A
Animesh Jain
Researcher at Amazon.com
Publications - 48
Citations - 577
Animesh Jain is an academic researcher from Amazon.com. The author has contributed to research in topics: Relativistic Heavy Ion Collider & Magnet. The author has an hindex of 10, co-authored 47 publications receiving 404 citations. Previous affiliations of Animesh Jain include Indian Institute of Technology (BHU) Varanasi & Indian Institute of Science.
Papers
More filters
Proceedings ArticleDOI
Gist: efficient data encoding for deep neural network training
TL;DR: This paper investigates widely used DNNs and finds that the major contributors to memory footprint are intermediate layer outputs (feature maps), and introduces a framework for DNN-layer-specific optimizations that significantly reduce this source of main memory pressure on GPUs.
Proceedings ArticleDOI
DeftNN: addressing bottlenecks for DNN execution on GPUs via synapse vector elimination and near-compute data fission
Parker Hill,Animesh Jain,Mason Hill,Babak Zamirai,Chang-Hong Hsu,Michael A. Laurenzano,Scott Mahlke,Lingjia Tang,Jason Mars +8 more
TL;DR: DeftNN is a GPU DNN execution framework that targets the key architectural bottlenecks of DNNs on GPUs to automatically and transparently improve execution performance, and is composed of two novel optimization techniques– synapse vector elimination, a technique that identifies non-contributing synapses in the DNN and carefully transforms data and removes the computation and data movement.
Journal ArticleDOI
Experimental analysis of nanofuel additives with magnetic fuel conditioning for diesel engine performance and emissions
Rashmi Rekha Sahoo,Animesh Jain +1 more
TL;DR: In this paper, a single-cylinder diesel engine with and without magnetic fuel conditioning at various locations on fuel line was tested with a 0.5% CuO (wt/wt) mass fraction nanoparticles and diesel fuel.
Proceedings ArticleDOI
Concise loads and stores: the case for an asymmetric compute-memory architecture for approximation
Animesh Jain,Parker Hill,Shih-Chieh Lin,Muneeb Khan,Emdadul Haque,Michael A. Laurenzano,Scott Mahlke,Lingjia Tang,Jason Mars +8 more
TL;DR: This paper introduces a novel approximate computing technique that decouples the format of data in the memory hierarchy from the format in the compute subsystem to significantly reduce the cost of storing and moving bits throughout theMemory hierarchy and improve application performance.
Posted Content
Efficient Execution of Quantized Deep Learning Models: A Compiler Approach.
TL;DR: This paper addresses the challenges of executing quantized deep learning models on diverse hardware platforms by proposing an augmented compiler approach that created a new dialect called Quantized Neural Network (QNN) that extends the compiler's internal representation with a quantization context.