SUPERCOMPUTERS TRACK HUMAN GENOME

September 1, 2000

FEATURES & COMMENTARY

Pasadena, CA — Andrew Pollack reports that Kwang-I Yu, president of Paracel Inc., will not say which secretive government agency buys his company’s specialized supercomputers. “We sell to the federal government,” he demurs.

But J. Craig Venter, the president of Celera Genomics, is less circumspect. Paracel’s machines, he said, are used by the National Security Agency, the code-breaking unit that eavesdrops on other nations. And Dr. Venter should know. Celera Genomics, the company that mapped the human genome, acquired Paracel for nearly $250 million in stock in June.

Paracel’s machines are designed to sift documents rapidly for particular words, phrases or strings of letters – an ability of obvious use to those filtering communications for important nuggets.

But what attracted Celera is that Paracel’s technology is also good for breaking another code – the genetic one. The machines are being snapped up to sift the blizzard of data being generated by the Human Genome Project and various private genomics efforts. “They’re all character strings,” said Dr. Yu, comparing the gene sequence to a text message.

The need for Paracel’s machines stems from biology’s shift from a “wet” science performed in test tubes to at least partly a “dry” one, in which much of the crucial analysis is done on computers. This has given rise to a field called bioinformatics, the use of computer science in life science.

And bioinformatics must handle volumes of data so huge they could bring ordinary computers to their knees. The human genome is comprised of three billion chemical units represented by the letters A, C, T, and G – a string that would stretch from Boston to London if written in letters of the size in this article. A scientist studying a particular sequence of DNA might search through the entire human genome, as well as those of other animals and bacteria, to find similar sequences. And genomics companies might want to do thousands of such searches a day.

“There’s a class of algorithms that would take forever on a regular computer,” said Edward Kiruluta, vice president for research and development at DoubleTwist, a bioinformatics company in Oakland, Calif. Even with its Paracel machine, he said, some analyses take weeks. On regular computers, they would take months. And speed is of the essence, especially in analyzing data from the publicly financed Human Genome Project, which makes new sequences available each day to everyone at the same time.

“As the genome data comes out, you want to analyze it as fast as you can, make the discoveries first and protect the intellectual property,” said Martin D. Leach, director for bioinformatics at CuraGen, which uses genomics to develop drugs.

Computers, of course, are doubling in speed for a given cost every 18 months or so, according to Moore’s Law, named after Intel’s co-founder, Gordon Moore. But the volume of genomic data is growing even faster, owing to automated DNA-sequencing machines. GenBank, the government-run public domain database of DNA sequences, more than doubled in size the last six months to more than 8.6 billion chemical units.

I.B.M. estimates that the market for hardware and software for life sciences will grow from $3.5 billion now to more than $9 billion by 2002. Carolyn Kovac, who heads a newly formed life sciences division at I.B.M., said biologists had replaced physicists as the main scientific users of supercomputers.

Sun Microsystems and Compaq Computer are also trying to develop products for the life sciences, with either supercomputers or servers linked together in big clusters. And start-ups like Parabon Computation of Fairfax, Va., and Entropia of San Diego, are trying to harness tens of thousands of home and business computers to work on genomics when they would otherwise be idle by distributing portions of the tasks to these computers over the Internet.

Paracel’s GeneMatcher machine, however, is especially built for genomic searchers. It has 7,000 processors arranged in a way best suited to matching character strings. Dr. Yu likened the process to having each letter of the sequence being studied in a separate processor strung along the inside of a hose. The database to be searched flows over these letters like water through a hose, at the rate of 30 million characters a second, and each processor sees whether it has a match as each letter flows by.

With a list price of $360,000, GeneMatcher would be too much for general-purpose computing. But for the task for which it was designed, it can be up to 1,000 times as fast as a Pentium-based computer, making it cost-effective, Dr. Yu said.

The market for machines like GeneMatcher, which are called genomics accelerators, is tiny. Paracel sells only one or two a month, so few in fact that it assembles its machines by hand in one room of a high-rise office here.

