3. Amazon is part of the
semiconductor and
electronics industry
We design our own silicon
devices, and we source from
a global supply chain
Amazon has multiple, globally
distributed silicon teams, for:
• Datacenter infrastructure
• Consumer devices
• Robotics and AI
• And more
We benefit from AWS in our
own IC development
4. Amazon Silicon
AWS Graviton/2
Powerful and efficient server
chip for modern applications
AWS Inferentia
Machine learning hardware
and software at scale
AWS Nitro System
Cloud hypervisor, network,
storage, and security
100% Developed in the Cloud: RTL GDSII
5. Our own journey—our own digital transformation
2011 2015
Annapurna Startup
Formed 2011, Israel
Started with on-prem datacenter
2014
AWS silicon
optimizations
Formed 2014, Austin
Born in the cloud
US expands
deployment in AWS
Israel expands
productivity via AWS
AWS
“One Team”
Acquisition of
Annapurna
Multi-site
Development
Hybrid
Model
On-prem data center Less infrastructure in on-
prem data center
Multi-site
Development
US expands
deployment in AWS
Multiple end-to-end
silicon projects in AWS
2016 2017 Today
Full SoC Development
in the cloud
Latest semiconductor
fab 7nm process
Multi-site
On-prem data center
only for emulators
6. Why AWS?
• Innovate faster
• Collaborate better
• Reduce risk
• Reduce cost
24 geographic regions
A region is a physical location
in the world where we have
multiple Availability Zones
76 Availability Zones
Distinct locations that are
engineered to be insulated
from failures in other
Availability Zones
Network
AWS offers highly reliable, low latency, and high
throughput network connectivity. This is achieved with
a fully redundant 100 Gbps network that circles the
globe.
7. Why EDA on AWS?
• Innovate faster – Prototype, design, and verify complex systems-on-chip, using
scalable cloud resources for Electronic Design Automation (EDA).
• Collaborate better – Work seamlessly and securely with third-party partners
including IP providers, EDA software vendors, and manufacturing service
providers (foundries, OSATs, contract and original device manufacturers).
• Reduce risk – Advanced silicon and system verification is hard, and getting harder.
Mistakes can cost millions if not billions of dollars for a larger companies.
• Reduce cost – Stop wasting CAPEX on IT, and stop wasting valuable engineering
time.
9. Electronic Design Automation Infrastructure
TRADITIONAL EDA IT STACK
EDA data center
Remote desktop
License managers
Workload schedulers
Directory services
Compute nodes
Shared file storage
10. EDA Infrastructure on AWS
AWS Cloud
Virtual Private Cloud on AWS
Fabrication, third-party IP
providers, collaborators
On-Prem Data Center
Job Master
Login Server
Directory Services
Remote Desktop
License Manager
User Access
Infrastructure Support
Compute
Storage
EC2 and Auto Scaling
Amazon EC2
Spot Instances
Containers
Home FS (EFS) Local diskScratch space
Amazon S3 and
Amazon Glacier
Machine Learning
and Analytics
AWS Snowball
AWS Direct
Connect
On AWS, secure and well-
optimized EDA clusters
can be automatically
created, operated, and
torn down in minutes
Encryption everywhere -
with your own keys!
12. High clock speed compute instances: z1d
Up to 4 GHz sustained, all-turbo performance
• Z1d instances are optimized for memory-intensive,
compute-intensive applications
• Custom Intel Xeon Scalable processor
• Up to 4 GHz sustained, all-turbo performance
• Up to 385GiB DDR4 memory
• Enhanced networking, up to 25 GB throughput
EDA stack on AWS
3D graphics virtual workstation
License managers and cluster
head nodes with job schedulers
Cloud-based, auto-scaling HPC clusters
Shared file storage Storage cache
13. High bandwidth instances: C5n, M5n, R5n
100 Gbps network performance
• C5n, M5n, and R5n instances offer up to 100 Gbps of
network bandwidth
• Significant improvements in maximum bandwidth, packet
per seconds, and packets processing
• Purpose-built to run network bound workloads including
distributed cluster and database workloads, HPC, real-time
communications and video streaming
EDA stack on AWS
3D graphics virtual workstation
License managers and cluster
head nodes with job schedulers
Cloud-based, auto-scaling HPC clusters
Shared file storage Storage cache
16. File system options for EDA
Amazon EFS
Scalable, elastic, cloud-
native file system for Linux
Amazon EC2 Amazon EBS
• Build your own self-managed NFS server
• AWS Marketplace
Amazon FSx for
Lustre
Fully managed shared file systems
for high performance computing
workloads
17. Fully managed high performance shared file system
SSD-basedParallel
distributed
file system
Supports hundreds
of thousands of
cores
Massively scalable performance
• 100+ GiB/ s throughput
• Millions of IOPS
• Consistent low latencies
Amazon FSx for Lustre EDA stack on AWS
Job Master
Login Server
Directory Services
Remote Desktop
License Manager
User Access
Infrastructure Support
Compute
EC2 and Auto Scaling
Amazon EC2
Spot Instances
Containers
Storage
Home FS (EFS) Local diskScratch space
18. Mapping storage to EDA data types
Tools
Project
IP libraries
Workspaces
Scratch
Home
TEMPORARYPERSISTENT
Amazon FSx for Lustre
Data type Storage solutions
READ-ONLYREAD-WRITEREAD-WRITE
DIY/ Marketplace
NFS server
DIY/ Marketplace
NFS server
Amazon FSx for Lustre
DIY/ Marketplace
NFS server
Amazon EFS
Amazon FSx for Lustre
19. Technology partners for silicon design: examples
Design
verification
Synthesis
Physical
layout
Physical
verification
Tape out/
manufacturing
Power/ signal
analysis
Design
specification
Silicon
validation
Front-end design Back-end design Production and test
Phase
Full flow providers
Design flow and file management
IP providers Extraction, power, EM
Quality, yield, and failure analysis High Volume Manufacturing (HVM)
20. Cloud enables secure collaboration
Yield/ Process
data
EDA Vendor
Secure
Chamber
IP Provider
Secure
Chamber
Semicon
Secure Chamber
Foundry
Secure Chamber
Fabless Semiconductor
Design Company
Foundry
AWS Cloud
AWS Cloud
Equipment
Provider
Secure
Chamber
AWS CloudAWS Cloud
EDA
Tools
IP and
Libraries
PDK
GDSII
Curated OT and
Predictive data
Updated ML models
AWS Cloud
EDA/ IP collaboration
cloud chamber
AWS Cloud
Yield analytics cloud
chamber
AWS Cloud
Foundry IP merge
cloud chamber
AWS Cloud
Digital twin cloud
chambers
AWS Cloud
Transient, as needed
connections
Persistent connections to
chambered environments
21. Scale-out computing on AWS
aws.amazon.com/solutions/scale-out-computing-on-aws
Framework behind Amazon Devices
Lab126 HPC environment
Enables engineers/ scientists with
minimal cloud and/ or Linux
experience
Official AWS Solution:
“Vetted, technical reference
implementations designed to help
you solve common problems and
build faster”
22. Scale-out computing on AWS
aws.amazon.com/solutions/scale-out-computing-on-aws
Amazon
S3
Amazon Elastic
File System
Amazon FSx
for Lustre
Users
(access web UI,
DCV, ssh to
scheduler)
Elastic Load
Balancing (manages
access)
DCV graphical
sessions
web UI
Amazon EC2
(scheduler
instance)
python scripts
(used to run jobs)
Amazon EC2
Auto Scaling
(launch
instances to run
jobs)
Amazon
Elasticsearch
Service
(stores job & host
information)
AWS Secrets Manager
(stores cluster
information)
(storage options
for either
persistent or
ephemeral data)
• AWS Solution
• EDA/ HPC environment on AWS
• Easy installation in your AWS account
• Amazon EC2 Integration
• Simple job submission
• OS agnostic and AMI support
• Desktop cloud visualization
• Automatic errors handling
• Web UI
• 100% customizable
• Persistent and unlimited storage
• Centralized user-management
• Support for network licenses
• EFA support
• Simple cost/ budget management
• Detailed cluster analytics
• Used in production
27. Challenges from COVID-19 Pandemic
• Advance technology IC design project 6nm
• Security control
• EDA workload
• Computing scale
• Data Storage
• Pandemic impacts project schedule
unexpectedly
• Work from Home
28. AWS Cloud to Solve Problem
• Security environment certified by foundry
• Build up infrastructure quickly
• Give proven EDA running experience
• Smooth data transfer between cloud & office
29. Access
Control
List
EDA Architecture on AWS Overview
VPC
Private subnet Amazon
CloudWatch
Public subnet
Secure
FTP
Computing
Box
EDA
License
AWS Cloud (US-West-1)
Project Data
Foundry Data
EDA Tool
Internet
Local Site
BJ Site
Desktop
Cloud
Virtualization
AWS
CloudTrail
AWS
GuardDuty
33. EDA Infrastructure on AWS – AI/ML
AWS Cloud
Virtual Private Cloud on AWS
Fabrication, third-party IP
providers, collaborators
On-Prem Data Center
Job Master
Login Server
Directory Services
Remote Desktop
License Manager
User Access
Infrastructure Support
Compute
Storage
EC2 and Auto Scaling
Amazon EC2
Spot Instances
Containers
Home FS (EFS) Local diskScratch space
Amazon S3 and
Amazon Glacier
Machine Learning
and Analytics
AWS Snowball
AWS Direct
Connect
On AWS, secure and well-
optimized EDA clusters
can be automatically
created, operated, and
torn down in minutes
Encryption everywhere -
with your own keys!
36. Tips for EDA in the cloud
1. Leadership Alignment
2. Think big but start small
• Don’t try to do seamless bursting or cloud-native workflows at first.
• Start with EDA workloads or projects that are important but not critical and have few on-prem dependencies.
• Build a controlled cloud environment with qualified flows and a trained set of users who know they are in the cloud.
3. Stay familiar
• Start with a familiar environment to leverage your staff’s expertise where you can.
• Use the AWS integration in commercial schedulers that are commonly found in EDA.
• If you are using NetApp, consider NetApp in the cloud.
• Leverage your DDM solutions to keep design data and libraries in sync.
4. Use EC2 Spot!
• You’re scale-out flows are likely already fault tolerant.
5. Centralize your data
• The more data sources you keep in AWS, the more options you have for machine learning and analytics.
6. Train your builders
• You already have the people you need to succeed in the cloud. Enable them.
1/For the second year in a row, we have been voted by the HPC community as the Best Cloud HPC Platform. We picked up 5 other awards as well to get our tally up to 6 this year.
2/The HPCWire Readers’ Choice and Editors’ Choice Awards have been around for 16 years and are a regular, well anticipated announcement at the Supercomputing show. Readers’ Choice Awards are voted on by 26,000+ subscribers of HPCWire.com.
3/ Also note that AWS/Amazon employees were not allowed to vote for AWS, so this is purely based our customer and partner feedback
4/HPC wire evaluated AWS’s C5n instances powered by 2nd Generation Intel Xeon Scalable Processors. with 100Gb networking and the AWS Nitro system.
Amazon EC2 z1d instances deliver high single thread performance due to a custom Intel® Xeon® Scalable processor with a sustained all core frequency of up to 4.0 GHz, the fastest of any cloud instance. z1d provides both high compute performance and high memory, which is ideal for electronic design automation (EDA), gaming, Optical Proximity Correction (computational lithography), HPC, financial, actuarial, data analytics, and certain relational database workloads with high per-core licensing costs.
z1d is also ideal for applications requiring high single-threaded performance and high memory usage.
Amazon EC2 z1d instances deliver high single thread performance due to a custom Intel® Xeon® Scalable processor with a sustained all core frequency of up to 4.0 GHz, the fastest of any cloud instance. z1d provides both high compute performance and high memory, which is ideal for electronic design automation (EDA), gaming, Optical Proximity Correction (computational lithography), HPC, financial, actuarial, data analytics, and certain relational database workloads with high per-core licensing costs.
z1d is also ideal for applications requiring high single-threaded performance and high memory usage.
DRC Design Rule Checking (DRC)
SW>>potentially move text on Left to notes. Simple/Clean
