The Role and Potential of CPUs in Deep Learning

By Sparsh Mittal

April 14, 2021

In this invited guest piece, Sparsh Mittal provides perspective on the role of the Central Processing Unit (CPU) for deep learning workloads in an increasingly diverse processor space, reviewing use cases where the performance of the CPU excels and noting some of the architectural changes and directions spurred by deep learning applications. The article serves as an introduction to a new survey research paper (written by Mittal et al) published this April in IEEE Transactions on Neural Networks and Learning Systems.

Deep learning (DL) applications have unique architectural characteristics and efficiency requirements. Hence, the choice of computing system has a profound impact on how large a piece of the DL pie a user can finally enjoy. Even though accelerators may provide higher throughput than general-purpose computing systems (CPUs), there are several other metrics and usage scenarios on which CPUs are preferred or are superior. A recent survey paper I’ve coauthored with Poonam Rajput and Sreenivas Subramoney (A Survey of Deep Learning on CPUs: Opportunities and Co-optimizations) highlights the strengths of CPUs in DL, and identifies opportunities for further optimization.

CPU has its forte, and accelerator is not a panacea

Sparse DNNs are inefficient on massively parallel processors because of their irregular memory accesses and inability to leverage optimizations such as cache tiling and vectorization. Further, RNNs are difficult to parallelize due to the dependencies between the steps. Similarly, DNNs such as InceptionNet variants have filter shapes of 1×1, 3×3, 1×3, 3×1, etc., which lead to irregular memory accesses and variable-amount of parallelism across the layers. CPUs are more suitable for such applications with limited parallelism because of their advanced memory management techniques. For example, researchers from Rice University have shown that for fully connected networks over sparse datasets such as Amazon-670K, and Delicious-200K, the DL training problem can be modeled as a search problem. This allows replacing matrix multiplication operation with hash tables. Their technique on CPU provides higher performance than a TensorFlow-based implementation on GPU.

3D CNNs and even 2D CNNs with large batch-size require a massive amount of memory. Since CPU-managed hosts in the cloud and datacenter scenarios have much larger memory capacities than accelerators, running memory-hungry operations on CPUs is not merely attractive but often imperative. Accelerators such as TPU provide high throughput for large batch size; however, for applications requiring real-time inference, the use of large batch size is not preferred. At small batch sizes, CPUs generally provide competitive latency. There are a host of techniques that can be applied to further tune the DL applications on CPUs, for example, hardware-aware pruning, vectorization, cache tiling and approximate computing. Our survey paper summarizes many such techniques.

Across the board: From tiny wearables to large datacenters

IoT devices and wearables have tight power and area budgets, which precludes over-specialization. For example, a smartwatch chip cannot host separate accelerators for speech/audio/image/video processing. In smartphones running Android, the programming support for mobile GPU or DSP is not fully mature. In fact, on a typical mobile SoC, the theoretical peak performance of mobile CPUs equals that of mobile GPUs. Further, datacenters supporting web services such as social networks see a significant fluctuation in computing demand over time. CPUs can meet this variability in demand due to their high availability and efficiency for both DL and non-DL tasks. Finally, in extreme environments such as defense and medical, which require security certifications, CPUs are sometimes the only platform of choice.

Not missing the obvious: economy and ease of use

Accelerators require long design cycles and massive investment. Integrating them into existing ecosystems requires high costs and engineering work. By contrast, the hardware/software stack of CPUs is already well-established and understood. They can provide reasonable speedups on a broad range of applications. While large-scale companies have the resources to build and maintain their custom accelerators, CPUs (or GPUs) remain the most feasible platform for other companies.

Future outlook: brighter than you think

Going forward, merely increasing peak performance will not be sufficient; more revolutionary improvements are required to boost the performance of a broad range of DL applications, such as reinforcement learning and generative adversarial networks. Recent CPUs have begun to provide hardware support for low-precision computing. Once in-memory computing reaches maturity, the large caches of CPUs would turn into massive compute units. Development of open-source ISA such as RISC-V would further break the portability and proprietary barriers of accelerators.

The metrics of interest are numerous and varied, and so are the state-of-the-art DL models. We believe that instead of a “general-purpose processor versus accelerator” debate, the future will see a CPU-accelerator heterogeneous computing approach that brings together the best of both worlds.

About the Author

Dr. Sparsh Mittal is currently working as an assistant professor at IIT Roorkee, India. He received the B.Tech. degree from IIT, Roorkee, India and the Ph.D. degree from Iowa State University (ISU), USA. He has worked as a Post-Doctoral Research Associate at Oak Ridge National Lab (ORNL), USA and as an assistant professor at CSE, IIT Hyderabad. He was the graduating topper of his batch in B.Tech and his BTech project received the best project award. He has received a fellowship from ISU and a performance award from ORNL. He has published more than 90 papers at top venues. He is an associate editor of Elsevier’s Journal of Systems Architecture. He has given invited talks at ISC Conference at Germany, New York University, University of Michigan and Xilinx (Hyderabad).

