Swagath Venkataramani

发表

4-bit Quantization of LSTM-based Speech Recognition Models pdf

Brian Kingsbury, George Saon, Swagath Venkataramani, 2021, Interspeech.

Energy-Efficient Object Detection Using Semantic Decomposition

K. Roy, A. Raghunathan, Swagath Venkataramani, 2015, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

OnSRAM: Efficient Inter-Node On-Chip Scratchpad Management in Deep Learning Accelerators

K. Gopalakrishnan, Swagath Venkataramani, V. Srinivasan, 2022 .

Ultra-Low Precision 4-bit Training of Deep Neural Networks

Swagath Venkataramani, Vijayalakshmi Srinivasan, Kailash Gopalakrishnan, 2020, NeurIPS.

ScaleCom: Scalable Sparsified Gradient Compression for Communication-Efficient Distributed Training

Swagath Venkataramani, Vijayalakshmi Srinivasan, Songtao Lu, 2021, NeurIPS.

DyHard-DNN: Even More DNN Acceleration with Dynamic Hardware Reconfiguration

Mircea R. Stan, Swagath Venkataramani, Pradip Bose, 2018, 2018 55th ACM/ESDA/IEEE Design Automation Conference (DAC).

Accelerating Inference and Language Model Fusion of Recurrent Neural Network Transducers via End-to-End 4-bit Quantization

Brian Kingsbury, K. Gopalakrishnan, G. Saon, 2022, ArXiv.

Compensated-DNN: Energy Efficient Low-Precision Deep Neural Networks by Compensating Quantization Errors

Swagath Venkataramani, Vijayalakshmi Srinivasan, Shubham Jain, 2018, 2018 55th ACM/ESDA/IEEE Design Automation Conference (DAC).

OnSRAM: Efficient Inter-Node On-Chip Scratchpad Management in Deep Learning Accelerators

K. Gopalakrishnan, Swagath Venkataramani, V. Srinivasan, 2022, ACM Trans. Embed. Comput. Syst..

Exploring Spin-Transfer-Torque Devices for Logic Applications

K. Roy, A. Raghunathan, Swagath Venkataramani, 2014, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

Supplementary Materials for AQD: Towards Accurate Quantized Object Detection

Steven K. Esser, Andrew G. Howard, Dmitry Kalenichenko, 2021 .

Scalable-effort classifiers for energy-efficient machine learning

Jie Liu, Anand Raghunathan, Swagath Venkataramani, 2015, DAC.

Approximate computing: An integrated cross-layer framework

Swagath Venkataramani, Swagath Venkataramani, 2016 .

RaPiD: AI Accelerator for Ultra-low Precision Training and Inference

Joel Silberman, Swagath Venkataramani, Matthew M. Ziegler, 2021, 2021 ACM/IEEE 48th Annual International Symposium on Computer Architecture (ISCA).

Performance-driven Programming of Multi-TFLOP Deep Learning Accelerators*

Swagath Venkataramani, Vijayalakshmi Srinivasan, Kailash Gopalakrishnan, 2019, 2019 IEEE International Symposium on Workload Characterization (IISWC).

DeepTools: Compiler and Execution Runtime Extensions for RaPiD AI Accelerator

Hiroshi Inoue, Swagath Venkataramani, Moriyoshi Ohara, 2019, IEEE Micro.

SparCE: Sparsity Aware General-Purpose Core Extensions to Accelerate Deep Neural Networks

Anand Raghunathan, Swagath Venkataramani, Sanchari Sen, 2017, IEEE Transactions on Computers.

A Scalable Multi-TeraOPS Core for AI Training and Inference

Swagath Venkataramani, Vijayalakshmi Srinivasan, George Gristede, 2018, IEEE Solid-State Circuits Letters.

PACT: Parameterized Clipping Activation for Quantized Neural Networks pdf

Swagath Venkataramani, Zhuo Wang, Vijayalakshmi Srinivasan, 2018, ArXiv.

Hybrid 8-bit Floating Point (HFP8) Training and Inference for Deep Neural Networks

Swagath Venkataramani, Vijayalakshmi Srinivasan, Kailash Gopalakrishnan, 2019, NeurIPS.

SCALEDEEP: A scalable compute architecture for learning and evaluating deep networks

Pradeep Dubey, Anand Raghunathan, Swagath Venkataramani, 2017, 2017 ACM/IEEE 44th Annual International Symposium on Computer Architecture (ISCA).

Approximate computing and the quest for computing efficiency

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2015, 2015 52nd ACM/EDAC/IEEE Design Automation Conference (DAC).

AxNN: Energy-efficient neuromorphic systems using approximate computing

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2014, 2014 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED).

BiScaled-DNN: Quantizing Long-tailed Datastructures with Two Scale Factors for Deep Neural Networks

Swagath Venkataramani, Vijayalakshmi Srinivasan, Shubham Jain, 2019, 2019 56th ACM/IEEE Design Automation Conference (DAC).

Exploiting approximate computing for deep learning acceleration

Swagath Venkataramani, Vijayalakshmi Srinivasan, Kailash Gopalakrishnan, 2018, 2018 Design, Automation & Test in Europe Conference & Exhibition (DATE).

A 7nm 4-Core AI Chip with 25.6TFLOPS Hybrid FP8 Training, 102.4TOPS INT4 Inference and Workload-Aware Throttling

Joel Silberman, Swagath Venkataramani, Mingu Kang, 2021, 2021 IEEE International Solid- State Circuits Conference (ISSCC).

Approximate computing for spiking neural networks

Anand Raghunathan, Swagath Venkataramani, Sanchari Sen, 2017, Design, Automation & Test in Europe Conference & Exhibition (DATE), 2017.

Exploiting On-Device Image Classification for Energy Efficiency in Ambient-Aware Systems

Jie Liu, Xian-Sheng Hua, Swagath Venkataramani, 2015, Mobile Cloud Visual Media Computing.

SALSA: Systematic logic synthesis of approximate circuits

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2012, DAC Design Automation Conference 2012.

Very Low Voltage (VLV) Design

Eric Cheng, Rajiv V. Joshi, Arun Paidimarri, 2017, 2017 IEEE International Conference on Computer Design (ICCD).

Accurate and Efficient 2-bit Quantized Neural Networks

Swagath Venkataramani, Vijayalakshmi Srinivasan, Zhuo Wang, 2019, MLSys.

STAG: Spintronic-Tape Architecture for GPGPU cache hierarchies

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2014, 2014 ACM/IEEE 41st International Symposium on Computer Architecture (ISCA).

Efficient embedded learning for IoT devices

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2016, 2016 21st Asia and South Pacific Design Automation Conference (ASP-DAC).

Compensated-DNN: Energy Efficient Low-Precision Deep Neural Networks by Compensating Quantization Errors

Swagath Venkataramani, Vijayalakshmi Srinivasan, Shubham Jain, 2018, 2018 55th ACM/ESDA/IEEE Design Automation Conference (DAC).

Approximate storage for energy efficient spintronic memories

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2015, 2015 52nd ACM/EDAC/IEEE Design Automation Conference (DAC).

Efficient AI System Design With Cross-Layer Approximate Computing

Swagath Venkataramani, Xiao Sun, Mingu Kang, 2020, Proceedings of the IEEE.

ASLAN: Synthesis of approximate sequential circuits

Kaushik Roy, Anand Raghunathan, Arnab Raha, 2014, 2014 Design, Automation & Test in Europe Conference & Exhibition (DATE).

A 3.0 TFLOPS 0.62V Scalable Processor Core for High Compute Utilization AI Training and Inference

Swagath Venkataramani, Xiao Sun, Mingu Kang, 2020, 2020 IEEE Symposium on VLSI Circuits.

Memory and Interconnect Optimizations for Peta-Scale Deep Learning Systems

Swagath Venkataramani, Philip Heidelberger, Vijayalakshmi Srinivasan, 2019, 2019 IEEE 26th International Conference on High Performance Computing, Data, and Analytics (HiPC).

Deep Compression of Pre-trained Transformer Models

Kaoutar El Maghraoui, Chia-Yu Chen, Naigang Wang, 2022, NeurIPS.

