HPC Technique Propels Deep Learning at Scale

By Tiffany Trader

February 21, 2017

Researchers from Baidu’s Silicon Valley AI Lab (SVAIL) have adapted a well-known HPC communication technique to boost the speed and scale of their neural network training and now they are sharing their implementation with the larger deep learning community.

The technique, a modified version of the OpenMPI algorithm “ring all-reduce,” is being used at Baidu to parallelize the training of their speech recognition model, Deep Speech 2, across many GPU nodes. The two pieces of software Baidu is announcing today are the baidu-allreduce C library, as well as a patch for TensorFlow, which allows people who have already modeled in TensorFlow to compile this new version and use it for parallelizing across many devices. The codes are available on GitHub.

Ring all-reduce – all GPUs send data simultaneously

Baidu’s SVAIL team developed the approach about two years ago for their internal deep learning framework, named Gene and Majel (in tribute to the famous Star Trek creator and the actress who voiced the onboard computer interfaces for the series). The technique is commonplace in HPC circles, but underused within artificial intelligence and deep learning, according to Baidu.

Many of the researchers in the SVAIL group had come from the high performance computing space and recognized the competitive edge it offered.

“The algorithm is actually part of OpenMPI, but the OpenMPI implementation is not as fast,” comments Baidu Research Scientist Shubho Sengupta. “So the way we stumbled upon it was we started using OpenMPI for doing training and we realized it was not scaling to the extent that we want it to scale. I started digging through the OpenMPI source, found the algorithm, saw that it’s not very efficient, and reimplemented it.”

The SVAIL researchers wrote their own implementation of the ring algorithm for higher performance and better stability. The key distinction from the OpenMPI version is that the SVAIL implementation avoids extraneous copies between the CPU and GPU.

Explains Sengupta, “Once OpenMPI does the communication of these matrices, if the matrices are in GPU memory, it actually copies to CPU memory to do the reduction part of it – that’s actually quite wasteful. You don’t really need to do a copy, you could just write a small kernel that does the reduction in GPU memory space itself. And this especially helps when you are doing all-reduce within a node and all the GPUs are within a PCI root complex, then it doesn’t do any of the copies actually – it can just do everything in GPU memory space. This very simple idea of eliminating this copy resulted in this speedup in scaling over OpenMPI’s own implementation.”

Employing this algorithm along with SVAIL’s focus on fast networking (InfiniBand) and careful hardware-software codesign has enabled the team to get linear GPU scaling up to 128 GPUs, an achievement that was detailed in their December 2015 paper, “Deep Speech 2: End-to-End Speech Recognition in English and Mandarin.”

With their internal implementation of ring all-reduce, the team achieves between a 2.3-21.4X speedup over OpenMPI (version 1.8.5) depending on the number of GPUs.

Sengupta notes that their implementation is fastest for a small number of GPUs. “At 8 GPUs it’s about 20x faster, then as you increase the number of GPUs, it drops because now you actually have to copy data to the CPU to send across the network. But for the internal framework, we can scale all the way up to 128 GPUs and get linear scaling.”

Comparison of two different all-reduce implementations. All times are in seconds. Performance gain is the ratio of OpenMPI all-reduce time to SVAIL’s all-reduce time. (Source: Deep Speech 2 paper)

Sengupta’s teammate Baidu Research Scientist Andrew Gibiansky says similar benefits can now be seen with TensorFlow: “In terms of the TensorFlow implementation, we get the same linear scaling path past eight. In terms of a comparison with running on a single GPU, it ends up being about 31x faster at 40 GPUs.”

After the Deep Speech 2 paper was published, the SVAIL team began getting requests from the community who wanted to know more about the implementation. Given that the algorithm is pretty tightly coupled to SVAIL’s proprietary deep learning framework, they needed to come up with a different way to release it, so they created two new implementations, one specifically for TensorFlow and one that is more general.

Gibiansky, who led the work on the TensorFlow patch, describes their multi-pronged approach to disseminating the information. “You can read the blog post [for a thorough technical explanation] and figure it out. If you’re using TensorFlow, you can use our modification to train your own models with this. And if you’re a deep learning author, you can look at our C library and integrate that. The goal is really to take this idea we’ve found to be really successful internally and try to start spreading it so that other people can also take advantage of it.”

Sengupta shares an interesting perspective on the opportunities to be mined for deep learning within HPC.

“With MPI – people [in deep learning] think that it is this old technology, that it is not relevant, but I think because of our work we have shown that you can build very fast collectives using MPI and that allows you to do synchronous gradient descent which converges faster, gives you deterministic results and you don’t need to do asynchronous gradient descent with parameter servers which was the dominant way of doing this when we first started,” says Sengupta.

As for the reduced-copy approach propagating back to MPI, Gibiansky notes that if you look at some of the other MPI implementations, they’re slowly moving their collectives to GPU versions. “MVPICH recently introduced an all-gather that doesn’t end up copying to CPU – so OpenMPI will probably get there, it just might take a while. Potentially giving this a little more visibility, we can spur that on.”

“There’s a lot of interest now in collectives and one thing we also realized is the all-reduce operation used in traditional HPC setups, it actually transfers data that’s actually not very large,” Sengupta adds. “What it usually does, when I talk to HPC people, it’s trying to figure out the status of something across a bunch of machines – while in deep learning we are transferring these large matrices – like 2048×2048, essentially 4 million 32-bit floating points. For the traditional HPC community, this is a very atypical input for all-reduce. The traditional HPC community does not actually use all-reduce with really large data sizes. I think with deep learning, more and more people are realizing that collective operations for really large matrices is also very important.”

A detailed explanation of ring all-reduce and Baidu’s GPU implementation is covered in this technical blog post, published today by Baidu Research. A variant of the technique is also used to provide high-performance node-local scaling for PaddlePaddle, the company’s open source deep learning framework.

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!

Supercomputer Simulations Validate NASA Crash Testing

February 17, 2020

Car crash simulation is already a challenging supercomputing task, requiring pinpoint estimation of how hundreds of components interact with turbulent forces and human bodies. Spacecraft crash simulation is far more diff Read more…

By Oliver Peckham

What’s New in HPC Research: Quantum Clouds, Interatomic Models, Genetic Algorithms & More

February 14, 2020

In this bimonthly feature, HPCwire highlights newly published research in the high-performance computing community and related domains. From parallel programming to exascale to quantum computing, the details are here. Read more…

By Oliver Peckham

The Massive GPU Cloudburst Experiment Plays a Smaller, More Productive Encore

February 13, 2020

In November, researchers at the San Diego Supercomputer Center (SDSC) and the IceCube Particle Astrophysics Center (WIPAC) set out to break the internet – or at least, pull off the cloud HPC equivalent. As part of thei Read more…

By Oliver Peckham

ORNL Team Develops AI-based Cancer Text Mining Tool on Summit

February 13, 2020

A group of Oak Ridge National Laboratory researchers working on the Summit supercomputer has developed a new neural network tool for fast extraction of information from cancer pathology reports to speed research and clin Read more…

By John Russell

Nature Serves up Another Challenge to Quantum Computing?

February 13, 2020

Just when you thought it was safe to assume quantum computing – though distant – would eventually succumb to clever technology, another potentially confounding factor pops up. It’s the Heisenberg Limit (HL), close Read more…

By John Russell

AWS Solution Channel

Challenging the barriers to High Performance Computing in the Cloud

Cloud computing helps democratize High Performance Computing by placing powerful computational capabilities in the hands of more researchers, engineers, and organizations who may lack access to sufficient on-premises infrastructure. Read more…

IBM Accelerated Insights

Intelligent HPC – Keeping Hard Work at Bay(es)

Since the dawn of time, humans have looked for ways to make their lives easier. Over the centuries human ingenuity has given us inventions such as the wheel and simple machines – which help greatly with tasks that would otherwise be extremely laborious. Read more…

Researchers Enlist Three Supercomputers to Apply Deep Learning to Extreme Weather

February 12, 2020

When it comes to extreme weather, an errant forecast can have serious effects. While advance warning can give people time to prepare for the weather as it did with the polar vortex last year, the absence of accurate adva Read more…

By Oliver Peckham

The Massive GPU Cloudburst Experiment Plays a Smaller, More Productive Encore

February 13, 2020

In November, researchers at the San Diego Supercomputer Center (SDSC) and the IceCube Particle Astrophysics Center (WIPAC) set out to break the internet – or Read more…

By Oliver Peckham

Eni to Retake Industry HPC Crown with Launch of HPC5

February 12, 2020

With the launch of its Dell-built HPC5 system, Italian energy company Eni regains its position atop the industrial supercomputing leaderboard. At 52-petaflops p Read more…

By Tiffany Trader

Trump Budget Proposal Again Slashes Science Spending

February 11, 2020

President Donald Trump’s FY2021 U.S. Budget, submitted to Congress this week, again slashes science spending. It’s a $4.8 trillion statement of priorities, Read more…

By John Russell

Policy: Republicans Eye Bigger Science Budgets; NSF Celebrates 70th, Names Idea Machine Winners

February 5, 2020

It’s a busy week for science policy. Yesterday, the National Science Foundation announced winners of its 2026 Idea Machine contest seeking directions for futu Read more…

By John Russell

Fujitsu A64FX Supercomputer to Be Deployed at Nagoya University This Summer

February 3, 2020

Japanese tech giant Fujitsu announced today that it will supply Nagoya University Information Technology Center with the first commercial supercomputer powered Read more…

By Tiffany Trader

Intel Stopping Nervana Development to Focus on Habana AI Chips

February 3, 2020

Just two months after acquiring Israeli AI chip start-up Habana Labs for $2 billion, Intel is stopping development of its existing Nervana neural network proces Read more…

By John Russell

Lise Supercomputer, Part of HLRN-IV, Begins Operations

January 29, 2020

The second phase of the build-out of HLRN-IV – the planned 16 peak-petaflops supercomputer serving the North-German Supercomputing Alliance (HLRN) – is unde Read more…

By Staff report

IBM Debuts IC922 Power Server for AI Inferencing and Data Management

January 28, 2020

IBM today launched a Power9-based inference server – the IC922 – that features up to six Nvidia T4 GPUs, PCIe Gen 4 and OpenCAPI connectivity, and can accom Read more…

By John Russell

Julia Programming’s Dramatic Rise in HPC and Elsewhere

January 14, 2020

Back in 2012 a paper by four computer scientists including Alan Edelman of MIT introduced Julia, A Fast Dynamic Language for Technical Computing. At the time, t Read more…

By John Russell

Cray, Fujitsu Both Bringing Fujitsu A64FX-based Supercomputers to Market in 2020

November 12, 2019

The number of top-tier HPC systems makers has shrunk due to a steady march of M&A activity, but there is increased diversity and choice of processing compon Read more…

By Tiffany Trader

SC19: IBM Changes Its HPC-AI Game Plan

November 25, 2019

It’s probably fair to say IBM is known for big bets. Summit supercomputer – a big win. Red Hat acquisition – looking like a big win. OpenPOWER and Power processors – jury’s out? At SC19, long-time IBMer Dave Turek sketched out a different kind of bet for Big Blue – a small ball strategy, if you’ll forgive the baseball analogy... Read more…

By John Russell

Intel Debuts New GPU – Ponte Vecchio – and Outlines Aspirations for oneAPI

November 17, 2019

Intel today revealed a few more details about its forthcoming Xe line of GPUs – the top SKU is named Ponte Vecchio and will be used in Aurora, the first plann Read more…

By John Russell

Dell Ramps Up HPC Testing of AMD Rome Processors

October 21, 2019

Dell Technologies is wading deeper into the AMD-based systems market with a growing evaluation program for the latest Epyc (Rome) microprocessors from AMD. In a Read more…

