impact of advanced electronics for dod today · 7/19/2017 · 65 nm cmos. 14 nm cmos pixel with...
TRANSCRIPT
Impact of Advanced Electronics for DoD Today
Dr. Jay Lewis, Deputy DirectorDARPA Microsystems Technology Office (MTO)
Electronics Resurgence Initiative Workshop – Day 2July 19, 2017
Reference herein to any specific commercial product, process, or service by trade name, trademark or other trade name, manufacturer or otherwise, does not necessarily constitute or imply endorsement by
DARPA, the Defense Department or the U.S. government, and shall not be used for advertising or product endorsement purposes.
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Leading-edge ASICs under development in MTO programs could deliver revolutionary capabilities to the warfighter
CLASS Disguise and dynamically vary signals for inexpensive LPI/LPD comms
SHIELD Verify the authenticity of components at every point in the supply chain
CLASIC Distinguish and classify RF signals for 180 hours on a cellphone battery
Enable real-time machine learning for object recognition on UAVs
DAHI 10x higher dynamic range arbitrary waveform generator for RF applications
ACT** Capture unprecedented volumes of RF data at 64Gs/sec for next-gen arrays
ReImagine Collect different data in a single camera frame with a reconfigurable ROIC
RF-FPGA A software-defined front end that works for 20GHz or below
Build trusted circuits through split integration
DISTRIBUTION A. Approved for public release: distribution unlimited.
ASIC – application specific integrated circuit RF – radio frequency
Example ASIC* Capability
ROIC – readout ICLPI/LPD – low probability of intercept/detection
*ASICs from MTO programs
SPADE
UPSIDE
**Program acronyms are later slide
Leading-edge ASICs under development in MTO programs could deliver revolutionary capabilities to the warfighter
CLASS Disguise and dynamically vary signals for inexpensive LPI/LPD comms
SHIELD Verify the authenticity of components at every point in the supply chain
CLASIC Distinguish and classify RF signals for 180 hours on a cellphone battery
Enable real-time machine learning for object recognition on UAVs
DAHI 10x higher dynamic range arbitrary waveform generator for RF applications
ACT** Capture unprecedented volumes of RF data at 64Gs/sec for next-gen arrays
ReImagine Collect different data in a single camera frame with a reconfigurable ROIC
RF-FPGA A software-defined front end that works for 20GHz or below
Build trusted circuits through split integration
DISTRIBUTION A. Approved for public release: distribution unlimited.
ASIC – application specific integrated circuit RF – radio frequency
Example ASIC* Capability
ROIC – readout ICLPI/LPD – low probability of intercept/detection
*ASICs from MTO programs
SPADE
UPSIDE
**Program acronyms on later slide
Leading-edge ASICs under development in MTO programs could deliver revolutionary capabilities to the warfighter
CLASS Disguise and dynamically vary signals for inexpensive LPI/LPD comms
SHIELD Verify the authenticity of components at every point in the supply chain
CLASIC Distinguish and classify RF signals for 180 hours on a cellphone battery
UPSIDE Enable real-time machine learning for object recognition on UAVs
DAHI 10x higher dynamic range arbitrary waveform generator for RF applications
ACT Capture unprecedented volumes of RF data at 64Gs/sec for next-gen arrays
ReImagine Collect different data in a single camera frame with a reconfigurable ROIC
RF-FPGA A software-defined front end that works for 20GHz or below
SPADE Build trusted circuits through split integration
DISTRIBUTION A. Approved for public release: distribution unlimited.
ASIC – application specific integrated circuitRF – radio frequency
Example ASIC* Capability
ROIC – readout ICLPI/LPD – low probability of intercept/detection*ASICs from MTO programs
6.8m
m
6.5mm
5mm
4mm
Area: 44.2mm2
50% Less Area: 20mm2
40% Less Power
32nm 14nm
ACT
• Enable digital re-use forphased arrays
• Capture unprecedentedvolumes of RF data at 64Gs/sec for next-gen arrays
• Leverage the world’s best digital beamforming system
32nm SOI vs. 14nm FinFet
Leading-edge ASICs under development in MTO programs could deliver revolutionary capabilities to the warfighter
CLASS Disguise and dynamically vary signals for inexpensive LPI/LPD comms
SHIELD Verify the authenticity of components at every point in the supply chain
CLASIC Distinguish and classify RF signals for 180 hours on a cellphone battery
UPSIDE Enable real-time machine learning for object recognition on UAVs
DAHI 10x higher dynamic range arbitrary waveform generator for RF applications
ACT Capture unprecedented volumes of RF data at 64Gs/sec for next-gen arrays
ReImagine Collect different data in a single camera frame with a reconfigurable ROIC
RF-FPGA A software-defined front end that works for 20GHz or below
SPADE Build trusted circuits through split integration
DISTRIBUTION A. Approved for public release: distribution unlimited.
