Connect with us

Quantum

Intel and Argonne National Lab on exascale and their new Aurora supercomputer

Published

on

The scale of supercomputing has grown almost too large to comprehend, with millions of compute units performing calculations at rates requiring, for first time, the exa prefix — denoting quadrillions per second. How was this accomplished? With careful planning… and a lot of wires, say two people close to the project.

Having noted the news that Intel and Argonne National Lab were planning to take the wrapper off a new exascale computer called Aurora (one of several being built in the U.S.) earlier this year, I recently got a chance to talk with Trish Damkroger, head of Intel’s Extreme Computing Organization, and Rick Stevens, Argonne’s associate lab director for computing, environment and life sciences.

The two discussed the technical details of the system at the Supercomputing conference in Denver, where, probably, most of the people who can truly say they understand this type of work already were. So while you can read at industry journals and the press release about the nuts and bolts of the system, including Intel’s new Xe architecture and Ponte Vecchio general-purpose compute chip, I tried to get a little more of the big picture from the two.

Intel and Cray are building a $500 million ‘exascale’ supercomputer for Argonne National Lab

It should surprise no one that this is a project long in the making — but you might not guess exactly how long: more than a decade. Part of the challenge, then, was to establish computing hardware that was leagues beyond what was possible at the time.

“Exascale was first being started in 2007. At that time we hadn’t even hit the petascale target yet, so we were planning like three to four magnitudes out,” said Stevens. “At that time, if we had exascale, it would have required a gigawatt of power, which is obviously not realistic. So a big part of reaching exascale has been reducing power draw.”

Intel’s supercomputing-focused Xe architecture is based on a 7-nanometer process, pushing the very edge of Newtonian physics — much smaller and quantum effects start coming into play. But the smaller the gates, the less power they take, and microscopic savings add up quickly when you’re talking billions and trillions of them.

But that merely exposes another problem: If you increase the power of a processor by 1000x, you run into a memory bottleneck. The system may be able to think fast, but if it can’t access and store data equally fast, there’s no point.

“By having exascale-level computing, but not exabyte-level bandwidth, you end up with a very lopsided system,” said Stevens.

And once you clear both those obstacles, you run into a third: what’s called concurrency. High performance computing is equally about synchronizing a task between huge numbers of computing units as it is about making those units as powerful as possible. The machine operates as a whole, and as such every part must communicate with every other part — which becomes something of a problem as you scale up.

“These systems have many thousands of nodes, and the nodes have hundreds of cores, and the cores have thousands of computation units, so there’s like, billion-way concurrency,” Stevens explained. “Dealing with that is the core of the architecture.”

How they did it, I, being utterly unfamiliar with the vagaries of high performance computing architecture design, would not even attempt to explain. But they seem to have done it, as these exascale systems are coming online. The solution, I’ll only venture to say, is essentially a major advance on the networking side. The level of sustained bandwidth between all these nodes and units is staggering.

Making exascale accessible

While even in 2007 you could predict that we’d eventually reach such low-power processes and improved memory bandwidth, other trends would have been nearly impossible to predict — for example, the exploding demand for AI and machine learning. Back then it wasn’t even a consideration, and now it would be folly to create any kind of high performance computing system that wasn’t at least partially optimized for machine learning problems.

“By 2023 we expect AI workloads to be a third of the overall HPC server market,” said Damkroger. “This AI-HPC convergence is bringing those two workloads together to solve problems faster and provide greater insight.”

To that end the architecture of the Aurora system is built to be flexible while retaining the ability to accelerate certain common operations, for instance the type of matrix calculations that make up a great deal of certain machine learning tasks.

“But it’s not just about performance, it has to be about programmability,” she continued. “One of the big challenges of an exacale machine is being able to write software to use that machine. oneAPI is going to be a unified programming model — it’s based on an open standard of Open Parallel C++, and that’s key for promoting use in the community.”

Summit, as of this writing the most powerful single computing system in the world, is very dissimilar to many of the systems developers are used working on. If the creators of a new supercomputer want it to have broad appeal, they need to bring it as close to being like a “normal” computer to operate as possible.

“It’s something of a challenge to bring x86-based packages to Summit,” Stevens noted. “The big advantage for us is that, because we have x86 nodes and Intel GPUs, this thing is basically going to run every piece of software that exists. It’ll run standard software, Linux software, literally millions of apps.”

I asked about the costs involved, since it’s something of a mystery with a system like this how that a half-billion dollar budget gets broken down. Really I just thought it would be interesting to know how much of it went to, say, RAM versus processing cores, or how many miles of wire they had to run. Though both Stevens and Damkroger declined to comment, the former did note that “the backlink bandwidth on this machine is many times the total of the entire internet, and that does cost something.” Make of that what you will.

