Publications by Babak Falsafi | EPFL Graph Search

A Low-latency On-chip Cache Hierarchy for Load-to-use Stall Reduction in GPUs

Memory hierarchy in Graphics Processing Units (GPUs) is conventionally designed to provide high bandwidth rather than low latency. In particular, because of the high tolerance to load-to-use latency (i.e., the time that warps wait for data fetched by memor ...

Association for Computing Machinery (ACM)2025

Single-Address-Space FaaS with Jord

Babak Falsafi, Atri Bhattacharyya, Mathias Josef Payer, Yuanlong Li, Sanidhya Kashyap, Abhishek Bhattacharjee

Function-as-a-Service (FaaS) has emerged as a popular cloud paradigm that simplifies software development and deployment by providing scalable and event-driven function execution without the burden of managing servers. FaaS was originally created with a fu ...

ACM2025

Avant-Garde: Empowering GPUs with Scaled Numeric Formats

Babak Falsafi

The escalating computational and memory demands of deep neural networks have outpaced chip density improvements, making arithmetic density a key bottleneck for GPUs. Scaled numeric formats, such as FP8 and Microscaling (MX), improve arithmetic density by a ...

ACM2025

Server Architecture from Enterprise to Post-Moore

Babak Falsafi

Luiz Barroso started his career at Digital Equipment Corporation, investigating workload-optimized multiprocessor server architectures marketed to enterprises in the 1990s. These high-margin, low-volume products lost their market to more cost-effective ent ...

2024

Imprecise Store Exceptions

Babak Falsafi, Siddharth Gupta, Mathias Josef Payer, Yunho Oh, Yuanlong Li, Qingxuan Kang, Abhishek Bhattacharjee

Precise exceptions are a cornerstone of modern computing as they provide the abstraction of sequential instruction execution to programmers while accommodating microarchitectural optimizations. However, increasing compute capabilities in deep memory hierar ...

ACM2023

AstriFlash: A Flash-Based System for Online Services

Babak Falsafi, Mark Johnathon Sutherland, Siddharth Gupta, Lei Yan, Yunho Oh

Modern datacenters host datasets in DRAM to offer large-scale online services with tight tail-latency requirements. Unfortunately, as DRAM is expensive and increasingly difficult to scale, datacenter operators are forced to consider denser storage technolo ...

2023

Cooperative Concurrency Control for Write-Intensive Key-Value Workloads

Babak Falsafi, Alexandros Daglis, Mark Johnathon Sutherland

Key-Value Stores (KVS) are foundational infrastructure components for online services. Due to their latency-critical nature, today’s best-performing KVS contain a plethora of full-stack optimizations commonly targeting read-mostly, popularity-skewed worklo ...

ACM2023

SecureCells: A Secure Compartmentalized Architecture

Babak Falsafi, Atri Bhattacharyya, Siddharth Gupta, Mathias Josef Payer, Andrés Sánchez Marín, Yuanlong Li, Florian Hofhammer

Modern programs are monolithic, combining code of varied provenance without isolation, all the while running on network-connected devices. A vulnerability in any component may compromise code and data of all other components. Compartmentalization separates ...

2023

What's Missing in Agile Hardware Design? Verification!

Babak Falsafi

Springer Singapore Pte Ltd2023

Scale-out Systolic Arrays

Pascal Frossard, Babak Falsafi, Ahmet Caner Yüzügüler

Multi-pod systolic arrays are emerging as the architecture of choice in DNN inference accelerators. Despite their potential, designing multi-pod systolic arrays to maximize effective throughput/Watt-i.e., throughput/Watt adjusted when accounting for array ...

ASSOC COMPUTING MACHINERY2023

Effective Interplay Between Sparsity and Quantization: from Theory to Practice

Babak Falsafi, Martin Jaggi

The increasing size of deep neural networks (DNNs) necessitates effective model compression to reduce their computational and memory footprints. Sparsity and quantization are two prominent compression methods that have been shown to reduce DNNs' computatio ...