ASIC – application specific integrated circuitRF – radio frequency
Example ASIC* Capability
ROIC – readout ICLPI/LPD – low probability of intercept/detection*ASICs from MTO programs
SHIELD
• Ensure the authenticityof genuine military electronic components
• Tag electronics at low cost with an encrypted 100µm x 100µm ASIC
• Personalization of millions of die at wafer-scale
14nm FinFet
Full AES encryption~30 µm x 30 µm
100
µm
100 µm
MicroscopicSHIELD dielet
Leading-edge ASICs under development in MTO programs could deliver revolutionary capabilities to the warfighter
CLASS Disguise and dynamically vary signals for inexpensive LPI/LPD comms
SHIELD Verify the authenticity of components at every point in the supply chain
CLASIC Distinguish and classify RF signals for 180 hours on a cellphone battery
UPSIDE Enable real-time machine learning for object recognition on UAVs
DAHI 10x higher dynamic range arbitrary waveform generator for RF applications
ACT Capture unprecedented volumes of RF data at 64Gs/sec for next-gen arrays
ReImagine Collect different data in a single camera frame with a reconfigurable ROIC
RF-FPGA A software-defined front end that works for 20GHz or below
SPADE Build trusted circuits through split integration
DISTRIBUTION A. Approved for public release: distribution unlimited.
ASIC – application specific integrated circuitRF – radio frequency
Example ASIC* Capability
ROIC – readout ICLPI/LPD – low probability of intercept/detection*ASICs from MTO programs
25 µm
SOA digital ROIC pixel
layout using 65 nm CMOS
14 nm CMOS pixel with
computation
~10 µm
~6
µmImages courtesy: MIT Lincoln Laboratory
ReImagine
• Achieve full battlespaceawareness with a single reconfigurable ROIC
• Simultaneously collect diverse data types from multiple regions of interest
• ADC with signalprocessing in every pixel
14nm CMOS
Program Names
CLASS Computational Leverage Against Surveillance Systems
SHIELD Supply Chain Hardware Integrity for Electronics Defense
CLASIC Cognitive radio Low-energy signal Analysis Sensor Integrated Circuits
UPSIDE Unconventional Processing of Signals for Intelligent Data Exploitation
DAHI Dense Accessible Heterogeneous Integration
ACT Arrays at Commercial Timescales
ReImagine Reconfigurable Imaging
RF-FPGA Radio Frequency – Field Programmable Gate Array
DISTRIBUTION A. Approved for public release: distribution unlimited.
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Ongoing FY18 Programs in AdvancedComputing and Design
Dr. William Chappell, DirectorDARPA Microsystems Technology Office (MTO)
Electronics Resurgence Initiative Workshop – Day 2July 19, 2017
Reference herein to any specific commercial product, process, or service by trade name, trademark or other trade name, manufacturer or otherwise, does not necessarily constitute or imply endorsement by
DARPA, the Defense Department or the U.S. government, and shall not be used for advertising or product endorsement purposes.
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
$141 million in Current Efforts (FY18)
$75 millionOf New Funding
(FY18)
$216 MILLION TOTAL (FY18)
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
materials architectures designs
JUMP + Traditional Programs
NATIONAL ELECTRONICS CAPABILITY
2025 - 2030
materials architectures designs• JUMP – Joint University Microelectronics Program
• CHIPS – Common Heterogeneous Integration and IP Reuse Strategies
• HIVE – Hierarchical Identify Verify Exploit
• L2M – Lifelong Learning Machines
• N-ZERO - Near-Zero Power Radio Frequency Receivers
• CRAFT – Circuit Realization at Faster Time Scales
• SSITH – System Security Integrated Through Hardware and firmware
Traditional Programs Currently Funded
JUMP + Traditional Programs
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
8/22/2016JUMP
Approved
MTO Electronics Timeline
4/2017HIVE
Kickoff
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
4/2016CRAFT Kickoff
1/2017L2M
Approved
11/2015
N-ZEROKickoff
2016 2017 2018 2019
Today
4/2017SSITH
BAA Released
6/2016CHIPS
Approved
• Ultra Low Power Design
• Reduced Design Time
• Pseudolithic Design
• Broad University Support
• In Field Machine Learning
• Graph Processing
• Built in Security
Joint University Microelectronics Program(JUMP)
RF to THz Distributed Computing
Cognitive Computing
IntelligentMemory/Storage
Advanced IC Architectures
Devices/Materials
Industry
40% 60%
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Joint University Microelectronics Program (JUMP)
Predecessor Program STARnet
• 646 Graduate students
• 184 Faculty researchers
• 46 Universities
• 6,118 Research publications
Circuit Realization at Faster Time Scales (CRAFT)
Driving a design methodology that can be used to quickly design flexible, high
performance custom integrated circuits using leading-edge CMOS technology while
driving DoD to use the best commercial fabrication and design practices
Sharply reduce barriers to DoD use of leading-edge custom integrated circuits (ICs) for orders-of-magnitude higher performance
at low power for DoD systems.
DoD (Today)
CRAFT(Future)
Chip Design and Fabrication Time (weeks)
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
N-ZERO passive sensor wake-up
Devices are off (zero power consumption) yet continually alert.
10000
1000
100
10
10.01 0.1 1.0 10 100
Unattended Ground Sensors
1 mo.
1 yr.
10 yrs.
Life
time
(Day
s)
Event Activity (% of Time)
Battery leakage, active processing and
N-ZERO wake-up
Battery drainage by active wake-
up circuitry
• Continuous operation and near-zero power processing• Persistent sensing with greatly extended lifetime and reduced cost• Multiple sensing modalities with sensor fusion
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Acoustic Sensor Wake-up
Wake-up to generator and truck at > 5m with 12 nW of power consumptionS. Jeong, et al. "21.6 A 12nW always-on acoustic sensing and object recognition microsystem using frequency-domain feature extraction and SVM classification." Solid-State Circuits Conference (ISSCC), 2017 IEEE International. IEEE, 2017.
Acoustic Signal
Sound of Interest+
Noise
Identify
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Hierarchical identify Verify Exploit (HIVE)
Enabling the DoD to perform graph analytics at the edge of the battlefield and not rely on datacenters back in the United
States; providing greater situational awareness in addition to the ability to do
real time sensor fusion and exploitation at the lowest echelons
Next-generation server processor designed to find patterns in streaming data sets by using graph analytics
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
HIVE: An example of blending commercial and DoD interests
CUDA – compute unified device architectureTPU – tensor processing unitGPU – graphics processing unit
TA 3:Evaluator (evaluation framework 100-1000x improvement)
TA 1: Graph analytics accelerator(ref: TPU, GPU)
TA 2:Graph analytics toolkits (ref: Tensorflow, CUDA)
Graph Software
What should be accelerated?
GraphHardware
How should it be accelerated?
Define graph primitives
Create data format model
Define data flow model
Identify/Develop hardware
accelerators for each building block
Create memorycontroller which optimizes data
movement based on sparse mapping
Develop busarchitectures
to avoid congestion in
data movement
Accelerators Memory Scaling
Cyber security
NorthropGrumman
Georgia TechPacific Northwest
Intel
Qualcomm
DISTRIBUTION A. Approved for public release: distribution unlimited.
Lifelong Learning Machines (L2M)
Pursuing approaches for biologically-inspired artificial intelligence utilizing
flexible models to continue adapting during execution in the field
Continual Learning
Mechanisms
Fundamentally new machine learning mechanisms for machines that learn continuously as they operate
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Cortical Processor results
DISTRIBUTION A. Approved for public release: distribution unlimited.
Adaptive Architectures
Results • Found optimal result278x faster than gridsearch method*
• Slightly higheraccuracy than thehand designed(Neverova et.al. 2015)*12 iterations vs ~3300
System modifies architecture via hyper-parameters and finds best
fusion processing paradigm
Determine Optimal Fusion Architecture• 21 classes, 3 modalities (ChaLearn, gesture)
SRI (Chai)
Dense Captioning
Context: learning picture elements and relationships instead of whole
image captioningContext Sensitivity
Requires localization and recurrence
Stanford (Li)
Learn Architecture
Collaborative Machine Intelligence Mitigating Catastrophic Forgetting
Classification error:• Set 1 – Pedestrians and bikers• Set 2 – Skateboarders and cars
Predict Motion using Social LSTM
Average Error vs. Predicted Location
Lifelong Learning
Collaborative Machine Intelligence
Modeled Predictions
Multiple Interacting Neural Networks
Algorithmic Solutions
Each network sees some of others: interacts intelligently
with dynamic worldFirst steps toward solving
catastrophic forgetting
Stanford-UTK (Savarese) Stanford-UTK (Savarese)
Cortical Processor results
DISTRIBUTION A. Approved for public release: distribution unlimited.
Common Heterogeneous Integration and IP Reuse Strategies (CHIPS)
Market IP Proprietary IP
6 –
12 W
eeks
Integrate
CHIPS enables rapid integration of modular circuits at the die level
Developing the design tools and integration standards required to demonstrate modular
integrated circuit (IC) designs that leverage the best of the DoD and commercial designs and
technology
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Predecessor is DAHI
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
300mm diameter Si CMOS wafer (45nm node)
Successful testing identified optimal S/H circuit for ADC
(>65dB SFDR @ 2GHz)Frequency (Hz) 10 9
0 1 2 3 4 5 6 7
Rel
ativ
e Am
plitu
de (d
B)
-90
-80
-70
-60
-50
-40
-30
-20
-10
0FCLK_15 GHz, FCW_771_clk_ttune_128_dem_en_0
DAC with very low digital noise
(-70dBc)
0
20
40
60
80
R2C3
M0
R3C4
M0
R4C3
M0
R5C4
M0
R2C3
M0
R3C4
M0
R4C3
M0
R5C4
M0
R2C3
M0
R3C4
M0
R4C3
M0
R5C4
M0
R2C3
M1
R3C4
M1
R4C3
M1
R5C4
M1
R2C3
M0
R3C4
M0
R4C3
M0
R5C4
M0
R2C3
M0
R3C4
M0
R4C3
M0
R5C4
M0
AR_2 AR_2A AR_1 AR_1A AR_1B
HIC Redundancy: None HIC Redundancy: 2x
HBT
Arra
y -B
eta
at 1
mA
Beta_812_@1mA Beta_813_@1mA Beta_814_@1mA Beta_815_@1mA Beta_862_@1mA Beta_863_@1mA Beta_864_@1mA Beta_865_@1mA Beta_872_@1mA Beta_873_@1mA Beta_874_@1mA Beta_875_@1mA
High foundry integration yields; test
vehicles fully functional
99.94% HIC yield98% HBT post-integration
DAHI integration (Dec 2015): Si (45nm), InP (TF5 HBT), GaN (GaN20 HEMT)
Sources: DARPA, Northrop GrummanDAHI – Diverse Accessible Heterogeneous Integration
CHIPS challenge: make a usable interface standardSource: DARPA
LVDS
32nm_SerDes
65nm_SerDes28nm_SerDes
28nm_SerDesProprietary
PCIe
HBM
WideIO
Si Photonic
1.00E+03
1.00E+04
1.00E+05
1.00E+06
1.00E+07
1.00E+08
1.00E+09
1.00E+10
1.00E+11
1.00E-05 1.00E-04 1.00E-03 1.00E-02 1.00E-01 1.00E+00
BWD
/EPB
Interconnect Length (m)
MonoLithic PCB
What standards will allow CHIPS to bridge the gap?
Interface standards:Too many? Not enough? How to compare?
Too many solutions can hinder wider adoption
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺/𝑚𝑚𝑚𝑚𝐸𝐸𝐸𝐸𝑒𝑒𝑟𝑟𝑔𝑔𝑦𝑦/𝐺𝐺𝑏𝑏𝑏𝑏
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
CHIPS challenge: make a usable interface standardSource: DARPA
LVDS
32nm_SerDes
65nm_SerDes28nm_SerDes
28nm_SerDesProprietary
PCIe
HBM
WideIO
Si Photonic
1.00E+03
1.00E+04
1.00E+05
1.00E+06
1.00E+07
1.00E+08
1.00E+09
1.00E+10
1.00E+11
1.00E-05 1.00E-04 1.00E-03 1.00E-02 1.00E-01 1.00E+00
BWD
/EPB
Interconnect Length (m)
MonoLithic PCB
Interface standards:Too many? Not enough? How to compare?