Aurora, unlike its cousin El Capitan at Lawrence Livermore National Lab, will not be used for weapons development.

$600M Cray supercomputer will tower above the rest — to build better nukes

“Argonne is a science lab, and it’s open, not classified science,” said Stevens. “Our machine is a national user resource; We have people using it from all over the country. A large amount of time is allocated via a process that’s peer reviewed and priced to accommodate the most interesting projects. About two thirds is that, and the other third Department of Energy stuff, but still unclassified problems.”

Initial work will be in climate science, chemistry, and data science, with 15 teams between them signed up for major projects to be run on Aurora — details to be announced soon.

Read more: https://techcrunch.com/2019/11/18/intel-and-argonne-national-lab-on-exascale-and-their-new-aurora-supercomputer/

Quantum

Deep tech VCs on what they view as some of the most impactful young startups right now

Published

on

By

During this week’s Democratic debate, there was a lot of talk, unsurprisingly, about ensuring the future of this country’s children and grandchildren. Climate change was of particular interest to billionaire Tom Steyer, who said repeatedly that addressing it would be his top priority were he elected U.S. president.

As it happens, earlier the same day, we’d spent time on the phone with two venture capitalists who think of almost nothing else every day. The reason: they both invest in so-called deep tech, and they meet routinely with startups whose central focus is on making the world habitable for generations of people to come — as well as trying to produce outsize financial returns, of course.

The two VCs with whom we talked know each other well. Siraj Khaliq is a partner at the global venture firm Atomico, where he tries to find world-changing startups that are enabled by machine learning, AI, and computer vision. He has strong experience in the area, having cofounded The Climate Corporation back in 2006, a company that helps farmers optimize crop yield and that was acquired by Monsanto in 2013 for roughly $1 billion.

Seth Bannon is meanwhile a founding partner of Fifty Years, a nearly five-year-old, San Francisco-based seed-stage fund whose stated ambition is backing founders who want to solve the world’s biggest problems. The investors’ interests overlap so much that Khaliq is also one of Fifty Years’s investors.

From both, we wanted to know which companies or trends are capturing their imagination and, in some cases, their investment dollars. Following are excerpts from our extended conversation earlier this week. (We thought it was interesting; hopefully you will, too.)

TC: Seth, how would you describe what you’re looking to fund at your firm?

SB: There’s a Winston Churchill essay [penned nearly 100 years ago] called “Fifty Years Hence” that describes what we do. He predicts genomic engineering, synthetic biology, growing meat without animals, nuclear power, satellite telephony.  Churchill also notes that because tech changes so quickly that it’s important that technologists take a principled approach to their work. [Inspired by him] we’re backing founders who can make a ton of money while doing good and focusing on health, disease, the climate crisis . . .

TC: What does that mean exactly? Are you investing in software?

SB: We’re not so enthusiastic about pure software because it’s been so abstracted away that it’s become a commodity. High school students can now build an app, which is great, but it also means that competitive pressures are very high. There are a thousand funds focused on software seed investing. Fortunately, you can now launch a synthetic biology startup with seed funding, and that wasn’t possible 10 years ago. There are a lot of infrastructural advancements happening that makes [deep tech investing even with smaller checks] interesting.

TC: Siraj, you also invest exclusively on frontier, or deep tech, at Atomico . What’s your approach to funding startups?

SK: We do Series A [deals] onward and don’t do seed stage. We primarily focus on Europe. But there’s lot of common thinking between us and Seth. As a fund, we’re looking for big problems that change the world, sometimes at companies that won’t necessarily be big in five years but if you look out 10 years could be necessary for humanity. So we’re trying to anticipate all of these big trends and focus on three or four theses a year and talk as much as we can with academics and other experts to understand what’s going on. Founders then know we have an informed view.

Last year, we focused on synthetic biology, which is a becoming so broad a category that it’s time to start subdividing it. We were also doing AI-based drug discovery and quantum computing and we started to spend some time on energy as well. We also [continued an earlier focus on ] the future of manufacturing and industry. We see a number of trends that make [the latter] attractive, especially in Europe where manufacturing hasn’t yet been digitized.

TC: Seth, you mentioned synthetic biology infrastructure. Can you elaborate on what you’re seeing that’s interesting on this front?

SB: You’ve maybe heard of directed evolution, technology that allows biologists to use the power of evolution to get microbes or other biological machines to do what they want them to do that would have been impossible before. [Editor’s note: here, Bannon talks a bit about Frances Arnold, the Nobel Prize-winning chemist who was awarded the honor in 2018 for developing the technique.]