 

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industy updates delivered to you every week!

IBM Research Debuts 2nm Test Chip with 50 Billion Transistors

May 6, 2021

IBM Research today announced the successful prototyping of the world's first 2 nanometer chip, fabricated with silicon nanosheet technology on a standard 300mm bulk wafer. With ~50 billion transistors, the chip will enab Read more…

Supercomputer-Powered CRISPR Simulation Lights Path to Better DNA Editing

May 5, 2021

CRISPR-Cas9 – mostly just known as CRISPR – is a powerful genome editing tool that uses an enzyme (Cas9) to slice off sections of DNA and a guide RNA to repair and modify the DNA as desired, opening the door for cure Read more…

LRZ Announces New Phase of SuperMUC-NG Supercomputer with Intel’s ‘Ponte Vecchio’ GPU

May 5, 2021

At the Leibniz Supercomputing Centre (LRZ) in München, Germany – one of the constituent centers of the Gauss Centre for Supercomputing (GCS) – the SuperMUC-NG system has stood tall for several years, placing 15th on Read more…

HPC Simulations Show How Antibodies Quash SARS-CoV-2

May 5, 2021

Following more than a year of rapid-fire research and pharmaceutical development, nearly a billion COVID-19 vaccine doses have been administered around the world, with many of those vaccines proving remarkably effective Read more…

Crystal Ball Gazing at Nvidia: R&D Chief Bill Dally Talks Targets and Approach

May 4, 2021

There’s no quibbling with Nvidia’s success. Entrenched atop the GPU market, Nvidia has ridden its own inventiveness and growing demand for accelerated computing to meet the needs of HPC and AI. Recently it embarked on an ambitious expansion by acquiring Mellanox (interconnect)... Read more…

AWS Solution Channel

FLYING WHALES runs CFD workloads 15 times faster on AWS

FLYING WHALES is a French startup that is developing a 60-ton payload cargo airship for the heavy lift and outsize cargo market. The project was born out of France’s ambition to provide efficient, environmentally friendly transportation for collecting wood in remote areas. Read more…

2021 Winter Classic – Coaches Chat

May 4, 2021

The Winter Classic Invitational Student Cluster Competition raged for all last week and now we’re into the week of judging interviews. Time has been flying. So as we wait for results, let’s dive a bit deeper into t Read more…

IBM Research Debuts 2nm Test Chip with 50 Billion Transistors

May 6, 2021

IBM Research today announced the successful prototyping of the world's first 2 nanometer chip, fabricated with silicon nanosheet technology on a standard 300mm Read more…

LRZ Announces New Phase of SuperMUC-NG Supercomputer with Intel’s ‘Ponte Vecchio’ GPU

May 5, 2021

At the Leibniz Supercomputing Centre (LRZ) in München, Germany – one of the constituent centers of the Gauss Centre for Supercomputing (GCS) – the SuperMUC Read more…

Crystal Ball Gazing at Nvidia: R&D Chief Bill Dally Talks Targets and Approach

May 4, 2021

There’s no quibbling with Nvidia’s success. Entrenched atop the GPU market, Nvidia has ridden its own inventiveness and growing demand for accelerated computing to meet the needs of HPC and AI. Recently it embarked on an ambitious expansion by acquiring Mellanox (interconnect)... Read more…

Intel Invests $3.5 Billion in New Mexico Fab to Focus on Foveros Packaging Technology

May 3, 2021

Intel announced it is investing $3.5 billion in its Rio Rancho, New Mexico, facility to support its advanced 3D manufacturing and packaging technology, Foveros. Read more…

Supercomputer Research Shows Standard Model May Withstand Muon Discrepancy

May 3, 2021

Big news recently struck the physics world: researchers at the Fermi National Accelerator Laboratory (FNAL), in the midst of their Muon g-2 experiment, publishe Read more…

NWChemEx: Computational Chemistry Code for the Exascale Era

April 29, 2021

A team working on biofuel research is rewriting the decades-old NWChem software program for the exascale era. The new software, NWChemEx, will enable computatio Read more…

HPE Will Build Singapore’s New National Supercomputer

April 28, 2021

More than two years ago, Singapore’s National Supercomputing Centre (NSCC) announced a $200 million SGD (~$151 million USD) investment to boost its supercomputing power by an order of magnitude. Today, those plans come closer to fruition with the announcement that Hewlett Packard Enterprise (HPE) has been awarded... Read more…

Arm Details Neoverse V1, N2 Platforms with New Mesh Interconnect, Advances Partner Ecosystem

April 27, 2021

Chip designer Arm Holdings is sharing details about its Neoverse V1 and N2 cores, introducing its new CMN-700 interconnect, and showcasing its partners' plans t Read more…

Julia Update: Adoption Keeps Climbing; Is It a Python Challenger?

