This page contains a Flash digital edition of a book.
HPC predictions Silvina Grad-Freilich, parallel-computing marketing manager at MathWorks A


s HPC hardware continues to deliver greater compute power, more and more companies are looking to see how they


can harness these hardware advancements. Soſtware will be increasingly critical in this effort. Programmers dealing with large, complex problems need to be able to access additional computing capacity easily. In the current uncertain economic environment, businesses are forced to look very carefully at their ROI and determine how to best structure their IT resources to balance hardware, soſtware and cloud-based infrastructure. Finally, to support growing demand for using


HPC technologies, companies will need to work with academia to develop skilled programming talent. Typically, businesses and organisations


want to get the most mileage from HPC resources for the highest ROI for their investment. Advances in hardware technology have been, and continue to be, very rapid. GPUs, for example, deliver immense compute power within desktop machines. Organisations already have this power, however GPUs require time for data transfer and are notoriously technical to work with. Businesses need to ask what makes most


sense for them and how they can ensure, if they invest resources in the cloud or in a supercomputer, that their staff will be able to utilise those resources. Tere is a need for HPC expertise and the pipeline of new graduates in this area needs to build. It’s important that academic institutions and industry collaborate to build engineering curricula that inspire students to pursue engineering fields, including HPC.


Bill Nitzberg, CTO for PBS Works at Altair Engineering O


ne of the biggest talking


points in 2012 will be GPUs, partly because Intel’s going to get into the party with


a big splash, and partly because power and ‘green’ computing will continue to be trending topics and GPUs offer a possible solution. I do believe that a real change this year will


be that we’ll see more delivery in areas such as cloud and Big Data, rather than just hype. A lot of the excitement surrounding cloud is coming from the data centre world where it has huge value and I hope that some of the


practical things HPC can do with cloud will come into being. Tere are some sides of the market that are not discussed for HPC, but Big Data is slowly being relabelled for this sector and a few people in HPC are discovering what can be achieved. It’s not as exciting as the hype, but it’s real. What does this mean for HPC users? Te


seasoned HPC users won’t feel any change whatsoever this year, and have probably never felt any. Tings get a little bit faster and sometimes they get more complicated, sometimes they get easier. Tere is always this feeling amongst HPC users that the way they do their work could be improved, and I don’t think that will change. With the non-hardened user – the scientist or engineer that is just


beginning to explore HPC – I believe that we will see a greater use of GUIs for accessing resources. Te big benefit we got from the hype


surrounding cloud is that it made us realise there is a well-defined interface that can be placed between individuals and computers that ends the need to train scientists in computer science. If you look at the adoption curve for other


GUI technology, such as the iPad, you will see that most of what’s learned and developed comes from fellow users and I expect the same will be true in HPC amongst engineers. Tat’s going to be the biggest difference and I hope it means a lot more people will be able to get the most out of HPC.


Shai Fultheim, founder and CEO of ScaleMP I


believe that we’ll see three main areas of focus in 2012. First and foremost will be the entry-level users coming into the market.


Te main issue in technical computing is that the IT infrastructure required to run the off-the-shelf tools designed for the entry-level user is quite a complicated one. Tere is a focus, however, on resolving that last mile of problems that prevents users from doing everything that technical computing will allow them to do. You can see the evidence of that within that space – IBM’s purchase of Platform Computing, for example, is providing more infrastructure tools to the technical computing market. Virtualisation technology can be an accelerator for these entry-level organisations to grow and use


32 SCIENTIFIC COMPUTING WORLD


HPC to its fullest, without having the burden of complex IT infrastructures. Te second area of focus in 2012 is Big


Data. Last year was the year of Big Data, primarily because we defined the problem. Tis year we will start to solve the problem and this can be done in one of two ways: we can either bring the data issue to the compute, or move the compute to the data. Most of what’s been done so far was the latter and the infrastructures that have evolved over the past few years have reflected this. Transferring the data is a difficult prospect but that’s where virtualisation will come in – the execution environment can be centralised by virtualising the distributed resources as one. Lastly, we come to virtualisation for


technical computing. Tere is a difficulty when virtualisation and HPC are discussed together as there is the contradictory thought that HPC is about getting the best possible performance, while virtualisation is about efficiency regardless of the impact on performance. Tis is not always true, as in the case of virtualising multiple systems as one, and people are beginning to realise the possibilities. It is definitely becoming more of a trend and I believe we will see significant developments in this area.


www.scientific-computing.com


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  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48