_version_ 1866916118910730240
author Gallo, Manuel Le
Khaddam-Aljameh, Riduan
Stanisavljevic, Milos
Vasilopoulos, Athanasios
Kersting, Benedikt
Dazzi, Martino
Karunaratne, Geethan
Braendli, Matthias
Singh, Abhairaj
Mueller, Silvia M.
Buechel, Julian
Timoneda, Xavier
Joshi, Vinay
Egger, Urs
Garofalo, Angelo
Petropoulos, Anastasios
Antonakopoulos, Theodore
Brew, Kevin
Choi, Samuel
Ok, Injo
Philip, Timothy
Chan, Victor
Silvestre, Claire
Ahsan, Ishtiaq
Saulnier, Nicole
Narayanan, Vijay
Francese, Pier Andrea
Eleftheriou, Evangelos
Sebastian, Abu
author_facet Gallo, Manuel Le
Khaddam-Aljameh, Riduan
Stanisavljevic, Milos
Vasilopoulos, Athanasios
Kersting, Benedikt
Dazzi, Martino
Karunaratne, Geethan
Braendli, Matthias
Singh, Abhairaj
Mueller, Silvia M.
Buechel, Julian
Timoneda, Xavier
Joshi, Vinay
Egger, Urs
Garofalo, Angelo
Petropoulos, Anastasios
Antonakopoulos, Theodore
Brew, Kevin
Choi, Samuel
Ok, Injo
Philip, Timothy
Chan, Victor
Silvestre, Claire
Ahsan, Ishtiaq
Saulnier, Nicole
Narayanan, Vijay
Francese, Pier Andrea
Eleftheriou, Evangelos
Sebastian, Abu
contents The need to repeatedly shuttle around synaptic weight values from memory to processing units has been a key source of energy inefficiency associated with hardware implementation of artificial neural networks. Analog in-memory computing (AIMC) with spatially instantiated synaptic weights holds high promise to overcome this challenge, by performing matrix-vector multiplications (MVMs) directly within the network weights stored on a chip to execute an inference workload. However, to achieve end-to-end improvements in latency and energy consumption, AIMC must be combined with on-chip digital operations and communication to move towards configurations in which a full inference workload is realized entirely on-chip. Moreover, it is highly desirable to achieve high MVM and inference accuracy without application-wise re-tuning of the chip. Here, we present a multi-core AIMC chip designed and fabricated in 14-nm complementary metal-oxide-semiconductor (CMOS) technology with backend-integrated phase-change memory (PCM). The fully-integrated chip features 64 256x256 AIMC cores interconnected via an on-chip communication network. It also implements the digital activation functions and processing involved in ResNet convolutional neural networks and long short-term memory (LSTM) networks. We demonstrate near software-equivalent inference accuracy with ResNet and LSTM networks while implementing all the computations associated with the weight layers and the activation functions on-chip. The chip can achieve a maximal throughput of 63.1 TOPS at an energy efficiency of 9.76 TOPS/W for 8-bit input/output matrix-vector multiplications.
format Preprint
id arxiv_https___arxiv_org_abs_2212_02872
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle A 64-core mixed-signal in-memory compute chip based on phase-change memory for deep neural network inference
Gallo, Manuel Le
Khaddam-Aljameh, Riduan
Stanisavljevic, Milos
Vasilopoulos, Athanasios
Kersting, Benedikt
Dazzi, Martino
Karunaratne, Geethan
Braendli, Matthias
Singh, Abhairaj
Mueller, Silvia M.
Buechel, Julian
Timoneda, Xavier
Joshi, Vinay
Egger, Urs
Garofalo, Angelo
Petropoulos, Anastasios
Antonakopoulos, Theodore
Brew, Kevin
Choi, Samuel
Ok, Injo
Philip, Timothy
Chan, Victor
Silvestre, Claire
Ahsan, Ishtiaq
Saulnier, Nicole
Narayanan, Vijay
Francese, Pier Andrea
Eleftheriou, Evangelos
Sebastian, Abu
Emerging Technologies
The need to repeatedly shuttle around synaptic weight values from memory to processing units has been a key source of energy inefficiency associated with hardware implementation of artificial neural networks. Analog in-memory computing (AIMC) with spatially instantiated synaptic weights holds high promise to overcome this challenge, by performing matrix-vector multiplications (MVMs) directly within the network weights stored on a chip to execute an inference workload. However, to achieve end-to-end improvements in latency and energy consumption, AIMC must be combined with on-chip digital operations and communication to move towards configurations in which a full inference workload is realized entirely on-chip. Moreover, it is highly desirable to achieve high MVM and inference accuracy without application-wise re-tuning of the chip. Here, we present a multi-core AIMC chip designed and fabricated in 14-nm complementary metal-oxide-semiconductor (CMOS) technology with backend-integrated phase-change memory (PCM). The fully-integrated chip features 64 256x256 AIMC cores interconnected via an on-chip communication network. It also implements the digital activation functions and processing involved in ResNet convolutional neural networks and long short-term memory (LSTM) networks. We demonstrate near software-equivalent inference accuracy with ResNet and LSTM networks while implementing all the computations associated with the weight layers and the activation functions on-chip. The chip can achieve a maximal throughput of 63.1 TOPS at an energy efficiency of 9.76 TOPS/W for 8-bit input/output matrix-vector multiplications.
title A 64-core mixed-signal in-memory compute chip based on phase-change memory for deep neural network inference
topic Emerging Technologies
url https://arxiv.org/abs/2212.02872