Stephane Chevobbe

Stephane Chevobbe

Auto-generated profile, based on publicly available information

Stephane Chevobbe

CEA IRFU, Université Paris-Saclay, route de Saclay, 91191 Gif sur Yvette Cedex, France

Research Focus & Metrics

Academic Career Length

18Years (2009-Present)

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

Representative Publications

View All Publications
7 January 2019 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
LMLaurent Millet
Laurent Millet
SCStephane Chevobbe
Stephane Chevobbe
CACaaliph Andriamisaina
Caaliph Andriamisaina
LBLamine Benaissa
Lamine Benaissa
EDEdouard Deschaseaux
Edouard Deschaseaux
MDMarc Duranton
Marc Duranton
This paper presents a 3-D stacked vision chip featuring in-focal-plane read-out tightly coupled with flexible computing architecture for configurable high-speed image analysis. The chip architecture is based on a scalable standalone structure integrating image sensor on the top tier and processing elements (PEs) plus memories in the bottom tier. By using 3-D stacking partitioning, our prototype benefits from backside illuminated pixels sensitivity, a fully parallel communication between image sensor and PEs for low-latency performances, while leaving enough room in the bottom tier to embed advanced computing features. One scalable structure embeds a 16x16 pixel array (or 64x64 pixels in high-resolution mode), associated with an 8-bit single instruction multiple data (SIMD) processor; fabricated in dual 130-nm 1P6M CMOS process. This paper exhibits a 5500 frames/s and 85 giga operations per second (GOPS)/W in low-resolution mode, with large kernels capabilities through eight directions interpixel communication. Multiflow capability is also demonstrated to execute different programs in different areas of the vision chip.

1 June 2025 in Journal of Instrumentation
Verified
Journal Article
Closed access
DNDamien Neyret
Damien Neyret
PBP. Baron
P. Baron
FBF. Bouyjou
F. Bouyjou
MBM. Bregant
M. Bregant
SCS. Chevobbe
S. Chevobbe
FCF. William da Costa
F. William da Costa
The SALSA chip is a future readout ASIC foreseen for the MPGD detectors, developed in the framework of the EIC collider project, to equip the MPGD trackers of the EPIC experiment. It is designed to be versatile, to be adapted to other usages of MPGD detectors like TPC or photon detectors. It integrates a frontend block and an ADC for each of the 64 channels, associated to a configurable DSP processor meant to correct data and reduce the raw data flux to limit the output bandwidth. It will be compatible with the continuous readout foreseen for the EPIC DAQ, but will also work in a triggered environment. Several prototypes are already produced in order to qualify the different blocks of the chip, in particular the frontend, the ADC and the clock generation. The next 32-channel prototype is currently under development and is planed to be produced in 2025. The final prototype will be produced and tested from 2026 for a production of the SALSA chip at the horizon of 2027.

1 January 2009 in International Journal of Reconfigurable Computing
Verified
Journal Article
Open access
PDF file
SCStéphane Chevobbe
Stéphane Chevobbe
SGStéphane Guyetant
Stéphane Guyetant
A predictive dynamic reconfiguration management service is described here, targeting a new generation of multicore SoC that embed multiple heterogeneous reconfigurable cores. The main goal of the service is to hide the reconfiguration overheads, thus permitting more dynamicity for reconfiguring. We describe the implementation of the reconfiguration service managing three heterogeneous cores; functional results are presented on generated multithreaded applications.

8 February 2023 in Sensors
Verified
Journal Article
Open access
PDF file
MLMaria Lepecq
Maria Lepecq
TDThomas Dalgaty
Thomas Dalgaty
FWFabre William
Fabre William
SCStéphane Chevobbe
Stéphane Chevobbe
Three-dimensional-integrated focal-plane array image processor chips offer new opportunities to implement highly parallelised computer vision algorithms directly inside sensors. Neural networks in particular can perform highly complex machine vision tasks, and therefore their efficient implementation in such imagers are of significant interest. However, studies with existing pixel-processor array chips have focused on the implementation of a subset of neural network components—notably convolutional kernels—on pixel processor arrays. In this work, we implement a continuous end-to-end pipeline for a convolutional neural network from the digitisation of incoming photons to the output prediction vector on a macropixel processor array chip (where a single processor acts on group of pixels). Our implementation performs inference at a rate between 265 and 309 frames per second, directly inside of the sensor, by exploiting the different levels of parallelism available.

1 April 2016
Verified
Conference Paper
Closed access
CACeline Azar
Celine Azar
SCStephane Chevobbe
Stephane Chevobbe
YLYves Lhuillier
Yves Lhuillier
JDJean-Philippe Diguet
Jean-Philippe Diguet
Embedded platforms are expected to integrate thousands of cores in the near future, and one of the major difficulties remains in scaling the interconnection network. We propose a novel Scalable NETwork concept, named SNet, that broadens the NoCs space to include a combined software/hardware dynamic routing methodology. It exploits a distributed, adaptive, non-supervised routing strategy, based on the ACO algorithm (Ant Colony Optimization), to define routing paths among communicating tasks. Data transmission is achieved by a small footprint hardware module, named DMC (Direct Management of Communications). SNet has the advantage of being highly flexible, and enables to define a wide range of routing topologies to match application requirements. In this paper, we apply the SNet paradigm to a manycore architecture and we evaluate its performances by running a high number of test cases.