Giuseppe Tagliavini

Giuseppe Tagliavini


Research Focus & Metrics

Academic Career Length

16Years (2011-Present)

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

Representative Publications

View All Publications
15 May 2019 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
APAntonio Pullini
Antonio Pullini
DRDavide Rossi
Davide Rossi
ILIgor Loi
Igor Loi
Giuseppe TagliaviniGiuseppe Tagliavini
Giuseppe Tagliavini
LBLuca Benini
Luca Benini
This paper presents Mr.Wolf, a parallel ultra-low power (PULP) system on chip (SoC) featuring a hierarchical architecture with a small (12 kgates) microcontroller (MCU) class RISC-V core augmented with an autonomous IO subsystem for efficient data transfer from a wide set of peripherals. The small core can offload compute-intensive kernels to an eight-core floating-point capable of processing engine available on demand. The proposed SoC, implemented in a 40-nm LP CMOS technology, features a 108-μW fully retentive memory (512 kB). The IO subsystem is capable of transferring up to 1.6 Gbit/s from external devices to the memory in less than 2.5 mW. The eight-core compute cluster achieves a peak performance of 850 million of 32-bit integer multiply and accumulate per second (MMAC/s) and 500 million of 32-bit floating-point multiply and accumulate per second (MFMAC/s) -1 GFlop/s--with an energy efficiency up to 15 MMAC/s/mW and 9 MFMAC/s/mW. These building blocks are supported by aggressive on-chip power conversion and management, enabling energy-proportional heterogeneous computing for always-on IoT end nodes improving performance by several orders of magnitude with respect to traditional single-core MCUs within a power envelope of 153 mW. We demonstrated the capabilities of the proposed SoC on a wide set of near-sensor processing kernels showing that Mr.Wolf can deliver performance up to 16.4 GOp/s with energy efficiency up to 274 MOp/s/mW on real-life applications, paving the way for always-on data analytics on high-bandwidth sensors at the edge of the Internet of Things.

18 March 2021 in IEEE Transactions on Computers
Verified
Journal Article
Closed access
ABAlessio Burrello
Alessio Burrello
AGAngelo Garofalo
Angelo Garofalo
NBNazareno Bruschi
Nazareno Bruschi
Giuseppe TagliaviniGiuseppe Tagliavini
Giuseppe Tagliavini
DRDavide Rossi
Davide Rossi
FCFrancesco Conti
Francesco Conti
The deployment of Deep Neural Networks (DNNs) on end-nodes at the extreme edge of the Internet-of-Things is a critical enabler to support pervasive Deep Learning-enhanced applications. Low-Cost MCU-based end-nodes have limited on-chip memory and often replace caches with scratchpads, to reduce area overheads and increase energy efficiency - requiring explicit DMA-based memory transfers between different levels of the memory hierarchy. Mapping modern DNNs on these systems requires aggressive topology-dependent tiling and double-buffering. In this work, we propose DORY (Deployment Oriented to memoRY) - an automatic tool to deploy DNNs on low cost MCUs with typically less than 1MB of on-chip SRAM memory. DORY abstracts tiling as a Constraint Programming (CP) problem: it maximizes L1 memory utilization under the topological constraints imposed by each DNN layer. Then, it generates ANSI C code to orchestrate off- and on-chip transfers and computation phases. Furthermore, to maximize speed, DORY augments the CP formulation with heuristics promoting performance-effective tile sizes. As a case study for DORY, we target GreenWaves Technologies GAP8, one of the most advanced parallel ultra-low power MCU-class devices on the market. On this device, DORY achieves up to 2.5× better MAC/cycle than the GreenWaves proprietary software solution and 18.1× better than the state-of-the-art result on an STM32-H743 MCU on single layers. Using our tool, GAP-8 can perform end-to-end inference of a 1.0-MobileNet-128 network consuming just 63 pJ/MAC on average @ 4.3 fps - 15.4× better than an STM32-H743. We release all our developments - the DORY framework, the optimized backend kernels, and the related heuristics - as open-source software.

6 October 2021 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
DRDavide Rossi
Davide Rossi
FCFrancesco Conti
Francesco Conti
MEManuel Eggiman
Manuel Eggiman
AMAlfio Di Mauro
Alfio Di Mauro
Giuseppe TagliaviniGiuseppe Tagliavini
Giuseppe Tagliavini
LBLuca Benini
Luca Benini
The Internet-of-Things (IoT) requires endnodes with ultra-low-power always-on capability for a long battery lifetime, as well as high performance, energy efficiency, and extreme flexibility to deal with complex and fast-evolving near-sensor analytics algorithms (NSAAs). We present Vega, an IoT endnode system on chip (SoC) capable of scaling from a 1.7- $\mu \text{W}$ fully retentive cognitive sleep mode up to 32.2-GOPS (at 49.4 mW) peak performance on NSAAs, including mobile deep neural network (DNN) inference, exploiting 1.6 MB of state-retentive SRAM, and 4 MB of non-volatile magnetoresistive random access memory (MRAM). To meet the performance and flexibility requirements of NSAAs, the SoC features ten RISC-V cores: one core for SoC and IO management and a nine-core cluster supporting multi-precision single instruction multiple data (SIMD) integer and floating-point (FP) computation. Vega achieves the state-of-the-art (SoA)-leading efficiency of 615 GOPS/W on 8-bit INT computation (boosted to 1.3 TOPS/W for 8-bit DNN inference with hardware acceleration). On FP computation, it achieves the SoA-leading efficiency of 79 and 129 GFLOPS/W on 32- and 16-bit FP, respectively. Two programmable machine learning (ML) accelerators boost energy efficiency in cognitive sleep and active states.

1 August 2015
Verified
Conference Paper
Closed access
DRDavide Rossi
Davide Rossi
FCFrancesco Conti
Francesco Conti
AMAndrea Marongiu
Andrea Marongiu
APAntonio Pullini
Antonio Pullini
ILIgor Loi
Igor Loi
LBLuca Benini
Luca Benini
Presents a collection of slides covering the following: PULP platform; Internet of Things; IoT applications; near-sensor processing; near-threshold multiprocessing; minimum energy operation; extended OpenRISC core; body biasing; ULP memory implementation; power management; heterogeneous memory architecture; fractal heterogeneity; and convolutional neural networks.

16 April 2021 in IEEE Transactions on Emerging Topics in Computing
Verified
Journal Article
Closed access
AGAngelo Garofalo
Angelo Garofalo
Giuseppe TagliaviniGiuseppe Tagliavini
Giuseppe Tagliavini
FCFrancesco Conti
Francesco Conti
LBLuca Benini
Luca Benini
DRDavide Rossi
Davide Rossi
Heavily quantized fixed-point arithmetic is becoming a common approach to deploy Convolutional Neural Networks (CNNs) on limited-memory low-power IoT end-nodes. However, this trend is narrowed by the lack of support for low-bitwidth in the arithmetic units of state-of-the-art embedded Microcontrollers (MCUs). This work proposes a multi-precision arithmetic unit fully integrated into a RISC-V processor at the micro-architectural and ISA level to boost the efficiency of heavily Quantized Neural Network (QNN) inference on microcontroller-class cores. By extending the ISA with nibble (4-bit) and crumb (2-bit) SIMD instructions, we show near-linear speedup with respect to higher precision integer computation on the key kernels for QNN computation. Also, we propose a custom execution paradigm for SIMD sum-of-dot-product operations, which consists of fusing a dot product with a load operation, with an up to 1.64 × peak MAC/cycle improvement compared to a standard execution scenario. To further push the efficiency, we integrate the RISC-V extended core in a parallel cluster of 8 processors, with near-linear improvement with respect to a single core architecture. To evaluate the proposed extensions, we fully implement the cluster of processors in GF22FDX technology. QNN convolution kernels on a parallel cluster implementing the proposed extension run 6 × and 8 × faster when considering 4- and 2-bit data operands, respectively, compared to a baseline processing cluster only supporting 8-bit SIMD instructions. With a peak of 2.22 TOPs/s/W, the proposed solution achieves efficiency levels comparable with dedicated DNN inference accelerators and up to three orders of magnitude better than state-of-the-art ARM Cortex-M based microcontroller systems such as the low-end STM32L4 MCU and the high-end STM32H7 MCU.

9 July 2019 in IEEE Transactions on Biomedical Circuits and Systems
Verified
Journal Article
Closed access
VKVictor Kartsch
Victor Kartsch
Giuseppe TagliaviniGiuseppe Tagliavini
Giuseppe Tagliavini
MGMarco Guermandi
Marco Guermandi
SBSimone Benatti
Simone Benatti
DRDavide Rossi
Davide Rossi
LBLuca Benini
Luca Benini
Advancements in Digital Signal Processing (DSP) and machine learning techniques have boosted the popularity of Brain-Computer Interfaces (BCIs), where electroencephalography (EEG) is a widely accepted method to enable intuitive human-machine interaction. Nevertheless, the evolution of such interfaces is currently hampered by the unavailability of embedded platforms capable of delivering the required computational power at high energy efficiency and allowing for a small and unobtrusive form factor. To fill this gap, we developed BioWolf, a highly wearable (40x20x2mm) BCI platform based on Mr. Wolf, a Parallel Ultra Low Power system on chip featuring nine RISC-V cores with DSP-oriented instruction set extensions. BioWolf also integrates a commercial 8-channel medical-grade analog-to-digital converter, and an ARM-Cortex M4 microcontroller unit (MCU) with Bluetooth low-energy connectivity. To demonstrate the capabilities of the system, we implemented and tested a BCI featuring Canonical Correlation Analysis (CCA) of steady-state visual evoked potentials (SSVEP). The system achieves an average information transfer rate of 1.46 bits per second (bps) (aligned with the state-of-the-art of bench-top systems). Thanks to the reduced power envelope of the digital computational platform, which consumes less than the analog front-end, the total power budget is just 6.31 mW, providing up to 38 hrs operation (65 mAh battery). To our knowledge, our design is the first to explore the significant energy boost of a parallel MCU with respect to single-core MCUs for CCA-based BCI.