2022

Cerebros: Evading the RPC Tax in Datacenters

Babak Falsafi, Alexandros Daglis, Arash Pourhabibi Zarandi, Mark Johnathon Sutherland

The emerging paradigm of microservices decomposes online services into fine-grained software modules frequently communicating over the datacenter network, often using Remote Procedure Calls (RPCs). Ongoing advancements in the network stack have exposed the ...

2021

Rebooting Virtual Memory with Midgard

Babak Falsafi, Atri Bhattacharyya, Siddharth Gupta, Mathias Josef Payer, Yunho Oh, Abhishek Bhattacharjee

Computer systems designers are building cache hierarchies with higher capacity to capture the ever-increasing working sets of modern workloads. Cache hierarchies with higher capacity improve system performance but shift the performance bottleneck to addres ...

2021

Equinox: Training (for Free) on a Custom Inference Accelerator

Babak Falsafi, Ahmet Caner Yüzügüler, Mario Paulo Drumond Lages De Oliveira, Arash Pourhabibi Zarandi, Louis Coulon, Martin Jaggi

DNN inference accelerators executing online services exhibit low average loads because of service demand variability, leading to poor resource utilization. Unfortunately, reclaiming idle inference cycles is difficult as other workloads can not execute on a ...

ACM2021

Data transformer apparatus

Babak Falsafi, Christoph Koch, Hussein Kassir, Mario Paulo Drumond Lages De Oliveira, Arash Pourhabibi Zarandi, Mark Johnathon Sutherland, Zilu Tian, Siddharth Gupta

Data transformer apparatus comprising at least a dispatcher module (D), a reader module (R), a converter module (C) and a writer module (W). The dispatcher module (D) is configured to: receive a data transformation request (DTR) including: a first informat ...

2021

Exploiting Errors for Efficiency: A Survey from Circuits to Applications

Damien Zufferey, Babak Falsafi, Sasa Misailovic, Eva Darulova, Dorde Jevdic, Alexandros Daglis, Mattia Cacciotti

When a computational task tolerates a relaxation of its specification or when an algorithm tolerates the effects of noise in its execution, hardware, system software, and programming language compilers or their runtime systems can trade deviations from cor ...

ASSOC COMPUTING MACHINERY2020

The NEBULA RPC-Optimized Architecture

Babak Falsafi, Alexandros Daglis, Mark Johnathon Sutherland, Dionysios Pnevmatikatos, Siddharth Gupta

Large-scale online services are commonly structured as a network of software tiers, which communicate over the datacenter network using RPCs. Ongoing trends towards software decomposition have led to the prevalence of tiers receiving and generating RPCs wi ...

2020

Optimus Prime: Accelerating Data Transformation in Servers

Babak Falsafi, Christoph Koch, Hussein Kassir, Mario Paulo Drumond Lages De Oliveira, Arash Pourhabibi Zarandi, Mark Johnathon Sutherland, Zilu Tian, Siddharth Gupta

Modern online services are shifting away from monolithic applications to loosely-coupled microservices because of their improved scalability, reliability, programmability and development velocity. Microservices communicating over the datacenter network req ...

ACM2020

RPCValet: NI-Driven Tail-Aware Balancing of µs-Scale RPCs

Babak Falsafi, Alexandros Daglis, Mark Johnathon Sutherland

Modern online services come with stringent quality requirements in terms of response time tail latency. Because of their decomposition into fine-grained communicating software layers, a single user request fans out into a plethora of short, μs-scale RPCs, ...

ACM2019

Mitigating Load Imbalance in Distributed Data Serving with Rack-Scale Memory Pooling

Babak Falsafi, Stanko Novakovic, Boris Robert Grot, Alexandros Daglis, Edouard Bugnion, Dmitrii Ustiugov

To provide low-latency and high-throughput guarantees, most large key-value stores keep the data in the memory of many servers. Despite the natural parallelism across lookups, the load imbalance, introduced by heavy skew in the popularity distribution of k ...

2019