HIGH PERFORMANCE COMPUTING
include some specialisation. However, there are other examples like Density Functional Theory (DFT) codes, which are material science codes that have a handful of algorithms, which, if you could accelerate them, you would get a lot of bang for your buck. It is possible that we will see both kinds of specialisation, the GPU kind, which is very broad, or some more narrow specialisations that might be targeted to one application.
How does architecture specialisation change the way HPC systems are designed and purchased? It may be the case that in the future we have to decide how much of the capital acquisition budget will go into working together with a company to add specialisations to the machine through non-recurring engineering expenses. It might change the model of acquisition, where you have to make a decision about how much of your budget you’re willing to put into R&D, as opposed to just strictly acquisition and site preparation costs. If you look at the mega data centre
market, it is already happening, so it is a question of when is HPC going to catch up?
Microsoft Research have Project Catapult, which has FPGAs integrated throughout the interconnect on the machine to do processing in-network.
Google has its TPU and it is already on its third generation. Amazon has its own specialised chip that it is designing using Arm IP – that’s one way to reduce the costs of specialisation, to use technology from the embedded IP ecosystem. So the mega datacentres are already doing this, it is a forgone conclusion that this is an approach that is being adopted, the question is, how do you adopt it productively for scientific computing?
How much benefit do CMOS replacements offer? The answer is that we don’t know what is physically possible, but we do know the fundamental limit in physics for digital computing; it is the Landauer limit. We are many orders of magnitude above the Landauer limit. There is a lot of room at the bottom but we don’t know the physical limits of the devices we can construct. The materials that we have to construct
these devices, the pace at which we are able to discover those devices and the expense it takes to create just one of those devices as a demonstration – the process is incredibly slow and very artisanal. Because of the urgency of the issue,
we have started a lab-wide initiative for Beyond Moore’s Law Microelectronics to industrialise the process using modelling and simulation of candidate materials, using something called the Materials
‘How do we create customisations that are effective for scientific computing?’
Project. This aims to optimise the search for candidate materials. You say what characteristics you want to optimise, then the materials project framework can automate that search for better materials. Sifting through tens of thousands of materials, it can find the handful that have that optimised property. You then conduct device-scale simulation. Researchers do full ab initio material science simulations using a code called LS3DF, which is able to do those kinds of device-scale simulations. It takes a whole supercomputer to be able to do it, but it is so much better to simulate the behaviour of the device before you construct it, because it is so costly to fabricate them.
The full version of this interview can be found on our website. In the last part Shalf discusses the unique capabilities of the DOE to accelerate these areas of research and also highlights how new computing paradigms can compliment digital computing by accelerating specific workloads.
Unveiling the secrets of life
IVO SBALZARINI DISCUSSES HOW RESEARCHERS ARE DEVELOPING COMPUTATIONAL METHODS AND SOFTWARE SYSTEMS TO UNDERSTAND BIOLOGICAL PROCESSES ON AN ALGORITHMIC BASIS
How would you describe your work? Over the past decades and centuries Biology has studied the components of living systems in isolation. This meant studying individual plants and animals, or studying molecules, genes, proteins and so on over the last couple of decades. An individual protein is not life, nor is a gene, but when they all come together,
www.scientific-computing.com | @scwmagazine
they form life. Systems biology is the science that seeks explanations of how this happens. This is, of course, a very hard problem, because the interactions are non-linear. We have millions or billions of components that interact in such a system. Interactions are regulated across a wide variation of length scales and time scales from molecular and intramolecular interactions, all the way up to interactions between animals or between ecosystems. And we are dealing with physics that is only partially known. There you want to leverage the power
of computing, on the one hand to simulate these systems on the computer, so using large HPC-type simulations where you take a hypothetical interaction – how you believe such a system works – and reconstitute that in a large computer simulation to show that it produces the
behaviour that you see. One of the more famous examples of this is the Human Brain Project, where the goal is to simulate a human brain, or part of a human brain, neuron by neuron, or even ion channel by ion channel, to answer the question of whether the collective interaction of all these neurons is enough to generate intelligence.
How do these challenges differ from classical biology? Systems biology started out around 2000 by systematic experimentation. A lot of robotic experimentation, robotised microscopes and high-throughput experiments with pipetting robots and liquid-handling robots that were able to do thousands of experiments in an automated and very systematic fashion – producing enormous datasets.
June/July 2019 Scientific Computing World 5
g
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32