So we’re excited to back [related] startups. One, Solugen, enzymatically makes industrial chemicals [by combining genetically modified enzymes with organic compounds, like plant sugars]. Hydrogen peroxide is a $6 billion dollar industry, and it’s currently made through a petroleum-based process in seven-football-field-long production plants that sometimes explode and kill people.

TC: Is this then akin to Zymergen, which develops molecules in order to create unique specialty materials?

SB: Zymergen mainly works as a kind of consultant to help companies engineer strains that they want. Solugen is a vertically integrated chemicals company, so it [creates its formulations], then sells directly into industry.

TC: How does this relate to new architectures?

SB: The way to think about it is that there’s a bunch of application-level companies, but as synthetic biology companies start to take off, there’s a bunch of emerging infrastructure layer companies. One of these is Ansa Biotechnologies, which has a fully enzymatic process for writing DNA. Like Twist, which went public, they make DNA to sell to customers in the biotech industry. But whereas Twist is using a chemical process to make DNA, Ansa’s approach is fully enzymatic. [Editor’s note: More on the competition in this emerging space here.]

Also, if you look at plant-based alternatives to meat, they’re more sustainable but also far more expensive than traditional beef. Why is that? Well plant-based chicken is more expensive because the processing infrastructure being used is more than 10 years behind real chicken processing, where you’ll see robot arms that cut up chicken so efficiently that it looks like a Tesla factory.

[Alternative meat] companies are basically using these extruders built in the ’70s because the industry has been so small, and that’s because there’s been a lot of skepticism from the investment community in these companies. Or there was. The performance of Beyond Meat’s IPO ended it. Now there’s a rush of founders and dollars into that space, and whenever you have a space where the core infrastructure has been neglected, there’s opportunity. A former mechanical engineer with Boeing has started a company, Rebellyous Foods, to basically build the AWS for the plant-based food industry, for example. She’s using [the machines she’s building] to sell plant-based chicken nuggets, [but that’s the longer-term plan].

TC: Siraj, you say last year you started to spend time on energy. What’s interesting to you as it relates to energy?

SK: There’s been some improvement in how we capture emissions, but [carbon emissions] are still very deleterious to our health and the planet’s health, and there are a few areas to think about [to address the problem]. Helping people measure and control their consumption is one approach, but also we think about how to produce new energy, which is a shift we [meaning mankind] need to undertake. The challenge [in making that shift] is often [capital expenditures]. It’s hard for venture investors to back companies that are [building nuclear reactors], which makes government grants the best choice for early innovation oftentimes. There is one company, Seaborg, that has figured out a clever reactor. It’s not a portfolio company but it’s [compelling].

SB: We also really like what Seaborg is doing. These [fourth generation] nuclear companies have a whole host of approaches that allow for smaller, safer reactors that you wouldn’t mind having in your backyard. But Siraj put his finger on it: as an early-stage deep tech investor, we have to consider the capital plan of a company, and if it needs to raise billions of dollars, early investors will get really diluted, so early-stage venture just isn’t the best fit.

TC: There are other areas you like, though, because costs have fallen so much.

SB: Yes. Satellite telephony used to be one of those areas. Some of the satellites in space right now cost $350 million [to launch] and took three to four years to build, which would be really hard for any early-stage investor to fund. But now, a new generation of companies is building satellites for one-tenth of the cost in months, not years. That’s a game changer. They can iterate faster. They can build a better product. They don’t have to raise equity to build and launch either; they can raise from a debt financier [from whom they can] borrow money and pay it back over time. That model isn’t available to a company like Uber or Lyft, because those companies can’t say, ‘X is going to cost us Y dollars and it will pay back Z over time.’

TC: What of concerns that all these cheap satellites are going to clog up the sky pretty quickly?

SB: It’s a real concern. Most [of today’s satellites] are low earth satellites, and the closer to the earth they are, the brighter they are; they reflect the sun more, the more satellites we’re seeing instead of stars. I do think it’s incumbent on all of these companies to think about how they are contributing to the future of humanity. But when you connect the unconnected, educational outcomes improve, health improves, inequality decreases, and the stability of governments improves, so maybe the developed world needs to sacrifice a bit. I think that’s a reasonable tradeoff. If on the other hand, we’re putting up satellites to help people buy more crap . . .

TC: It’s like the argument for self-driving cars in a way. Life becomes more efficient, but they’ll require far more energy generation, for example. There are always second-order consequences.

SK: But think of how many people are killed in driving accidents, versus terrorist attacks. Humans have many great qualities, but being able to drive a lethal machine consistently isn’t one of them. So when we take that into perspective, it’s really important that we build autonomous vehicles.

You [voice] a legitimate concern, and often when there are step changes, there are discontinuities along the way that lead to side effects that aren’t great. That comes down to several things. First, infrastructure will have to keep up. We’ll also have to create regulations that don’t lead to the worst outcomes. One our investments, Lilium in Munich, has built an entirely electric air taxi service that’s built on vertical takeoff. It’s nimble. It’s quiet enough to operate in city environments.

On roads, cars are constrained by 2D terrain and buildings, but [in the air] if you can do dynamic air traffic control, it opens up far much efficient transport. If you can get from downtown London to Heathrow [airport] in five minutes versus 50 minutes in a Tesla? That’s far more energy efficient.

Read more: https://techcrunch.com/2020/01/17/deep-tech-vcs-on-what-they-view-as-some-of-the-most-impactful-young-startups-right-now/

Continue Reading

Quantum

Congratulations to QECDT Cohort 1!

Published

on

By

Quantum Engineering CDT students begin to submit their theses and complete their vivas

The very first cohort to take part in the QECDT PhD programme have begun to complete their studies, with some having already successfully passed their vivas. A few of our students have now left Bristol to begin their quantum career with companies such as Renishaw and Rigetti. Others have taken up Post Doctoral positions within the QETLabs community here in Bristol. The QECDT management team are extremely proud of our first cohort for reaching this milestone and would like to say a huge congratulations to them!

To the students who are continuing to write-up and prepare for their viva, we’d like to wish you the best of luck! The entire cohort has helped to shape the QECDT over the last four years and we look forward to maintaining our relationship with them for many years to come.

Source: http://www.bristol.ac.uk/quantum-engineering/news/congratulations-to-qecdt-cohort-1.html

Continue Reading

Quantum

Submissions for The Europas Tech Startups Awards are now open

Published

on

By

Submissions for The Europas Awards 2020 have now opened. We’re back for our 11th year of recognizing the hottest tech startups across the European tech scene, as supported by TechCrunch.

The awards evening will be held on 25 June 2020 in London, U.K., at the Museum of the Home (formerly known as the Geffrye Museum). Earlier in the day, we will be running a series of Pathfounder workshops aimed at giving Series A and late-stage seed startups practical advice on scaling up. The working agenda is here.

We’re actively scouting the top startups in each category, but you can nominate a startup, accelerator or venture fund that you think deserves to be recognized for their achievements in the last 12 months (1 January 2019 – 1 January 2020).

To be eligible to apply for the startup portion of the awards, you must be a late-stage seed or Series A startup. If you are bootstrapped, you must be generating sustainable revenues of at least £500,000 a year. For unicorns minted in the last year (1 January 2019 – 1 January 2020), you can apply to Unicorn of the Year.

This year, there are 22 categories, refreshed to ensure we recognize those startups working in tech’s most compelling fields. This year, ag tech and food tech have their own categories. We have included three tech-specific awards: Hottest AI Startup, Hottest Blockchain Project and Hottest Quantum Computing startup.

As we roll into our 11th year of The Europas, and into a new decade, we’re very conscious that collectively, we have only 10 years left to reach the UN’s ambitious Sustainable Development Goals. We want to recognize those startups striving to create a more sustainable, equitable future. To that end, three new categories have been added, including Hottest GreenTech Startup, Hottest Social Innovation and Hottest Sustainability Tech. Sustainability Tech includes any startup working in an industry for which they’ve built a sustainable product or service.

We are also introducing one new final award, “Pathfounder of the Year,” to honor a person who has made a significant contribution to the tech industry. The contribution may be over a lifetime, or it may have occurred in the past year.

The application is here. Early-bird tickets are here.

The timeline for The Europas are:

14 January: Entries open to the 2020 Europas.
25 March: Entries close.
14 April: Long list announced. Public voting opens.
15 May: Public voting closes. Judges deliberation begins.
8 June: Short list announced.
25 June: Winners announced at The Europas Awards evening.

The complete list of categories are:

Hottest AgTech Startup
Hottest B2B / SaaS Startup
Hottest CyberTech Startup
Hottest EdTech Startup
Hottest FinTech Startup
Hottest FoodTech Startup
Hottest GreenTech Startup
Hottest HealthTech Startup
Hottest Mobility Tech Startup
Hottest PropTech Startup
Hottest Public, Civic, GovTech Startup
Hottest SpaceTech Startup
Hottest Social Innovation
Hottest Sustainability Tech Startup
Hottest AI Startup
Hottest Blockchain Project
Hottest Quantum Startup
Hottest Accelerator
Hottest Seed Fund
Hottest VC Fund
Unicorn of the Year
Pathfounder of the Year

Read more: https://techcrunch.com/2020/01/15/submissions-for-the-europas-tech-startups-awards-are-now-open/

Continue Reading

Trending