January 13, 2021

The rapid adoption of Julia, the open source, high level programing language with roots at MIT, shows no sign of slowing according to data from Julialang.org. I Read more…

Intel Launches 10nm ‘Ice Lake’ Datacenter CPU with Up to 40 Cores

April 6, 2021

The wait is over. Today Intel officially launched its 10nm datacenter CPU, the third-generation Intel Xeon Scalable processor, codenamed Ice Lake. With up to 40 Read more…

CERN Is Betting Big on Exascale

April 1, 2021

The European Organization for Nuclear Research (CERN) involves 23 countries, 15,000 researchers, billions of dollars a year, and the biggest machine in the worl Read more…

HPE Launches Storage Line Loaded with IBM’s Spectrum Scale File System

April 6, 2021

HPE today launched a new family of storage solutions bundled with IBM’s Spectrum Scale Erasure Code Edition parallel file system (description below) and featu Read more…

10nm, 7nm, 5nm…. Should the Chip Nanometer Metric Be Replaced?

June 1, 2020

The biggest cool factor in server chips is the nanometer. AMD beating Intel to a CPU built on a 7nm process node* – with 5nm and 3nm on the way – has been i Read more…

Saudi Aramco Unveils Dammam 7, Its New Top Ten Supercomputer

January 21, 2021

By revenue, oil and gas giant Saudi Aramco is one of the largest companies in the world, and it has historically employed commensurate amounts of supercomputing Read more…

Quantum Computer Start-up IonQ Plans IPO via SPAC

March 8, 2021

IonQ, a Maryland-based quantum computing start-up working with ion trap technology, plans to go public via a Special Purpose Acquisition Company (SPAC) merger a Read more…

Can Deep Learning Replace Numerical Weather Prediction?

March 3, 2021

Numerical weather prediction (NWP) is a mainstay of supercomputing. Some of the first applications of the first supercomputers dealt with climate modeling, and Read more…

Leading Solution Providers

Contributors

Livermore’s El Capitan Supercomputer to Debut HPE ‘Rabbit’ Near Node Local Storage

February 18, 2021

A near node local storage innovation called Rabbit factored heavily into Lawrence Livermore National Laboratory’s decision to select Cray’s proposal for its CORAL-2 machine, the lab’s first exascale-class supercomputer, El Capitan. Details of this new storage technology were revealed... Read more…

AMD Launches Epyc ‘Milan’ with 19 SKUs for HPC, Enterprise and Hyperscale

March 15, 2021

At a virtual launch event held today (Monday), AMD revealed its third-generation Epyc “Milan” CPU lineup: a set of 19 SKUs -- including the flagship 64-core, 280-watt 7763 part --  aimed at HPC, enterprise and cloud workloads. Notably, the third-gen Epyc Milan chips achieve 19 percent... Read more…

Programming the Soon-to-Be World’s Fastest Supercomputer, Frontier

January 5, 2021

What’s it like designing an app for the world’s fastest supercomputer, set to come online in the United States in 2021? The University of Delaware’s Sunita Chandrasekaran is leading an elite international team in just that task. Chandrasekaran, assistant professor of computer and information sciences, recently was named... Read more…

New Deep Learning Algorithm Solves Rubik’s Cube

July 25, 2018

Solving (and attempting to solve) Rubik’s Cube has delighted millions of puzzle lovers since 1974 when the cube was invented by Hungarian sculptor and archite Read more…

African Supercomputing Center Inaugurates ‘Toubkal,’ Most Powerful Supercomputer on the Continent

February 25, 2021

Historically, Africa hasn’t exactly been synonymous with supercomputing. There are only a handful of supercomputers on the continent, with few ranking on the Read more…

GTC21: Nvidia Launches cuQuantum; Dips a Toe in Quantum Computing

April 13, 2021

Yesterday Nvidia officially dipped a toe into quantum computing with the launch of cuQuantum SDK, a development platform for simulating quantum circuits on GPU-accelerated systems. As Nvidia CEO Jensen Huang emphasized in his keynote, Nvidia doesn’t plan to build... Read more…

The History of Supercomputing vs. COVID-19

March 9, 2021

The COVID-19 pandemic poses a greater challenge to the high-performance computing community than any before. HPCwire's coverage of the supercomputing response t Read more…

HPE Names Justin Hotard New HPC Chief as Pete Ungaro Departs

March 2, 2021

HPE CEO Antonio Neri announced today (March 2, 2021) the appointment of Justin Hotard as general manager of HPC, mission critical solutions and labs, effective Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire