Thanks for the Memory

By George Heidekat

October 20, 2006

Computational scientists and other computer users in the future may hum this tune of gratitude, though perhaps few will appreciate how today's supercomputing helped to realize those even more capable hard drives of the future, as well as better iPods, DVDs, and who knows what?

Technology has made steady advances in the ability to write and store vast quantities of information in smaller, more manageable forms, but current technologies are approaching their limit. How much smaller can you make grains of magnetized material before they start to act like the mind of an overstressed, aging human who steps out of the shower and can't remember if he's washed his hair?

Not much, which is why researchers are seeking a way to magnetize a single nanoparticle reliably, so that such a particle — rather than “grains” used in current magnetic-storage media — can record one bit of digital data. The door to this breakthrough is understanding what happens at the quantum level for each atom in a nanoparticle. Laboratory methods, for all their usefulness, can't provide this information. Over the past year, however, Malcolm Stocks, Yang Wang and their colleagues — using PSC's Cray XT3 — have shown that computational simulations can do the job.

“Experimenters can tell us a lot about the magnetic and electrical properties of a nanoparticle as a whole,” says PSC physicist Wang, “but simulations are the only way we have to determine the magnetic moments and electrical charges of individual atoms and therefore to understand the underlying mechanisms that drive the particle's electro-magnetic behavior.”

Quantum-level simulations of even the smallest nanoparticles present huge challenges. “Consider that even a five nanometer-sized cube of iron contains on the order of 12,000 atoms,” says Oak Ridge National Labs physicist Stocks, who leads a team — including Wang, Aurelian Rusanu, Don Nicholson and Markus Eisenbach of Oak Ridge — who have carried out large-scale nanoparticle simulations with powerful software they developed and the XT3.

Along with proving the feasibility of their approach, they have with their most recent simulation — a 14,400 atom iron-platinum nanoparticle — revealed unexpected electro-magnetic features. They found that a boundary region of nanoparticle atoms isolates the particle interior from quantum disturbances — a finding that opens a new window for design and development of nanostructured magnetic materials.

The Superparamagnetic Limit

Steady increases in data-storage density — which have doubled annually for decades — and the similar trend in faster processors have driven us into the information age. Today's magnetic storage can record data at 100 gigabits per square inch, and the next milestone is a terabit, a trillion bits per square inch. Roughly, that's all the x-ray data in a large hospital, or about 5,000 trees worth of printed paper, on one square-inch of magnetic material.

The dominant storage media are films of magnetic particles, composed of complex alloys of various metals, on fast-spinning hard disks. Whether the information is a rap tune on an iPod or a cryptogram on a CIA server, it's written to the disk by magnetizing minute patches of these particles, one patch for each bit of information. The smaller and fewer the particles that can be given a uniform magnetic field to represent binary “1” or “0,” the greater the density.

The roadblock that looms — called the superparamagnetic limit — is that as the grains become smaller, heat effects (called thermal fluctuations) disturb the magnetic moments, resulting in noise, false sensor response, and long-term data loss. Storing information reliably — 10 years is a typical industry standard — requires stable magnetization at room temperature, but at grain sizes much smaller than current technologies, thermal fluctuations, even at room temperature, tend to scramble the data.

Proposed new designs, such as a patterned magnetic nanoparticle media in which each bit is stored in a single nanoparticle, require better understanding of quantum-mechanical behaviors. “There are big pieces of the underlying physics,” says Stocks, “that we don't understand.”

“We want to be able to predict for each nanoparticle,” says Wang, “how much energy it takes to flip the magnetic moment from up to down. And because thermal fluctuations can flip the moment, we need to know the smallest size a particle can be and resist the fluctuation, and we want to understand the interactions between nanoparticles — how close they can be one to another — because we don't want to flip one particle and affect another one.”

An Electromagnetic Screening Region

To get some of these answers, the Oak Ridge-PSC team turned to LSMS (the locally self-consistent multiple scattering method), powerful software they developed to simulate the quantum properties of solid-state materials. When implemented at PSC in 1998 (on the Cray T3E), LSMS broke the teraflop barrier. It was the first research software (as opposed to benchmark codes) to sustain performance over a teraflop (a trillion computations a second), an achievement that won the 1998 Gordon Bell Prize for high-performance computing.

On the XT3, running on 2,048 processors, LSMS zooms at over eight teraflops, more than 80 percent of the XT3's theoretical peak. “A powerful innovation of LSMS is its linear scaling,” says Stocks. “As the number of atoms increases, computing time increases by only the same multiple.” Without the innovations of LSMS, a realistically sized particle of tens-of-thousands of atoms would be too large to simulate. With a conventional approach, the amount of computing increases by the number of atoms cubed (N^3) — so that as the particle size changes, for instance, from 1,000 to 10,000 atoms, the computing would take 1,000 (10^3) times as long, rather than, as with LSMS, only ten times.

“Quantum-mechanical calculations are very time-consuming,” says Stocks, “even on big computers. What we've done with LSMS is develop methods that allow you to do calculations for a sufficient number of atoms so the nanoparticle that the theory people talk about gets very close to the size that the experimentalists can make and measure.”

In 2005, Stocks, Wang and colleagues used LSMS to carry out one of the first projects on PSC's then newly installed XT3. They calculated the electronic and magnetic structure of an iron nanoparticle embedded in an iron-aluminide matrix, a total of 16,000 atoms. This was the first quantum-based calculation of a physical system several nanometers in scale, and it successfully laid the groundwork for their next simulation.

The 2005 calculations were limited by an assumption that the magnetic moments for each atom of the nanoparticle were aligned in the same direction. “In reality,” says Yang, “because of surrounding materials, that won't be the case.” This year, taking the next step, they simulated an iron-platinum nanoparticle — at three different sizes, 2.5, 3.86 and 5.0 nanometers — in a matrix of surrounding alloy, totaling in each case 14,400 atoms, without any a priori assumptions about alignment of the magnetic moments. They focused on iron-platinum, says Wang, because industry leaders are interested in its potential as a storage medium. The computations occupied 1,200 XT3 processors — 12 atoms assigned to each processor — running for more than 50 hours.

“Iron-platinum,” says Stocks, “is an important material because of its high magnetic anisotropy compared to most metals. This means it takes more energy for the magnetic moment to flip. As particle size gets smaller, this is more and more important.”

The result of the LSMS simulations is a detailed picture of a realistic nanoparticle's electronic and magnetic structure. Atoms in the particle interior settle into an organized pattern of essentially the same amount of net charge and magnetic moment as bulk crystalline iron-platinum. Iron atoms gain electrons from the platinum atoms to become negatively charged, and conversely the platinum atoms become positively charged.

The most interesting finding is a boundary region, just a few atoms in thickness, that isolates the nanoparticle from disruptive quantum effects from the surrounding alloy. In this region, about four angstroms, net charge and magnetic moment fluctuate from influence of the alloy. The four-angstrom width of this region was constant for all three simulations, suggesting that it is independent of particle size. “The perturbed surface of the particle,” says Stocks, “is like a fishbowl, separating the interior from outside effects. This is a piece of knowledge we didn't have before, and it will prove useful.”

This success points to more challenging tasks ahead. Nanotechnology — such as new discoveries in nanostructured “spintronic” devices, which may be able to harness the electron's spin as well as its charge — represent problems well beyond even the latest LSMS simulation. Meeting those challenges will require characterizing electronic and magnetic structures on scales approaching millions of atoms, rather tens-of-thousands, and the computational cost will rise by a factor of hundreds.

Eventually, simulations such as these will aid in developing the more powerful computers needed for the next round of simulations. “Without the magnetic storage advances to date,” says Stocks, “we wouldn't be able to do these calculations, which will help develop the next level of materials for the disk drives we'll need for the next level of calculations.”

“Ultimately,” says Wang, “the success of direct quantum mechanical simulation of nanostructured electronic devices relies on the advent of petaflop computing technology.”

For more information, including graphics, visit http://www.psc.edu/science/2006/nano.html.

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!

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 power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device 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 the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

Nvidia Appoints Andy Grant as EMEA Director of Supercomputing, Higher Education, and AI

March 22, 2024

Nvidia recently appointed Andy Grant as Director, Supercomputing, Higher Education, and AI for Europe, the Middle East, and Africa (EMEA). With over 25 years of high-performance computing (HPC) experience, Grant brings a 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…

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…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Houston We Have a Solution: Addressing the HPC and Tech Talent Gap

March 15, 2024

Generations of Houstonian teachers, counselors, and parents have either worked in the aerospace industry or know people who do - the prospect of entering the fi Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. 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…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. 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…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

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 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire