HTC, Big Data and the God Particle

By Miha Ahronovitz

March 29, 2013

Who has seen the wind?

Neither I nor you:

But when the leaves hang trembling,

The wind is passing through.

— Christina Rossetti 1830–1894

The ancient sages say that a prophecy is conveyed in vision, in appearance, in sight, in revelation and in word. I often compare older predictions to what actually happens after the facts are known. For example I looked at IDC HPC predictions for 2010 in 2013 and I found them to be surprisingly accurate. HPCwire does not make prophecies, yet with astonishing intuition it captured all the significant events in performance computing for the last three decades.

The concept of High Throughput Computing (HTC) appeared on June 27, 1997, edition of HPCwire, during an interview with Miron Livny – a professor at University of Wisconsin, Madison – with Alan Beck, the HPCwire editor in chief at the time.

This month, NCSA’s (National Center for Supercomputing Applications) Advanced Computing Group (ACG) will begin testing Condor, a software system developed at the University of Wisconsin that promises to expand computing capabilities through efficient capture of cycles on idle machines. The software, operating within an HTC (High Throughput Computing) rather than a traditional HPC (High Performance Computing) paradigm, organizes machines into clusters, called pools, or collections of clusters called flocks, that can exchange resources. Condor then hunts for idle workstations to run jobs. When the owner resumes computing, Condor migrates the job to another machine.

In response to a question from Alan about the work they are doing at NCSA, Miron described the 1997 HTC plans:

“We play a dual role with respect to NCSA. On one hand, we’re a regional partner, with over 500 workstations here on campus. These will provide a source of cycles to NCSA and a testbed for scientists who would like to see how well their applications work in an HTC environment. On the other hand, we’re also an enabling technology, where our experience in building and maintaining Condor will contribute to the construction of the National Technology Grid. Thus, we hope we can soon move from a campus-wide to a nation-wide HTC system.”

By far the most enthusiastic response for HTC came from the high energy physics community. At that time, Fermilab in Batavia, 20 miles east of Chicago was the leading high energy physics facility. The Large Hadron Collider (LHC) at CERN was built from 1988 to 2008 and started operations in 2009. Before LHC, Fermilab’s Tevatron circular particle collider was the biggest in the world. For about thirty years, it was the number one particle collider serving high energy physics where the need for big data was prevalent long before this term became a media buzzword. On September 30, 2011, the Tevatron was closed during an emotional ceremony:

The Tevatron was the highest energy particle collider in the world until 2009, and is credited with discovering the top quark, two types of sigma baryon, the “Cascade B” Xi Baryon, and the “doubly strange” Omega-sub-b particle, as well as embarking on the hunt for the elusive “God particle,” throughout its storied history.

The Large Hadron Collider (LHC), which so far has cost nine billion dollars, just shut down for the next two years starting February 2013. This is called the long shut down (LS1) When it will reopen in 2015, it will start colliding protons at an unprecedented 14 TeV, (teraelectronvolt)

Scientists won’t be idle during the tunnel’s shutdown: CERN’s mass-storage systems are hanging onto 100 quadrillion bytes of data to analyze, most of which was acquired over the past year. This is equivalent to 700 years of high definition movies.

CERN Research Director Sergio Bertolucci says: “Other experiments here have ongoing analyses, so I’m looking forward to many interesting results emerging as LS1 progresses.”

OSG was created in response to the vast experience accumulated at Fermi Lab answering at the time the needs of high energy physics researchers. Since 2004, OSG and its underlying HTC technology proved itself “magna cum laude.” It directly contributed to the creation of the world’s largest computational grid at CERN, powerful enough to make the search for Higgs successful.

According to Lothar Bauerdick – OSG Executive Director – the name “Open Science Grid” was coined by Ruth Pordes during a walk in the vineyards around CERN in 2003. She is an associate head computing at Fermilab. An Oxford trained physicist – no matter what official title she holds – Ruth is recognized as the soul of OSG. Her 2004 slides from Condor Week sketched the first portrait of OSG to the outside world.

