A cloud-scale acceleration architecture

By Adrian M. Caulfield, Eric S. Chung, Andrew Putnam, Hari Angepat, J. Fowers, M. Haselman, Stephen Heil, Matt Humphrey, P. Kaur, Joo-Young Kim and 8 more. 2016

604 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Hardware and Architecture, Model Architectures, ML Systems & Hardware

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Hardware and Architecture

In Model Architectures

In ML Systems & Hardware

Citations in the corpus

In-datacenter performance analysis of a tensor processing unit (2017, methodology): “A revised version of Catapult was deployed at larger scale in 2016 [9].”
Chasing Carbon: The Elusive Environmental Footprint of Computing (2020, background): “cation of systems that have billions of transistors and 1,000 higher energy efﬁciency [3]. For salient applications, such as AI [4]–[9], molecular dynamics [10], video encoding [11], and cryptography [12], systems now comprise specialized hardware accelerators that provide orders-of-magnitude higher”
Seer: Leveraging Big Data to Navigate the Complexity of Performance Debugging in Cloud Microservices (2019, methodology): “Apart from software heterogeneity, datacenter hardware is also becoming increasingly heterogeneous as special-purpose architectures [20–22, 39, 55] and FPGAs are used to accelerate critical operations [19, 25, 40, 75].”
An Open-Source Benchmark Suite for Microservices and Their Hardware-Software Implications for Cloud & Edge Systems (2019, background): “This includes analyzing the cycle breakdown in modern servers, examining whether big or small cores are preferable [25, 35, 41, 42, 46– 48], determining the pressure microservices put on instruction caches [37, 52], and exploring the potential they have for hardware acceleration [24, 27, 38, 49, 71]”
The future of computing beyond Moore’s Law (2020, background): “…of diverse acceleration in the AI and ML markets (e.g. Google TPU [7], Nervana’s AI architecture [8], Facebook’s Big Sur [9]) and other forms of compute-in-network acceleration for mega-data centres (e.g. Microsoft’s FPGA Conﬁgurable Cloud and Project Catapult for FPGA-accelerated search [10]).”
Azure Accelerated Networking: SmartNICs in the Public Cloud (2018, methodology): “Our experience on AccelNet as well as other projects within Microsoft, such as Bing Ranking [7, 8] for web search, LZ77 for data compression [20], and BrainWave [17] for machine learning, demonstrate that programming FPGAs is very tractable for production-scale cloud workloads.”
IncBricks: Toward In-Network Computation with an In-Network Cache (2017, methodology): “FPGA accelerators like Catapult [17] could also ﬁt our scenario, but the general purpose processors in the network accelerator are more ﬂexible and easier to program compared with a FPGA.”