By John Russell

IBM Unveils Latest Achievements in AI Hardware

December 13, 2019

“The increased capabilities of contemporary AI models provide unprecedented recognition accuracy, but often at the expense of larger computational and energet Read more…

By Oliver Peckham

SC19: Welcome to Denver

November 17, 2019

A significant swath of the HPC community has come to Denver for SC19, which began today (Sunday) with a rich technical program. As is customary, the ribbon cutt Read more…

By Tiffany Trader

D-Wave’s Path to 5000 Qubits; Google’s Quantum Supremacy Claim

September 24, 2019

On the heels of IBM’s quantum news last week come two more quantum items. D-Wave Systems today announced the name of its forthcoming 5000-qubit system, Advantage (yes the name choice isn’t serendipity), at its user conference being held this week in Newport, RI. Read more…

By John Russell

Leading Solution Providers

SC 2019 Virtual Booth Video Tour

AMD
AMD
ASROCK RACK
ASROCK RACK
AWS
AWS
CEJN
CJEN
CRAY
CRAY
DDN
DDN
DELL EMC
DELL EMC
IBM
IBM
MELLANOX
MELLANOX
ONE STOP SYSTEMS
ONE STOP SYSTEMS
PANASAS
PANASAS
SIX NINES IT
SIX NINES IT
VERNE GLOBAL
VERNE GLOBAL
WEKAIO
WEKAIO

Jensen Huang’s SC19 – Fast Cars, a Strong Arm, and Aiming for the Cloud(s)

November 20, 2019

We’ve come to expect Nvidia CEO Jensen Huang’s annual SC keynote to contain stunning graphics and lively bravado (with plenty of examples) in support of GPU Read more…

By John Russell

51,000 Cloud GPUs Converge to Power Neutrino Discovery at the South Pole

November 22, 2019

At the dead center of the South Pole, thousands of sensors spanning a cubic kilometer are buried thousands of meters beneath the ice. The sensors are part of Ic Read more…

By Oliver Peckham

Fujitsu A64FX Supercomputer to Be Deployed at Nagoya University This Summer

February 3, 2020

Japanese tech giant Fujitsu announced today that it will supply Nagoya University Information Technology Center with the first commercial supercomputer powered Read more…

By Tiffany Trader

Top500: US Maintains Performance Lead; Arm Tops Green500

November 18, 2019

The 54th Top500, revealed today at SC19, is a familiar list: the U.S. Summit (ORNL) and Sierra (LLNL) machines, offering 148.6 and 94.6 petaflops respectively, Read more…

By Tiffany Trader

Azure Cloud First with AMD Epyc Rome Processors

November 6, 2019

At Ignite 2019 this week, Microsoft's Azure cloud team and AMD announced an expansion of their partnership that began in 2017 when Azure debuted Epyc-backed instances for storage workloads. The fourth-generation Azure D-series and E-series virtual machines previewed at the Rome launch in August are now generally available. Read more…

By Tiffany Trader

Intel’s New Hyderabad Design Center Targets Exascale Era Technologies

December 3, 2019

Intel's Raja Koduri was in India this week to help launch a new 300,000 square foot design and engineering center in Hyderabad, which will focus on advanced com Read more…

By Tiffany Trader

In Memoriam: Steve Tuecke, Globus Co-founder

November 4, 2019

HPCwire is deeply saddened to report that Steve Tuecke, longtime scientist at Argonne National Lab and University of Chicago, has passed away at age 52. Tuecke Read more…

By Tiffany Trader

Cray Debuts ClusterStor E1000 Finishing Remake of Portfolio for ‘Exascale Era’

October 30, 2019

Cray, now owned by HPE, today introduced the ClusterStor E1000 storage platform, which leverages Cray software and mixes hard disk drives (HDD) and flash memory Read more…

By John Russell

  • arrow
  • Click Here for More Headlines
  • arrow
Do NOT follow this link or you will be banned from the site!
Share This