Another key contributor is Frank Wuerthwein, a star physics researcher from UC San Diego. He is “developing, deploying, and now operating a worldwide distributed computing system for high throughput computing with large data volumes expect to grow to Exabytes by 2020.” He is one of the most active evangelists for HTC and OSG and has the credibility because he himself is involved in the day-to-day search for dark matter.

The Open Science Grid (OSG) has access to over 60,000 processors without owning any one of them. The OSG is the real life National Science Grid, predicted in the HPCwire article sixteen years ago.

Next >>

In an interview with ATKearney – the management consulting group – Fermilab’s head of the Scientific Computing Division and a leading physicist searching for Higgs in America – Rob Roser said: “2012 has been the most exciting year in physics in the last 50 years with the confirmation of the God particle, but the computing advancements contributed significantly to this excitement.”

Roser says that big data is hard:

People who are successful at big data are those who are not overwhelmed by it. They are managing the data – the data’s not managing them.

The Higgs particle discovery is clearly explained in the abstract of this CERN paper from December 2012 signed by 2,932 Atlas collaborators:

Nearly 50 years ago, theoretical physicists proposed that a field permeates the universe and gives energy to the vacuum. This field was required to explain why some, but not all, fundamental particles have mass. The ATLAS experiment at the Large Hadron Collider at CERN has now observed the production of a new particle with a mass of 126 giga-electron volts and decay signatures consistent with those expected for the Higgs particle.

We are used to judge organizations in terms of ROI – return on investment. I would argue that the (ROI) of Higgs boson discovery is infinite. The relativity theory or Maxwell’s electromagnetic laws – just to mention randomly a few scientific breakthroughs – have infinite ROI. They generate wealth for many generations to come and for every citizen of the planet.

Figure 1 Open Science Grid real time usage http://display.grid.iu.edu/

Higgs particle research at the Large Hadron Collider generated in 2012, twenty five millions of gigabytes (23.8 petabytes which is approximated to 25 petabytes in popular media). The forecast increase is 20 percent annually.

Figure 2

“Nothing shocks me. I’m a scientist,” said Harrison Ford acting as Indiana Jones in Steven Spielberg’s 1981 movie Raiders of the Lost Ark.

Like Indiana Jones, CERN’s computer scientists tackled the challenge. They built an unprecedented massive grid, using the same HTC technology described in the 1997 article from HPCwire. The Worldwide LHC Computing Grid (WLCG) “is the world’s largest computing grid. It is based on two main grids – the European Grid Infrastructure in Europe, and Open Science Grid in the US – but has many associated regional and national grids (such as TWGrid in Taiwan and EU-IndiaGrid, which supports grid infrastructures across Europe and Asia).” CERN explains why this solution was selected:

This grid-based infrastructure is the most effective solution to the data-analysis challenge of the LHC, offering many advantages over a centralized system. Multiple copies of data can be kept at different sites, ensuring access for all scientists independent of geographical location; there is no single point of failure; computer centers in multiple time zones ease round-the-clock monitoring and the availability of expert support; and resources can be distributed across the world, for funding and sociological reasons.

It is about time to give HTC the credit. Without this technology, the Higgs boson particle would have been relegated to an elegant theory. Sooner or later, technological advances in computer science ought to be successful to prove the God particle exists. It so happened that the first breakthrough technology to get there is HTC.

Great physicists are like great prophets. Professor Higgs and his colleagues are among them. They detect “the unrevealed,” those whispers requiring great effort to comprehend. WLCG delivered the proof of their thought experiments, while they are still alive. Never mind Nobel prize speculations, this by itself is a huge reward. High Throughput Computing (HTC) and High Performance Computing (HPC) complement each other. This is a comment of Damien Hocking, the CTO of Corelium Software:

The logical progression of HTC is to keep improving until it can effectively process HPC workloads, and that’s where portability will come from. CPUs evolved as HTC devices and current software design presents HTC-like workloads to the OS (which is itself HTC software by design), but on a much shorter timescale. HPC workloads run today because the HTC architecture underneath is so efficient.

