:00
Using Cloud forGlobal Collaboration – Demonstrations in CAE and EDA
David Pellerin,HPC Business Development Principal,Amazon Web Services
16:00
16:45
MentorGraphics Design Collaboration
Julian Sun,Business Development Director,Mentor Graphics
David Pellerin,HPC Business Development Principal,Amazon Web Services
16:45
AWS Kinesis – Big Data Management and Analytics in Manufacturing
Ken Chan,Solutions Architect,Amazon Web Services
David Pellerin,HPC Business Development Principal,Amazon Web Services
17:15
ClosingRemarks and Q&A
James Tien,Sales & Business Development Manager,Amazon Web Services
76. Using Cloud for Global
Collaboration
Use-Cases in CAE and EDA
April 10, 2014
David Pellerin, Principal Business Development Manager, HPC
Amazon Web Services
78. Global Collaboration for Global Manufacturing
Cloud provides a
global, distributed,
secure, and scalable
environment for
collaborative design
and manufacturing
79. Collaboration is More Secure in the Cloud
Bring the users to the data, don’t send the data
to the users
80. Collaboration is More Secure in the Cloud
Bring the users to the data, don’t send the data
to the users
81. Secure Remote Access
Data and computation hosted in a secure, customer-managed virtual private cloud, with
controlled access via a wide variety of client devices.
Virtual Private Cloud
Powered by NVIDIA GPUs
82. NVIDIA GRID K520 in AWS Cloud
Product Name GRID K520
GPUs 2 x GK104 GPUs
CUDA cores 3,072 (1,536 per GPU)
Core Clocks 800 MHz
Memory Size 8GB GDDR5 (4GB per GPU)
HW Video Encoder 2x h.264 (1 per GPU)
Power Consumption 225W
Supported APIs
OpenGL 4.3, DirectX 9, 10, 11, CUDA
5.5, OpenCL 1.1, NVFBC, NVIFR,
NVENC
88. AWS Test Drive
• Provides softwarevendors with a
controlled, secure, convenient
environment for product
evaluation and training
• Any application listed on AWS
Test Drive is available for
purchasefrom the ISV, and can
be deployed on AWS if desired
89. Cloud-based PLM with fast deployment
and simplified scalability
Dynamically scale PLM infrastructure
up and down based on project needs
90. Agenda
13:30
AWS Cloud forIT Enterprise – Overview and Case Studies
Tom O'Reilly, Head of Hong Kong & Taiwan,Amazon Web Services
14:00
15:00
AWS Cloud for Design and Simulation – Case Studies in CAE and EDA
David Pellerin,HPC Business Development Principal,Amazon Web Services
15:15
16:00
Using Cloud forGlobal Collaboration – Demonstrations in CAE and EDA
David Pellerin,HPC Business Development Principal,Amazon Web Services
16:00
16:45
MentorGraphics Design Collaboration
Julian Sun,Business Development Director,Mentor Graphics
David Pellerin,HPC Business Development Principal,Amazon Web Services
16:45
AWS Kinesis – Big Data Management and Analytics in Manufacturing
Ken Chan,Solutions Architect,Amazon Web Services
David Pellerin,HPC Business Development Principal,Amazon Web Services
17:15
ClosingRemarks and Q&A
James Tien,Sales & Business Development Manager,Amazon Web Services
91. Design Collaboration
Featuring Mentor Graphics
April 10, 2014
Julian Sun, Business Development Director, Mentor Graphics
David Pellerin, Principal Business Development Manager, HPC
Amazon Web Services
93. Agenda
13:30
AWS Cloud forIT Enterprise – Overview and Case Studies
Tom O'Reilly, Head of Hong Kong & Taiwan,Amazon Web Services
14:00
15:00
AWS Cloud for Design and Simulation – Case Studies in CAE and EDA
David Pellerin,HPC Business Development Principal,Amazon Web Services
15:15
16:00
Using Cloud forGlobal Collaboration – Demonstrations in CAE and EDA
David Pellerin,HPC Business Development Principal,Amazon Web Services
16:00
16:45
MentorGraphics Design Collaboration
Julian Sun,Business Development Director,Mentor Graphics
David Pellerin,HPC Business Development Principal,Amazon Web Services
16:45
AWS Kinesis – Big Data Management and Analytics in Manufacturing
Ken Chan,Solutions Architect,Amazon Web Services
David Pellerin,HPC Business Development Principal,Amazon Web Services
17:15
ClosingRemarks and Q&A
James Tien,Sales & Business Development Manager,Amazon Web Services
94. Big Data Management
For manufacturing
April 10, 2014
David Pellerin, Principal Business Development Manager, HPC
Amazon Web Services
95. Motivator: reduce the time spent searching for data
