by Phillip Stanley-Marbell on Oct 19, 2021 | Tags: Accelerators, Approximate Computing, Numerical Format, Sensors, Uncertainty
In Part 1 of this two-part post, I looked at some of the existing and possible avenues for computer architecture research relating to tracking uncertainty in computations, using the blackscholes benchmark from the PARSEC suite of computer architecture research benchmark applications as a working example. In this post, I’ll outline some existing and possible future paths for computer architects in computation with uncertainty. Just as architectural support and microarchitectural implementations of floating-point number representations improved the ease of implementation of real-valued computations, architectural and microarchitectural support for representations of uncertainty could enable new approaches to trustworthy computation on empirical data.
Read more...
by Simla Burcu Harma, Mario Drumond, Babak Falsafi on Sep 20, 2021 | Tags: Accelerators, Machine Learning, Numerical Format, Tools
DNN training is emerging as a popular compute-intensive workload. This blog post provides an overview of the recent research on numerical encoding formats for DNN training, and presents the Hybrid Block Floating-Point (HBFP) format which reduces silicon provisioning...
Read more...
by Payman Behnam and Mahdi Nazm Bojnordi on Apr 21, 2020 | Tags: Accelerators, Numerical Format
Motivation Number systems and computer arithmetics are essential for designing efficient hardware and software architecture. In particular, real-valued computation constitutes a crucial component in almost all forms of today’s computing systems from mobile devices to...
Read more...
