no code implementations • 14 Jun 2018 • Behzad Salami, Osman Unsal, Adrian Cristal
Machine Learning (ML) is making a strong resurgence in tune with the massive generation of unstructured data which in turn requires massive computational resources.
no code implementations • 29 Mar 2019 • Behzad Salami, Osman S. Unsal, Adrian Cristal Kestelman
In consequence, we achieve 40% of the BRAM power saving through undervolting below the minimum safe voltage level, with a negligible NN accuracy loss, thanks to the substantial fault coverage by the built-in ECC.
2 code implementations • 30 Nov 2019 • Muhammet Abdullah Soyturk, Konstantinos Parasyris, Behzad Salami, Osman Unsal, Gulay Yalcin, Leonardo Bautista Gomez
During our study we corrupt the L1-Dcache of the simulated system and we monitor the behavior of the two types of fault maps on the resiliency of six benchmarks.
Performance Hardware Architecture
no code implementations • 26 Dec 2019 • Kamyar Givaki, Behzad Salami, Reza Hojabr, S. M. Reza Tayaranian, Ahmad Khonsari, Dara Rahmati, Saeid Gorgin, Adrian Cristal, Osman S. Unsal
This paper experimentally evaluates the resilience of the training phase of DNNs in the presence of voltage underscaling related faults of FPGAs, especially in on-chip memories.
no code implementations • 4 May 2020 • Behzad Salami, Erhan Baturay Onural, Ismail Emir Yuksel, Fahrettin Koc, Oguz Ergin, Adrian Cristal Kestelman, Osman S. Unsal, Hamid Sarbazi-Azad, Onur Mutlu
We evaluate the reliability-power trade-off for such accelerators.
no code implementations • 10 May 2020 • Behzad Salami, Osman Unsal, Adrian Cristal
Based on the characterization of these undervolting faults, we propose fault mitigation techniques that can effectively improve the resilience behavior of such accelerator.
no code implementations • 10 Nov 2022 • Aditya Manglik, Minesh Patel, Haiyu Mao, Behzad Salami, Jisung Park, Lois Orosa, Onur Mutlu
Resistive Random-Access Memory (RRAM) is well-suited to accelerate neural network (NN) workloads as RRAM-based Processing-in-Memory (PIM) architectures natively support highly-parallel multiply-accumulate (MAC) operations that form the backbone of most NN workloads.