It says it has more than 30 customers, including drug companies like Novartis, Bayer and AstraZeneca, as well as some biotechnology and bioinformatics companies. No customer has more than two machines, other than Celera, which bought four last year.

But sales are growing. Paracel, which was privately held before its acquisition by Celera, had sales of $14.2 million in 1999, up from $5.9 million the year before. Dr. Yu said Paracel was not profitable because of investments but had been in the past.

GeneMatcher accounted for $5.2 million of Paracel’s 1999 sales, and genomics software $1 million. Textfinder machines, used by the government agency and a few other customers to search text databases, accounted for $8 million.

Paracel’s main competitor, TimeLogic, based in Incline Village, Nev., had sales of $3 million in 1999 and was profitable, according to its chief executive, James W. Lindelien. Sales of its DeCypher accelerator rose 350 percent in 1999 and should rise 50 percent this year, he said.

Compugen, an Israeli company that just went public, also sells an accelerator but is de-emphasizing that business.

Some experts think that conventional server farms can do the job less expensively than genomics accelerators can and can also be used for other tasks. The National Center for Biotechnology Information, which runs GenBank, uses about 140 conventional computers that are tied together. It is possible, said one expert, to define the problem so that you don’t have to compare everything with everything else in the entire database.

Indeed, history is not on the side of dedicated hardware. Machines designed for artificial intelligence, computer-aided design, database storage and graphics were all eventually replaced by general-purpose computers, which became fast enough to handle those jobs and cost less because of their huge production volumes. The large sales volumes of general-purpose computers also generate more money for research and development at the companies that produce them, allowing them to improve their machines faster than companies that design specialized computers.

Paracel’s acquisition by Celera will give the company more financial resources to help it keep up. But some experts were shocked that Celera would spend so much money for such a tiny, niche company.

But Dr. Venter of Celera said the Paracel machines would be important. “There’s no university or pharmaceutical company that has the compute capability they need to deal with our data or anyone else’s genomic data,” he said.

Celera must persuade companies to pay for its genomic data when GenBank offers much of the same human genome data free. One way to compete is to offer more and better data. But another, Dr. Venter said, is to use the Paracel machines to offer Celera customers faster searches than they could get using GenBank.

Paracel technology can also be used to compare the sequence of amino acids that make up a protein and could become part of a protein analysis machine that is being designed by Applied Biosystems, Celera’s sister company. Both Celera and Applied Bio, formerly known as PE Biosystems, are subsidiaries of the PE Corporation. Celera has one of the largest computer centers in the world, costing more than $50 million, and made up mainly of Compaq servers. By May, Celera had bought only $1.9 million of Paracel machines. But Dr. Venter said that while the Compaq machines were used to assemble the genome sequence in the correct order, the Paracel machines would be used to search the database.

PE and Paracel have long been partners. In 1996 PE bought a 14 percent stake in Paracel for $4.5 million. Dr. Yu himself owned about 17 percent of Paracel, giving him Celera stock worth about $35 million when the the acquisition closed in June.

Dr. Yu, 50, grew up in Hong Kong, Taiwan and Malaysia and came to the United States for college. After earning his doctorate in computer science from the California Institute of Technology, he joined TRW, the defense contractor, where he worked on development of the text-searching chip. In 1992 TRW spun out Paracel to commercialize the technology and before the buyout held 13 percent of the company.

The acquisition by Celera will benefit Paracel, but it could also cause genomics companies that compete with Celera to avoid buying from Paracel. TimeLogic is already playing up this factor in its sales pitch. And to counter Paracel’s alliance with Celera, TimeLogic is expected to announce soon that its machines will be distributed by Sun Microsystems.

Dr. Yu said he did not think customers would desert Paracel. And Paracel will obtain information from Celera that could help it design better machines.

“I really want to be part of something that’s the definitive provider of tools to mine that information,” he said. “I can get there faster by merging with Celera.”

============================================================

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!

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 XL — were added to the benchmark suite as MLPerf continues 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 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…

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…

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…

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