Value Prediction (VP) is a microarchitectural technique that speculatively breaks data dependencies to increase the available Instruction Level Parallelism (ILP) in general purpose processors. Despite recent proposals, VP remains expensive and has intricate interactions with several stages of the classical superscalar pipeline. In this paper, we re...
International audience
Embedding Artificial Intelligence onto low-power devices is a challenging task that has been partly overcome with recent advances in machine learning and hardware design. Presently, deep neural networks can be deployed on embedded targets to perform different tasks such as speech recognition,object detection or Human Activity Recognition. However, ...
Multicore parallelism involve inter-tasks interferences leading to execution timing uncertainties. This is important for safety concerns in industrial applications (automotive for instance). Existing solutions limit significantly the use of available computing resources for low-criticality tasks to keep strong timing guarantees on high-criticality ...
Emerging Non Volatile Memories (NVMs) are considered as potential candidates for replacing SRAM in future processor architectures. They offer higher density and near-zero leakage power, which is particularly interesting to reduce the overall system energy consumption. Nevertheless, NVMs can suffer from higher access costs in latency and dynamic ene...
Modern architectures adopt large on-chip cache memory hierarchies with more than two levels. While this improves performance, it has a certain cost in area and power consumption. In this paper, we consider an emerging non volatile memory technology, namely the Spin-Transfer Torque Magnetic RAM (STT-MRAM), with a powerful cache replacement policy in...
This paper deals with hardware implementations for lattice-based cryptography. Various CPA and CCA secure algorithms for LWE, RLWE and MLWE problems have been studied, parallelized, implemented and compared on FPGA using high-level synthesis. The impact of PRNG choices on the implementations performances and costs is also evaluated. HLS allows us t...
International audience
Driver drowsiness is one of the major causes of accidents and fatal road crashes, causing a high human andeconomic cost. Recently, automatic drowsiness detection has begun to be recognized as a promising solution, receiving growing attention from industry and academics. In this letter, we propose to embed a convolutional neural network (CNN)-based ...