Workload-aware Automatic Parallelization for Multi-GPU DNN Training

Swagath Venkataramani, Wonyong Sung, Sungho Shin, 2018, ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

A Scalable Multi- TeraOPS Deep Learning Processor Core for AI Trainina and Inference

Joel Silberman, Swagath Venkataramani, Matthew M. Ziegler, 2018, 2018 IEEE Symposium on VLSI Circuits.

SAPPHIRE: An always-on context-aware computer vision system for portable devices

Paramvir Bahl, Xian-Sheng Hua, Swagath Venkataramani, 2015, 2015 Design, Automation & Test in Europe Conference & Exhibition (DATE).

Substitute-and-simplify: A unified design paradigm for approximate and quality configurable circuits

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2013, 2013 Design, Automation & Test in Europe Conference & Exhibition (DATE).

DyVEDeep: Dynamic Variable Effort Deep Neural Networks pdf

Anand Raghunathan, Swagath Venkataramani, Balaraman Ravindran, 2017, ArXiv.

Bridging the Accuracy Gap for 2-bit Quantized Neural Networks (QNN) pdf

Swagath Venkataramani, Zhuo Wang, Vijayalakshmi Srinivasan, 2018, ArXiv.

A 7-nm Four-Core Mixed-Precision AI Chip With 26.2-TFLOPS Hybrid-FP8 Training, 104.9-TOPS INT4 Inference, and Workload-Aware Throttling

Joel Silberman, Swagath Venkataramani, Vijayalakshmi Srinivasan, 2022, IEEE Journal of Solid-State Circuits.

DyHard-DNN: Even More DNN Acceleration with Dynamic Hardware Reconfiguration

Swagath Venkataramani, Pradip Bose, Alper Buyuktosunoglu, 2018, 2018 55th ACM/ESDA/IEEE Design Automation Conference (DAC).

Multiplier-less Artificial Neurons exploiting error resiliency for energy-efficient neural computing

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2016, 2016 Design, Automation & Test in Europe Conference & Exhibition (DATE).

Energy-Efficient Neural Computing with Approximate Multipliers

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2018, ACM J. Emerg. Technol. Comput. Syst..

Approximation through logic isolation for the design of quality configurable circuits

Anand Raghunathan, Swagath Venkataramani, Shubham Jain, 2016, 2016 Design, Automation & Test in Europe Conference & Exhibition (DATE).

Relax-and-Retime: A methodology for energy-efficient recovery based design

Anand Raghunathan, Swagath Venkataramani, Shankar Ganesh Ramasubramanian, 2013, 2013 50th ACM/EDAC/IEEE Design Automation Conference (DAC).

Approximate Computing and the Efficient Machine Learning Expedition

Hai Helen Li, J. Henkel, M. Tahoori, 2022, 2022 IEEE/ACM International Conference On Computer Aided Design (ICCAD).

Computing approximately, and efficiently

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2015, 2015 Design, Automation & Test in Europe Conference & Exhibition (DATE).

Dynamic Spike Bundling for Energy-Efficient Spiking Neural Networks

Anand Raghunathan, Swagath Venkataramani, Sarada Krithivasan, 2019, 2019 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED).

STAxCache: An approximate, energy efficient STT-MRAM cache

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2017, Design, Automation & Test in Europe Conference & Exhibition (DATE), 2017.

Quality programmable vector processors for approximate computing

Kaushik Roy, Anand Raghunathan, Swagath Venkataramani, 2013, 2013 46th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO).

Designing approximate circuits using clock overgating

Younghoon Kim, Kaushik Roy, Anand Raghunathan, 2016, 2016 53nd ACM/EDAC/IEEE Design Automation Conference (DAC).

Enhance DNN Adversarial Robustness and Efficiency via Injecting Noise to Non-Essential Neurons

Swagath Venkataramani, Zhenyu Liu, Garrett Gagnon, 2024, arXiv.org.

Efficient Management of Scratch-Pad Memories in Deep Learning Accelerators

Swagath Venkataramani, Vijayalakshmi Srinivasan, Kailash Gopalakrishnan, 2021, 2021 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS).

POSTER: Design Space Exploration for Performance Optimization of Deep Neural Networks on Shared Memory Accelerators

Swagath Venkataramani, Vijayalakshmi Srinivasan, Kailash Gopalakrishnan, 2017, 2017 26th International Conference on Parallel Architectures and Compilation Techniques (PACT).