I asked Damien to elaborate:

My workstation has eight cores, and right now is running 110 processes and over 1,500 threads and is still responding quickly, even though I’m editing code, running a simulation, committing to Github and typing this, all at the same time. I’d argue that’s HTC, and that HPC today is realized by building and running it on top of the HTC platform that modern hardware and OS-es provide.

The Center for High Throughput Computing at University of Wisconsin is continuously adding new nodes. The latest addition is an HPC-capable node with 768 cores cluster with InfiniBand, using a SLURM scheduler and with HTCondor backfilling unused HPC cores with HTC jobs.

Many ask: “How do you determine if an application is suitable for HTC or HPC?” For the technical minded, there is a detailed presentation by Zach Miller from the University of Wisconsin. If you really are serious to start, there is an 2013 OSG User School. The website says:

“… you will learn to use high-throughput computing (HTC) systems – at your own campus or using the national Open Science Grid (OSG) – to run large-scale computing applications that are at the heart of today’s cutting-edge science … you will also go to the XSEDE13 Conference, July 22–25, in San Diego. This is a premier event for the high-performance computing (HPC) community, offering students a view into another approach to large-scale computing.”

CHTC and OSG leaderships executed the plan consistently and with remarkable skill. The 1997 dream of a 500 desktop grid transformed into the reality of WLCG grid designed to meet CERN’s unique needs. WLCG has in excess of 150,000 cores and growing.

Quoting the science fiction writer Arthur C. Clarke:

Every revolutionary idea seems to evoke three stages of reaction. They may be summed up by the phrases: (1) It’s completely impossible. (2) It’s possible, but it’s not worth doing. (3) I said it was a good idea all along.

And…

Any sufficiently advanced technology is indistinguishable from magic.

Will HTC technology for big data generate a new business like FaceBook or Google? Rob Roser says yes. The World Wide Web was invented by physicists at CERN, but other people took over. “I am perfectly happy to let corporate America and others take over,” he says, “and develop technology that we can use.”

Subscribe to HPCwire's Weekly Update!

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

ARM, Fujitsu Targeting Open-source Software for Power Efficiency in 2-nm Chip

July 19, 2024

Fujitsu and ARM are relying on open-source software to bring power efficiency to an air-cooled supercomputing chip that will ship in 2027. Monaka chip, which will be made using the 2-nanometer process, is based on the Read more…

SCALEing the CUDA Castle

July 18, 2024

In a previous article, HPCwire has reported on a way in which AMD can get across the CUDA moat that protects the Nvidia CUDA castle (at least for PyTorch AI projects.). Other tools have joined the CUDA castle siege. AMD Read more…

Quantum Watchers – Terrific Interview with Caltech’s John Preskill by CERN

July 17, 2024

In case you missed it, there's a fascinating interview with John Preskill, the prominent Caltech physicist and pioneering quantum computing researcher that was recently posted by CERN’s department of experimental physi Read more…

Aurora AI-Driven Atmosphere Model is 5,000x Faster Than Traditional Systems

July 16, 2024

While the onset of human-driven climate change brings with it many horrors, the increase in the frequency and strength of storms poses an enormous threat to communities across the globe. As climate change is warming ocea Read more…

Researchers Say Memory Bandwidth and NVLink Speeds in Hopper Not So Simple

July 15, 2024

Researchers measured the real-world bandwidth of Nvidia's Grace Hopper superchip, with the chip-to-chip interconnect results falling well short of theoretical claims. A paper published on July 10 by researchers in the U. Read more…

Belt-Tightening in Store for Most Federal FY25 Science Budets

July 15, 2024

If it’s summer, it’s federal budgeting time, not to mention an election year as well. There’s an excellent summary of the curent state of FY25 efforts reported in AIP’s policy FYI: Science Policy News. Belt-tight Read more…