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺/𝑚𝑚𝑚𝑚𝐸𝐸𝐸𝐸𝑒𝑒𝑟𝑟𝑔𝑔𝑦𝑦/𝐺𝐺𝑏𝑏𝑏𝑏
CHIPS Convergence to a minimal set of standards is necessary
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
MTO Electronics Timeline
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
2016 2017 2018 2019
JUMPUniversity Driven
Page 3 InvestmentsIndustry Driven
• N-ZERO• CRAFT• HIVE• L2M• CHIPS• SSITH
Traditional Programs
8/22/2016JUMP
Approved
4/2017HIVE
Kickoff
4/2016CRAFT Kickoff
1/2017L2M
Approved
11/2015
N-ZEROKickoff
Today
4/2017SSITH
BAA Released
6/2016CHIPS
Approved
Linton SalmonMaterials and Integration
JUMP + Traditional Programs
materials architectures designs
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Workshop Summary:
Tom RondeauWade ShenArchitectures
JUMP + Traditional Programs
materials architectures designs
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Workshop Summary:
Andreas OlofssonDesigns
JUMP + Traditional Programs
materials architectures designs
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Workshop Summary:
Proposals Requested
Proposals Submitted
Partners Selected
Funding Released
OctSepAugJulJunMay Dec Jan Apr
Launch, Learn, & Organize
6/21: Industry Discussion7/11: Defense Base
Summit
Open Competition
9/12: Proposals Requested (Expected)
Complete Contracting
4/20: Start Work
Summer of Listening
7/18: 2-day workshop onMaterials, Architectures,
Designs
V
V
Nov
Completed Happening Now Fall 2017 Spring 2018
7 months
Defense Base Summit
2-dayWorkshop
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
MTO Electronics Resurgence Initiative Timeline
(HIVE) (CRAFT) (Possible)
Defense and National Needs
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
Teaming Session
Advantest Cadence IBM Microsemi Qualcomm ST Research
Allvia, Inc. Cold Logic Intel Microsoft Quantum Semiconductor Teledyne
Alphacore Esperanto Technologies Intermolecular MonolithIC 3D QuickLogic Teradeep
Analog Devices Ethaphase Intrinsix Moon RF2BITS Texas Instruments
Applied Materials Fault Tolerant Technology Jariet Nanoshift Sage Design
Automation TSMC
ARM Ferric, Inc. Kryptos Solutions Novati Siemens Vista Ventures
Astrileux Flex Innovation Kyndi Nuvotronics Silicon Storage Technologies Xilinx
Avalanche Technology Google MaXentric NVIDIA Silvaco
Bell Labs HP Mentor Graphics PARC SRI International
BroadPak HRL Micron Photia Synopsys
Commercial Attendees
AFRL Georgia Tech Portland State University of Arkansas
University of Texas at Dallas
Arizona State Harvard Purdue University of Illinois at Chicago
University of Washington
ARL MIT Sandia National Lab
University of Massachusetts USC
Berkeley National Lab
Naval Research Lab Stanford University of
Michigan Virginia Tech
BRIDG NC State UC Berkeley University of Minnesota
Columbia University Northwestern UC Davis University of New
Mexico
Cyclotron Road Notre Dame UC San Diego University of Pennsylvania
DRAPER Oak Ridge National Lab UC Santa Cruz University of
Texas at Dallas
University/Research Attendees
BAE
Boeing
Leidos
Lockheed Martin
Northrop Grumman
Raytheon
Rockwell Collins
Defense Industry Attendees
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
(HIVE) (CRAFT) (Possible)
Defense and National Needs
Teaming Session #111:00 am – 12:30 pm
Fol lowed by:Lunch (12:30PM, Imperial Bal lroom) | Teaming Session #2 (2:00PM, Club Regent) | Workshop Close (4:00PM)
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
(HIVE) (CRAFT) (Possible)
Defense and National Needs
Teaming Session #22:00 pm – 4:00 pm
The workshop wil l adjourn informal ly at 4:00pm. Thank you for attending.
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.
DISTRIBUTION STATEMENT A. Approved for public release. Distribution is unlimited.