Servizi per la didattica
PORTALE DELLA DIDATTICA

Approximate computing paradigms: from component to application-level

01UJWIU

A.A. 2019/20

Lingua dell'insegnamento

Inglese

Corsi di studio

Dottorato di ricerca in Ingegneria Informatica E Dei Sistemi - Torino

Organizzazione dell'insegnamento
Didattica Ore
Lezioni 30
Docenti
Docente Qualifica Settore h.Lez h.Es h.Lab h.Tut Anni incarico
Savino Alessandro   Ricercatore a tempo det. L.240/10 art.24-B ING-INF/05 10 0 0 0 2
Collaboratori
Espandi

Didattica
SSD CFU Attivita' formative Ambiti disciplinari
*** N/A ***    
2019/20
PERIOD: SETTEMBRE This Course introduces basic and advanced topics on Approximate Computing. The lectures will follow a bottom-up approach: from component, up to application-level. More in detail, all the three levels of existing approximate computing techniques will be presented: hardware, data and computation. In addition, the course also provides an overview on dependability aspects related with Approximate Computing. The comprehensive overview will allow students to resort to Approximate Computing to solve computational efficiency issues in non-accuracy bounded problems.
PERIOD: SEPTEMBER This Course introduces basic and advanced topics on Approximate Computing. The lectures will follow a bottom-up approach: from component, up to application-level. More in detail, all the three levels of existing approximate computing techniques will be presented: hardware, data and computation. In addition, the course also provides an overview on dependability aspects related with Approximate Computing. The comprehensive overview will allow students to resort to Approximate Computing to solve computational efficiency issues in non-accuracy bounded problems.
• General introduction o Definition(s) o Overview and classification of techniques o Metrics • Techniques for approximate computing o Data level approximation  Data representation - arithmetic  Adaptive Precision scaling  Less data & Less up-to-date data o Hardware level approximation  Exact integer and floating-point operators  Inexact operators  Voltage over-scaling and overclocking  Approximate memories  Custom floating-point & fixed-point o Computation level approximation  Computation skip (Fine grained, coarse grained)  Computation approximation (Algorithm selection, Memoization, CNN replacement) • Methods and tools for approximate computing
• General introduction o Definition(s) o Overview and classification of techniques o Metrics • Techniques for approximate computing o Data level approximation  Data representation - arithmetic  Adaptive Precision scaling  Less data & Less up-to-date data o Hardware level approximation  Exact integer and floating-point operators  Inexact operators  Voltage over-scaling and overclocking  Approximate memories  Custom floating-point & fixed-point o Computation level approximation  Computation skip (Fine grained, coarse grained)  Computation approximation (Algorithm selection, Memoization, CNN replacement) • Methods and tools for approximate computing
inizio 8 settembre dalla 10 alle 13
inizio 8 settembre dalla 10 alle 13
Modalità di esame:
Exam:
Esporta Word


© Politecnico di Torino
Corso Duca degli Abruzzi, 24 - 10129 Torino, ITALY
Contatti