SCALEing the CUDA Castle

July 18, 2024

In a previous article, HPCwire has reported on a way in which AMD can get across the CUDA moat that protects the Nvidia CUDA castle (at least for PyTorch AI pro Read more…

Aurora AI-Driven Atmosphere Model is 5,000x Faster Than Traditional Systems

July 16, 2024

While the onset of human-driven climate change brings with it many horrors, the increase in the frequency and strength of storms poses an enormous threat to com Read more…

Shutterstock 1886124835

Researchers Say Memory Bandwidth and NVLink Speeds in Hopper Not So Simple

July 15, 2024

Researchers measured the real-world bandwidth of Nvidia's Grace Hopper superchip, with the chip-to-chip interconnect results falling well short of theoretical c Read more…

Shutterstock 2203611339

NSF Issues Next Solicitation and More Detail on National Quantum Virtual Laboratory

July 10, 2024

After percolating for roughly a year, NSF has issued the next solicitation for the National Quantum Virtual Lab program — this one focused on design and imple Read more…

NCSA’s SEAS Team Keeps APACE of AlphaFold2

July 9, 2024

High-performance computing (HPC) can often be challenging for researchers to use because it requires expertise in working with large datasets, scaling the softw Read more…

Anders Jensen on Europe’s Plan for AI-optimized Supercomputers, Welcoming the UK, and More

July 8, 2024

The recent ISC24 conference in Hamburg showcased LUMI and other leadership-class supercomputers co-funded by the EuroHPC Joint Undertaking (JU), including three Read more…

Generative AI to Account for 1.5% of World’s Power Consumption by 2029

July 8, 2024

Generative AI will take on a larger chunk of the world's power consumption to keep up with the hefty hardware requirements to run applications. "AI chips repres Read more…

US Senators Propose $32 Billion in Annual AI Spending, but Critics Remain Unconvinced

July 5, 2024

Senate leader, Chuck Schumer, and three colleagues want the US government to spend at least $32 billion annually by 2026 for non-defense related AI systems.  T Read more…

Atos Outlines Plans to Get Acquired, and a Path Forward

May 21, 2024

Atos – via its subsidiary Eviden – is the second major supercomputer maker outside of HPE, while others have largely dropped out. The lack of integrators and Atos' financial turmoil have the HPC market worried. If Atos goes under, HPE will be the only major option for building large-scale systems. Read more…

Everyone Except Nvidia Forms Ultra Accelerator Link (UALink) Consortium

May 30, 2024

Consider the GPU. An island of SIMD greatness that makes light work of matrix math. Originally designed to rapidly paint dots on a computer monitor, it was then Read more…

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock_1687123447

Nvidia Economics: Make $5-$7 for Every $1 Spent on GPUs

June 30, 2024

Nvidia is saying that companies could make $5 to $7 for every $1 invested in GPUs over a four-year period. Customers are investing billions in new Nvidia hardwa Read more…

Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study

June 10, 2024

Nvidia had an explosive 2023 in data-center GPU shipments, which totaled roughly 3.76 million units, according to a study conducted by semiconductor analyst fir Read more…

AMD Clears Up Messy GPU Roadmap, Upgrades Chips Annually

June 3, 2024

In the world of AI, there's a desperate search for an alternative to Nvidia's GPUs, and AMD is stepping up to the plate. AMD detailed its updated GPU roadmap, w Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

Intel’s Next-gen Falcon Shores Coming Out in Late 2025 

April 30, 2024

It's a long wait for customers hanging on for Intel's next-generation GPU, Falcon Shores, which will be released in late 2025.  "Then we have a rich, a very Read more…

Leading Solution Providers

Contributors

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's l Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

IonQ Plots Path to Commercial (Quantum) Advantage

July 2, 2024

IonQ, the trapped ion quantum computing specialist, delivered a progress report last week firming up 2024/25 product goals and reviewing its technology roadmap. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

The NASA Black Hole Plunge

May 7, 2024

We have all thought about it. No one has done it, but now, thanks to HPC, we see what it looks like. Hold on to your feet because NASA has released videos of wh Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire