“Nigel was part of the “15.396: Energy Ventures” group where we developed an innovative business plan for an energy-related start-up. Kevin Brokish was also part of this team. Nigel is a brilliant technical mind with cutting edge transistor domain knowledge. His ability to integrate his knowledge with that of other talented engineers as part of a creative process is very impressive. What truly distinguished Nigel from his peers is his ability to analyze complex business problems, develop thoughtful solutions and effectively advocate for his innovative business ideas. Nigel was a leader for our team and deserves much credit for a business plan that received an “A” in the class.”
Activity
-
Congrats Veerbhan, Nigel, Daniel and the entire Quadric team for being recognized for the second consecutive year on the EE Times | Electronic…
Congrats Veerbhan, Nigel, Daniel and the entire Quadric team for being recognized for the second consecutive year on the EE Times | Electronic…
Liked by Nigel Drego
-
Learn about the latest AI/ML breakthrough - KAN enables dramatically smaller model sizes with orders of magnitude lower power consumption…
Learn about the latest AI/ML breakthrough - KAN enables dramatically smaller model sizes with orders of magnitude lower power consumption…
Liked by Nigel Drego
-
After a ten year run (4 with VMware and 6+ with Nyansa (now a part of VMware)) - I decided to leave VMware/Broadcom at the end of March. Too many…
After a ten year run (4 with VMware and 6+ with Nyansa (now a part of VMware)) - I decided to leave VMware/Broadcom at the end of March. Too many…
Liked by Nigel Drego
Experience & Education
Publications
-
Reduction of Variation-Induced Energy Overhead in Multi-core Processors
IEEE Transaction on Computer Aided Design
Core-to-core variability in future many-core chip multi-processors (CMPs) negatively impacts energy. Under-performing cores necessitate increasing the system voltage to maintain homogeneous core performance, introducing an energy overhead. Multiple supply voltages can be used to mitigate the impact of delay variation in CMPs. In this paper, we carefully analyze the use of a local search algorithm to pick near-optimal supply voltages while meeting a fixed performance target. With two system…
Core-to-core variability in future many-core chip multi-processors (CMPs) negatively impacts energy. Under-performing cores necessitate increasing the system voltage to maintain homogeneous core performance, introducing an energy overhead. Multiple supply voltages can be used to mitigate the impact of delay variation in CMPs. In this paper, we carefully analyze the use of a local search algorithm to pick near-optimal supply voltages while meeting a fixed performance target. With two system voltages, we prove our algorithm selects the global optimum and in the more general multiple voltage case we develop quantitative bounds. Using a custom simulation methodology on a real processor core, we show that two system voltages provide the most incremental benefit, reducing the energy overhead relative to a single voltage by 59-75% and total energy by 6-16%. Additionally, the worst 5-15% of cores in such systems necessitate increasingly larger amounts of incremental energy for a constant incremental performance gain. Therefore, turning off or disabling these cores is beneficial to a joint performance-energy metric.
Other authorsSee publication -
All-Digital Circuits for Measurement of Spatial Variation in Digital Circuits
IEEE Journal of Solid-State Circuits
Increased variation in CMOS processes due to scaling results in greater reliance on accurate variation models in developing circuit methods to mitigate variation. This paper investigates spatial variation in digital circuit performance: we describe a test-chip in 90 nm CMOS containing all-digital measurement circuits capable of extracting accurate variation data. Specifically, we use replicated 64-bit Kogge-Stone adders, ring oscillators (ROs) of varying gate type and stage length and an…
Increased variation in CMOS processes due to scaling results in greater reliance on accurate variation models in developing circuit methods to mitigate variation. This paper investigates spatial variation in digital circuit performance: we describe a test-chip in 90 nm CMOS containing all-digital measurement circuits capable of extracting accurate variation data. Specifically, we use replicated 64-bit Kogge-Stone adders, ring oscillators (ROs) of varying gate type and stage length and an all-digital, sub-picosecond resolution delay measurement circuit to provide this data. Measurement data from the test-chips indicate that 1) relative variation is significantly larger in low-voltage domains, 2) within-die variation is spatially uncorrelated, and 3) die-to-die (or global) variation is strongly correlated, but degrades toward uncorrelated as the power-supply voltage is lowered. Lastly, extended analysis of the data reveals that systematic effects such as layout pattern dependencies or circuit structure can be misinterpreted as random but spatially-correlated variation. This suggests that circuit designers will reap more benefit from design tools capable of modeling systematic, position-dependent variation rather than spatially correlated, distance-dependent variation.
