Swagath Venkataramani

Title

Principal Research Scientist, AIU Architecture and Compilers

Publications

Efficient Management of Scratch-Pad Memories in Deep Learning Accelerators
- - Subhankar Pal
  - Swagath Venkataramani
  - et al.
- 2021
- ISPASS 2021
A 7nm 4-Core AI Chip with 25.6TFLOPS Hybrid FP8 Training, 102.4TOPS INT4 Inference and Workload-Aware Throttling
- - Ankur Agrawal
  - Saekyu Lee
  - et al.
- 2021
- ISSCC 2021
Value Similarity Extensions for Approximate Computing in General-Purpose Processors
- - Younghoon Kim
  - Swagath Venkataramani
  - et al.
- 2021
- DATE 2021
Ultra-Low Precision 4-bit Training of Deep Neural Networks
- - Xiao Sun
  - Naigang Wang
  - et al.
- 2020
- NeurIPS 2020
ScaleCom: Scalable Sparsified Gradient Compression for Communication-Efficient Distributed Training
- - Chia-Yu Chen
  - Jiamin Ni
  - et al.
- 2020
- NeurIPS 2020
Efficient AI System Design with Cross-Layer Approximate Computing
- - Swagath Venkataramani
  - Xiao Sun
  - et al.
- 2020
- Proceedings of the IEEE
A 3.0 TFLOPS 0.62V Scalable Processor Core for High Compute Utilization AI Training and Inference
- - Jinwook Oh
  - Sae Kyu Lee
  - et al.
- 2020
- VLSI Circuits 2020
DyVEDeep: Dynamic Variable Effort Deep Neural Networks
- - Sanjay Ganapathy
  - Swagath Venkataramani
  - et al.
- 2020
- ACM TECS
Hybrid 8-bit floating point (HFP8) training and inference for deep neural networks
- - Xiao Sun
  - Jungwook Choi
  - et al.
- 2019
- NeurIPS 2019
Memory and Interconnect Optimizations for Peta-Scale Deep Learning Systems
- - Swagath Venkataramani
  - Vijayalakshmi Srinivasan
  - et al.
- 2019
- HiPC 2019

Patents

- 06 Nov 2023
- US
- 11810340
System And Method For Consensus-based Representation And Error Checking For Neural Networks
- 10 Oct 2023
- TW
- I818518
Single Function To Perform Combined Convolution And Select Operations
- 20 Aug 2023
- TW
- I813258
Reformatting Of Tensors To Provide Sub-tensors
- 05 Jun 2023
- US
- 11669489
Sparse Systolic Array Design
- 31 May 2023
- TW
- I804209
Exploiting Fine-grained Structured Weight Sparsity In Systolic Arrays
- 11 May 2023
- CN
- ZL202010150294.1
Programmable Data Delivery To A System Of Shared Processing Elements With Shared Memory
- 06 Mar 2023
- US
- 11599795
Reducing The Cost Of N Modular Redundancy For Neural Networks
- 16 Jan 2023
- US
- 11556450
Hybrid Data-model Parallelism For Efficient Deep Learning
- 09 Jan 2023
- US
- 11551054
System-aware Selective Quantization For Performance Optimized Distributed Deep Learning
- 07 Dec 2022
- JP
- 7190799
Low Precision Deep Neural Network Enabled By Compensation Instructions

Top collaborators

Xiaodong Cui

Principal Research Scientist

Alberto Mannari

Software Developer

Jinwook Jung

Research Staff Member

Mori Ohara

Deputy Director, IBM Research Tokyo, Distinguished Engineer, Chief SW Engineer for Hybrid Cloud on IBM HW