Amlan Ganguly

Dr. Amlan Ganguly

Department of Computer Engineering, Rochester Institute of Technology, Rochester, NY 14623, USA
orcid

Research Focus & Metrics

Academic Career Length

14Years (2013-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
21 March 2022 in IEEE Design & Test
Verified
Journal Article
Closed access
SBSathwika Bavikadi
Sathwika Bavikadi
ADAbhijitt Dhavlle
Abhijitt Dhavlle
Amlan GangulyAmlan Ganguly
Amlan Ganguly
AHAnand Haridass
Anand Haridass
HHHagar Hendy
Hagar Hendy
SDSai Manoj Pudukotai Dinakarrao
Sai Manoj Pudukotai Dinakarrao
Advanced computing systems have long been enablers for breakthroughs in artificial intelligence (AI) and machine learning (ML) algorithms, either through sheer computational power or form-factor miniaturization. However, as AI/ML algorithms become more complex and the size of data sets increases, existing computing platforms are no longer sufficient to bridge the gap between algorithmic innovation and hardware design. This article presents a survey about various ML accelerators.

23 July 2020 in IEEE Computer Architecture Letters
Verified
Journal Article
Closed access
PSPurab Ranjan Sutradhar
Purab Ranjan Sutradhar
MCMark Connolly
Mark Connolly
SBSathwika Bavikadi
Sathwika Bavikadi
SDSai Manoj Pudukotai Dinakarrao
Sai Manoj Pudukotai Dinakarrao
MIMark Indovina
Mark Indovina
Amlan GangulyAmlan Ganguly
Amlan Ganguly
Memory access latencies and low data transfer bandwidth limit the processing speed of many data intensive applications such as Convolutional Neural Networks (CNNs) in conventional Von Neumann architectures. Processing in Memory (PIM) is envisioned as a potential hardware solution for such applications as the data access bottlenecks can be avoided in PIM by performing computations within the memory die. However, PIM realizations with logic-based complex processing units within the memory present complicated fabrication challenges. In this letter, we propose to leverage the existing memory infrastructure to implement a programmable PIM (pPIM), a novel Look-Up-Table (LUT)-based PIM where all the processing units are implemented solely with LUTs, as opposed to prior LUT-based PIM implementations that combine LUT with logic circuitry for computations. This enables pPIM to perform ultra-low power & low-latency operations with minimal fabrication complications. Moreover, the complete LUT-based design offers simple `memory write' based programmability in pPIM. Enabling precision scaling further improves the performance and the power consumption for CNN applications. The programmability feature potentially makes it easier for online training implementations. Our preliminary simulations demonstrate that our proposed pPIM can achieve 2000x, 657.5x and 1.46x improvement in inference throughput per unit power consumption compared to state-of-the-art conventional processor architecture, Graphics Processing Unit (GPUs) and a prior hybrid LUT-logic based PIM respectively. Furthermore, precision scaling improves the energy efficiency of the pPIM approximately by 1.35x over its full-precision operation.

17 March 2021 in IEEE Transactions on Parallel and Distributed Systems
Verified
Journal Article
Closed access
PSPurab Ranjan Sutradhar
Purab Ranjan Sutradhar
SBSathwika Bavikadi
Sathwika Bavikadi
MCMark Connolly
Mark Connolly
SPSavankumar Prajapati
Savankumar Prajapati
MIMark A. Indovina
Mark A. Indovina
Amlan GangulyAmlan Ganguly
Amlan Ganguly
Processing in memory (PIM) architecture, with its ability to perform ultra-low-latency parallel processing, is regarded as a more suitable alternative to von Neumann computing architectures for implementing data-intensive applications such as Deep Neural Networks (DNN) and Convolutional Neural Networks (CNN). In this article, we present a Look-up Table (LUT) based PIM architecture aimed at CNN/DNN acceleration that replaces logic-based processing with pre-calculated results stored inside the LUTs in order to perform complex computations on the DRAM memory platform. Our LUT-based DRAM-PIM architecture offers superior performance at a significantly higher energy-efficiency compared to the more conventional bit-wise parallel PIM architectures, while at the same time avoids fabrication challenges associated with the in-memory implementation of logic circuits. Alongside, the processing elements can be programmed and re-programmed to perform virtually any operation, including operations of Convolutional, Fully Connected, Pooling, and Activating Layers of CNN/DNN. Furthermore, it is capable of operating on several combinations of bit-widths of the operand data and thereby offers a wider range of flexibility across performance, precision, and efficiency. Transmission Gate (TG) realization of the circuitry ensures minimal footprint from the PIM architecture. Our simulations demonstrate that the proposed architecture can perform AlexNet inference at a nearly 13× faster rate and 125× more efficiency compared to state-of-the-art GPU and also provides 1.35× higher throughput at 2.5× higher energy-efficiency than another recent DRAM-implemented LUT-based PIM architecture in its baseline operation mode. Moreover, it offers 12× higher frame-rate at 9× more efficiency per frame for the lowest operand precision setting, with respect to its own baseline operation mode.

26 August 2019 in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Verified
Journal Article
Closed access
AVAbhishek Vashist
Abhishek Vashist
AKAndrew Keats
Andrew Keats
SDSai Manoj Pudukotai Dinakarrao
Sai Manoj Pudukotai Dinakarrao
Amlan GangulyAmlan Ganguly
Amlan Ganguly
Wireless networks-on-chips (NoCs) (WiNoCs) have emerged as a possible solution to the nonscalable multihop data transmission paths in traditional wired NoC architectures. Using low-power transceivers in NoC switches, novel WiNoC architectures have been shown to achieve higher energy efficiency with improved peak bandwidth and reduced on-chip data transfer latency. However, using wireless interconnects for intrachip data transfer over an unguided medium introduces additional security vulnerabilities in on-chip communication arising from either external attackers or internal hardware Trojans. In this article, we propose a mechanism to make the wireless communication in a WiNoC secure against persistent jamming-based denial-of-service (DoS) attacks and eavesdropping (ED) from both external and internal attackers. Persistent jamming attacks on the on-chip wireless medium will cause interference in data transfer over the duration of the attack resulting in errors in contiguous bits, known as burst errors. Therefore, we use a burst-error correction code to monitor the rate of burst errors received over the wireless medium and deploy a machine-learning (ML) classifier to detect the persistent jamming attack and distinguish it from random burst errors. In the event of a persistent jamming attack, alternate routing strategies are proposed to avoid the DoS attack over the wireless medium, so that a secure data transfer can be sustained even in the presence of persistent jamming. In the event of an external ED attack, we deploy a low-latency and lightweight data scrambling method to secure communication over the wireless channel. In the case of an internal ED, we propose a mechanism to identify the attacker and prevent the attack. We evaluate the proposed techniques on a WiNoC in the presence of DoS and ED attacks from both internal and external attackers. On an average, 99.87% of the attack on DoS detection was achieved with the chosen ML classifier. A bandwidth degradation of < 3% is experienced in the event of both DoS and ED internal attacks. The wireless interconnects are disabled in the presence of a persistent external jamming DoS attack for security, therefore eliminating the advantages of the wireless interconnections making the performance of the WiNoC comparable with that of a wired NoC. Although scrambling overheads are incurred in the presence of an external ED attack, the overheads are minimized by adopting simple XOR-based encoding and decoding.

Verified
Journal Article
Open access
PDF file
Amlan GangulyAmlan Ganguly
Amlan Ganguly
MAM Meraj Ahmed
M Meraj Ahmed
RNRounak Singh Narde
Rounak Singh Narde
AVAbhishek Vashist
Abhishek Vashist
SSShahriar Shamim
Shahriar Shamim
MIMark Indovina
Mark Indovina
With aggressive scaling of device geometries, density of manufacturing faults is expected to increase. Therefore, yield of complex Multi-Processor Systems-on-Chips (MP-SoCs) will decrease due to higher probability of manufacturing defects especially, in dies with large area. Therefore, disintegration of large SoCs into smaller chips called chiplets will improve yield and cost of complex platform-based systems. This will also provide functional flexibility, modular scalability as well as the capability to integrate heterogeneous architectures and technologies in a single unit. However, with scaling of the number of chiplets in such a system, the shared resources in the system such as the interconnection fabric and memory modules will become performance bottlenecks. Additionally, the integration of heterogeneous chiplets operating at different frequencies and voltages can be challenging. State-of-the-art inter-chip communication requires power-hungry high-speed I/O circuits and data transfer over long wired traces on substrates. This increases energy consumption and latency while decreasing data bandwidth for chip-to-chip communication. In this paper, we explore the advances and the challenges of interconnecting a multi-chip system with millimeter-wave (mm-wave) wireless interconnects from a variety of perspectives spanning multiple aspects of the wireless interconnection design. Our discussion on the recent advances include aspects such as interconnection topology, physical layer, Medium Access Control (MAC) and routing protocols. We also present some potential paradigm-shifting applications as well as complementary technologies of wireless inter-chip communications.

14 February 2022 in IEEE Micro
Verified
Journal Article
Closed access
Amlan GangulyAmlan Ganguly
Amlan Ganguly
Sergi AbadalSergi Abadal
Sergi Abadal
ITIshan Thakkar
Ishan Thakkar
NJNatalie Enright Jerger
Natalie Enright Jerger
MRMarc Riedel
Marc Riedel
BTBaris Taskin
Baris Taskin
The computing world is witnessing a proverbial Cambrian explosion of emerging paradigms propelled by applications, such as artificial intelligence, big data, and cybersecurity. The recent advances in technology to store digital data inside a deoxyribonucleic acid (DNA) strand, manipulate quantum bits (qubits), perform logical operations with photons, and perform computations inside memory systems are ushering in the era of emerging paradigms of DNA computing, quantum computing, optical computing, and in-memory computing. In an orthogonal direction, research on interconnect design using advanced electro-optic, wireless, and microfluidic technologies has shown promising solutions to the architectural limitations of traditional von-Neumann computers. In this article, experts present their comments on the role of interconnects in the emerging computing paradigms, and discuss the potential use of chiplet-based architectures for the heterogeneous integration of such technologies.