Other authorsSee publication -
Lack of Spatial Correlation in MOSFET Threshold Voltage and Implications for Dynamic Voltage Scaling,
IEEE Transactions on Semiconductor Manufacturing
Due to increased variation in modern process technology nodes, the spatial correlation of variation is a key issue for both modeling and design. We have created a large array test-structure to analyze the magnitude of spatial correlation of threshold voltage (VT) in a 180 nm CMOS process. The data from over 50 k measured devices per die indicates that there is no significant within-die spatial correlation in VT. Furthermore, the across-chip variation patterns between different die also do not…
Due to increased variation in modern process technology nodes, the spatial correlation of variation is a key issue for both modeling and design. We have created a large array test-structure to analyze the magnitude of spatial correlation of threshold voltage (VT) in a 180 nm CMOS process. The data from over 50 k measured devices per die indicates that there is no significant within-die spatial correlation in VT. Furthermore, the across-chip variation patterns between different die also do not correlate. This indicates that Random Dopant Fluctuation (RDF) is the primary mechanism responsible for VT variation and that relatively simple Monte Carlo-type analysis can capture the effects of such variation. While high performance digital logic circuits, at high VDD, can be strongly affected by spatially correlated channel length variation, we note that subthreshold logic will be primarily affected by random uncorrelated VT variation.
Other authorsSee publication -
An All-Digital, Highly Scalable Architecture for Measurement of Spatial Variation in Digital Circuits
IEEE Asian Solid-State Circuits Conference
Increased variation in CMOS processes due to scaling results in greater reliance on accurate variation models in developing circuit methods to mitigate variation. This paper investigates specific variation parameters and their measurement approach for use in such models, leading to critical considerations in aggressive voltage scaling systems. We describe a test-chip in 90 nm CMOS containing all-digital measurement circuits capable of extracting accurate variation data. Specifically, we use…
Increased variation in CMOS processes due to scaling results in greater reliance on accurate variation models in developing circuit methods to mitigate variation. This paper investigates specific variation parameters and their measurement approach for use in such models, leading to critical considerations in aggressive voltage scaling systems. We describe a test-chip in 90 nm CMOS containing all-digital measurement circuits capable of extracting accurate variation data. Specifically, we use replicated 64-bit Kogge-Stone adders, ring-oscillators (ROs) of varying gate type and stage length and an all-digital, sub-picosecond resolution delay measurement circuit to provide spatial variation data for digital circuits. Measurement data from the test-chips indicate that 1) relative variation is significantly larger in low-voltage domains, 2) within-die variation is spatially uncorrelated, and 3) die-to-die (or global) variation is strongly correlated, but degrades toward uncorrelated as the power-supply voltage is lowered.
Other authorsSee publication -
A Test-Structure to Efficiently Measure Threshold-Voltage Variation in Large MOSFET Arrays
Int. Symp. on Quality Electronic Design
A test-structure comprising a dual-slope integrating analog-to-digital converter, auto-zeroing circuitry, digital control logic and a large array of devices under test (DUTs) has been developed to isolate threshold voltage variation. Threshold-voltage (VT) isolation is achieved by testing all DUTs in the subthreshold regime where drain-to-source current is an exponential function of VT. Spice simulations show that the structure is at least an order of magnitude more sensitive to VT variation…
A test-structure comprising a dual-slope integrating analog-to-digital converter, auto-zeroing circuitry, digital control logic and a large array of devices under test (DUTs) has been developed to isolate threshold voltage variation. Threshold-voltage (VT) isolation is achieved by testing all DUTs in the subthreshold regime where drain-to-source current is an exponential function of VT. Spice simulations show that the structure is at least an order of magnitude more sensitive to VT variation than to channel length variation. This, in combination with a hierarchical access scheme and leakage control system, allows efficient characterization of DeltaVT for ~70,000 NMOS and ~70,000 PMOS devices in a dense 2mm x 2mm DUT array.
Other authorsSee publication -
A Low-Skew, Low-Jitter Receiver Circuit for On-Chip Optical Clock Distribution
MIT
My Master's Thesis
-
Off-detector electronics for a high-rate CSC detector
IEEE Transactions on Nuclear Science
Data acquisition (DAQ) electronics are described for a system of high-rate cathode strip chambers (CSC) in the forward region of A Toroidal LHC ApparatuS (ATLAS) muon spectrometer. The system provides serial streams of control signals for switched capacitor array analog memories on the chambers and accepts a total of nearly 294 Gbit/s in serial raw data streams from 64 chambers in the design configuration. Processing of the data is done in two stages, leading to an output bandwidth of 2.56…
Data acquisition (DAQ) electronics are described for a system of high-rate cathode strip chambers (CSC) in the forward region of A Toroidal LHC ApparatuS (ATLAS) muon spectrometer. The system provides serial streams of control signals for switched capacitor array analog memories on the chambers and accepts a total of nearly 294 Gbit/s in serial raw data streams from 64 chambers in the design configuration. Processing of the data is done in two stages, leading to an output bandwidth of 2.56 Gbit/s. The architecture of the system is described, as are some important signal processing algorithms and hardware implementation details. Although designed for a specific application, the architecture is sufficiently general to be used in other contexts.
Other authors -
Patents
-
A machine perception and dense algorithm integrated circuit
Issued US10474398
-
Digital currency mining circuitry with adaptable difficulty compare capabilities
Issued US 9942046
Cryptographic hashing circuitry such as mining circuitry used to mine digital currency may be formed on an integrated circuit. The hashing circuitry may include sequential rounds of register and logic circuitry that perform operations of a cryptographic protocol. A final hash output from the hashing circuitry may be checked using a difficulty comparison circuit to determine whether the hash output satisfies predetermined difficulty criteria. The difficulty comparison circuit may be configured…
Cryptographic hashing circuitry such as mining circuitry used to mine digital currency may be formed on an integrated circuit. The hashing circuitry may include sequential rounds of register and logic circuitry that perform operations of a cryptographic protocol. A final hash output from the hashing circuitry may be checked using a difficulty comparison circuit to determine whether the hash output satisfies predetermined difficulty criteria. The difficulty comparison circuit may be configured as a hardwired comparison circuit having logic gates for checking only a subset of bits in the hash output. The comparison circuit may be adapted to change the number of bits that is checked based on a target number of bits for comparison set by the Bitcoin protocol. Candidate solutions found using the hardwired comparison circuit may then be fed to a host controller that checks the entire hash output to determine whether the candidate solution is valid.
Other inventorsSee patent -
Systems and methods for flexibly optimizing processing circuit efficiency
Issued US 9659123
Circuit design equipment may design logic for a circuit. The design equipment may discover optimized design constraints and an optimized clock signal frequency for the circuit. The design equipment may output the discovered optimized clock signal frequency and design constraints to circuit fabrication equipment for fabricating the corresponding circuit. The design equipment may discover the optimized clock signal frequency and design constraints by populating a cost function with different…
Circuit design equipment may design logic for a circuit. The design equipment may discover optimized design constraints and an optimized clock signal frequency for the circuit. The design equipment may output the discovered optimized clock signal frequency and design constraints to circuit fabrication equipment for fabricating the corresponding circuit. The design equipment may discover the optimized clock signal frequency and design constraints by populating a cost function with different clock signal frequencies and different design constraint values. The cost function may be, for example, a multi-dimensional surface. The design equipment may identify a global minimum of the cost function and may identify the clock signal frequency and design constraint values that correspond to the global minimum as the optimized clock frequency and optimized design constraints to provide to circuit fabrication equipment.
Other inventorsSee patent
Courses
-
Energy Ventures
15.401
-
Introduction to Finance
15.403
-
Power Electronics
6.334
-
Sustainable Energy
10.291J
Languages
-
English
-
Recommendations received
4 people have recommended Nigel
Join now to viewMore activity by Nigel
-
https://lnkd.in/gEXrvT7v I will be at the Automotive Computing Conference this week in Detroit, learning about the latest trends in SoC design for…
https://lnkd.in/gEXrvT7v I will be at the Automotive Computing Conference this week in Detroit, learning about the latest trends in SoC design for…
Liked by Nigel Drego
-
Quadric is getting noticed! Today we were happy to see investment firm #Activant Capital include Quadric in their coverage of new disruptive AI/ML…
Quadric is getting noticed! Today we were happy to see investment firm #Activant Capital include Quadric in their coverage of new disruptive AI/ML…
Liked by Nigel Drego
-
I'm excited to announce that I have accepted a Deep Learning Compiler Engineer position from Quadric! I'd like to give thanks to Michael Leonard for…
I'm excited to announce that I have accepted a Deep Learning Compiler Engineer position from Quadric! I'd like to give thanks to Michael Leonard for…
Liked by Nigel Drego
-
Not often I get to be a part of a “first time ever,” but for the first time ever, the Iraqi Ground Forces sent 8x General Officers to the National…
Not often I get to be a part of a “first time ever,” but for the first time ever, the Iraqi Ground Forces sent 8x General Officers to the National…
Liked by Nigel Drego
-
Salesforce laying of 700 people, but they have 1,000 openings. Public Companies offer stability and opportunity? May be a decade ago. Wall Street…
Salesforce laying of 700 people, but they have 1,000 openings. Public Companies offer stability and opportunity? May be a decade ago. Wall Street…
Liked by Nigel Drego
-
Choosing the right format may sound like a simple task. It isn’t, but that choice is essential for efficient edge inferencing. By Brian…
Choosing the right format may sound like a simple task. It isn’t, but that choice is essential for efficient edge inferencing. By Brian…
Liked by Nigel Drego
-
🚀 Shaping the Future of Drug Pricing: SmithRx Leading the Charge! 🚀 We're thrilled for our founder Jake Frenz and SmithRx to be featured in Ron…
🚀 Shaping the Future of Drug Pricing: SmithRx Leading the Charge! 🚀 We're thrilled for our founder Jake Frenz and SmithRx to be featured in Ron…
Liked by Nigel Drego
-
I'm very excited to share this news and super proud of the team! Thank you Kevin Nazemi Ibrahim El Tatawy Tony Brancato Ben Knapp TTV Capital FPV…
I'm very excited to share this news and super proud of the team! Thank you Kevin Nazemi Ibrahim El Tatawy Tony Brancato Ben Knapp TTV Capital FPV…
Liked by Nigel Drego
-
Our Halloween release dropped yesterday! We now support llama-2! Plus, we've added support for: - YOLOv3 & v4 - A pile of Google Mediapipe…
Our Halloween release dropped yesterday! We now support llama-2! Plus, we've added support for: - YOLOv3 & v4 - A pile of Google Mediapipe…
Liked by Nigel Drego
-
Operator fusion is an optimization technique that improves the execution speed of deep neural network models by treating successive operators as one,…
Operator fusion is an optimization technique that improves the execution speed of deep neural network models by treating successive operators as one,…
Liked by Nigel Drego
-
Still don't know why they call this Baby Llama, but why not?
Still don't know why they call this Baby Llama, but why not?
Liked by Nigel Drego
-
Oct. 24, 2023 at Sheraton HsinChu. Learn about Quadric's offering at this technical workshop that covers the Chimera GPNPU, the SDK and development…
Oct. 24, 2023 at Sheraton HsinChu. Learn about Quadric's offering at this technical workshop that covers the Chimera GPNPU, the SDK and development…
Liked by Nigel Drego
-
👋 Please join me for the Marketing Cloud Keynote, where I’ll try to squeeze in as many references to 2019 Toronto Raptors Championship as humanly…
👋 Please join me for the Marketing Cloud Keynote, where I’ll try to squeeze in as many references to 2019 Toronto Raptors Championship as humanly…
Liked by Nigel Drego
-
A couple days ago, we discussed how the Quadric architecture enables deep operator fusion, which reduces data movement resulting in increased perf…
A couple days ago, we discussed how the Quadric architecture enables deep operator fusion, which reduces data movement resulting in increased perf…
Shared by Nigel Drego
-
We just announced support for #llms on Quadric's General Purpose NPU cores. Our software stack enabled speedy porting of the newest #llama2 model!…
We just announced support for #llms on Quadric's General Purpose NPU cores. Our software stack enabled speedy porting of the newest #llama2 model!…
Liked by Nigel Drego
Other similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More