Aggregate data to a common platform, with common access tools
Improve manufacturing yields by accessing more data in a more timely manner
Speed up the yield improvement ramp up on new products
Improve steady state yield on existing products
Provide end-to-end visibility into:
Every test, every diagnostic
Data generated from all components of a product
Data generated internally, and from field deployments
Big Data Analytics in Manufacturing
96. Scenarios Accelerated Ingest-Transform-Load Continual Metrics/KPI Extraction Responsive Data Analysis
Software/
Technology
IT server , App logs ingestion IT operational metrics dashboards Devices / Sensor Operational
Intelligence
Digital Ad Tech./
Marketing
Advertising Data aggregation Advertising metrics like coverage, yield,
conversion
Analytics on User engagement with
Ads, Optimized bid/ buy engines
Financial Services Market/ Financial Transaction order data
collection
Financial market data metrics Fraud monitoring, and Value-at-Risk
assessment, Auditing of market order
data
Manufacturing Production line and field repair data
collection and aggregation
Yield and failure analysis, batch and
real-time
Production monitoring systems,
embedded controllers, device logs
Consumer Online/
E-Commerce
Online customer engagement data
aggregation
Consumer engagement metrics like
page views, CTR
Customer clickstream analytics,
Recommendation engines
Scenarios Across Industry Segments
1 2 3
97. Metrics from HGST Big Data Platform pilot project:
Collecting >2M manufacturing/testing input files daily
Collecting from ~500 tables across 6 databases tens of millions of records daily
HGST’s BDP is demonstrating early benefits:
Example: HGST
Development Engineer: demonstrated the joining of data sets for detailed
logistics tracking—analyses that is very difficult to conduct with current
systems
Ops Engineer: a recent production issue required detailed historical data. Current systems
did not have the required retention for this data. However, the team was able to pull the data
from the BDP in minutes, as opposed to 3+ weeks to pull the data from tape archive
Development Engineer: obtained technical data from the BDP in
hours as opposed to 3+ weeks to pull from tape archive
DATA SEARCH
PARTIES
YIELD
98. Kinesis Architecture
Amazon Web Services
AZ AZ AZ
Durable, highly consistent storage replicates data
across three data centers (availability zones)
Aggregate and
archive to S3
Millions of
sources producing
100s of terabytes
per hour
Front
End
Authentication
Authorization
Ordered stream
of events supports
multiple readers
Real-time
dashboards
and alarms
Machine learning
algorithms or
sliding window
analytics
Aggregate analysis
in Hadoop or a
data warehouse
Inexpensive: $0.028 per million puts
99. Sending & Reading Data from Kinesis Streams
HTTP Post
AWS SDK
LOG4J
Flume
Fluentd
Get* APIs
Kinesis Client Library
+
ConnectorLibrary
Apache Storm
Amazon Elastic
MapReduce
Sending Reading
100. Possible Use-Case in ASIC Production
Processing
Input
Yield analysis
Manufacturing production monitoring
and logging
Logging
Log4J
Appender
push to
Kinesis
ElasticMapReduce
Hive
Pig
Cascading
MapReduce
pull from
101. 107
Easy Administration
Managed service for real-time streaming
data collection,processingandanalysis.
Simply create a new stream,set the desired
level of capacity,andlet the service handle
the rest.
Real-time Performance
Perform continual processingonstreaming
big data. Processinglatencies fall to a few
seconds,comparedwiththe minutes or
hours associatedwithbatchprocessing.
High Throughput. Elastic
Seamlessly scale to matchyour data
throughput rate and volume. Youcaneasily
scale up to gigabytes per second. The service
will scale up or downbasedon your
operational or business needs.
S3, Redshift, & DynamoDB Integration
Reliably collect,process,andtransformall of
your data in real-time & deliver to AWS data
stores of choice,withConnectors for S3,
Redshift,and DynamoDB.
Build Real-time Applications
Client libraries that enable developers to
design and operate real-time streamingdata
processingapplications.
Low Cost
Cost-efficient for workloads of any scale. You
canget startedby provisioninga small
stream,and pay low hourly rates only for
what youuse.
Amazon Kinesis: Key Developer Benefits