Supercomputing to Feed the World

Print Friendly, PDF & Email

Sharan Kalwani writes that that supercomputers a critical tool for predicting severe weather that could threaten entire regions with famine.

Sharan Kalwani

Picking up on the thread of the last topic I shared with readers of insideHPC.com: “HPC benefits for Society,” I had touched upon using Earthquake models as one of the examples. Another area which is close to my heart and one perhaps most familiar to the community since the last several decades is weather forecasting. In particular one of the most compelling applications of HPC was the use in helping to predict the monsoon season in the SE Asian sub-continent of India.

But most folks may be rather unaware that despite the recent advances there in industrialization, the bulk of the Indian economy is driven largely by agriculture. By some estimates almost 70% of the vast population (1.2 billion) are engaged one way or the other in agrarian pursuits. Also given that it is one of the oldest civilizations, there is a lot of cultural references to how weather and climate have played a role in the farming communities.

Imagine this: if the farmer plants too early or the rains arrive too late, the crops can wither and die, thus providing for a very bleak outlook for millions of people. On the other hand if planted too late or the rains come early and are excessive, then the various crops suffer, by getting dislodged, loosening the rich fertile soil to the point of a total loss, washing away into the rivers and a poor harvest results in suffering again for the population.

Thus predicting the precise arrival of the monsoon to as close as possible, in different parts of the country, south to north and in an easterly direction is hugely critical for the well-being of a large section of humanity. Scientific methods and models for accurate climate progress are a must and do determine outcome of society progress. It was in this vein that the nation of India, via the Indian Meteorological Department back in the 1980s decided to go for supercomputing to aid it in this task. During those days, it was pretty much Cray Research Inc. systems which were the first choice for such an endeavor. Sadly due to political circumstances, the procurement got embroiled in a wrangle, which after several rounds of extremely high level government level negotiations and agreements, finally got resolved. Of course it took a few years after the installation of a Cray XMP, tweaking of the software models, training of the scientists, upgrading of the data collection, etc. to finally get the job of forecasting going.

There is now enough objective data on hand, which testified to the improvement of weather forecasts in various parts of the country. Indeed they also now supply many neighboring countries in the region, with updates as well. Over a period of a decade, the grain and crop yields improved, and of course via the use of other inputs (such as production of fertilizers, mechanized machinery, etc.) the food situation stabilized a great deal. One of the earliest anecdotal evidence in the early 1990s was the amazement of the general population that they noticed the weather temperatures started becoming accurate in the day to day city bulletins: now almost up to a 3 day outlook (earlier it was no more than a few hours). Models have been improved – although at too slow a pace for the general public at large – that now this HPC approach is taking root in more than just the monsoon predictions. Tropical storms are also important (India has a good long coast line almost 4800 miles), and cyclones are also quite rampant. Recently two Petaflop-capable machines have been installed in a special tropical weather department in the central part of the country.

What better testimony than for many hundred millions of simple farming folks, generations of families and entire villages, now happily putting their faith in using the fruits (pun intended) of HPC, seeing it as a way to really improve their lives.

Sharan Kalwani is a seasoned scientific, technical and computing professional. He has spent over 20+ years implementing many new and pioneering technologies from operating systems (*nix) , high performance computing (Cray, SGI, compute clusters), engineering applications (CAE simulations), networking (TCP/IP, Infiniband), operations (ITIL/ ITSM), scientific domain (BioInformatics) and project management. Sharan looks to increase the professional approach of every individual he interacts with. He enjoys teaching, contributing to STEM activities and publishing. He is a senior member of IEEE, ACM, Emeritus member of Michigan!/usr/group, and leads the SIG-Linux section of SEMCO.org. He is currently the Chair of the IEEE SE Michigan Education Society Chapter for 2017-2018 and heads the Professional Activities Committee for Engineers (PACE) for the same section. He has also a published author on the topic: “UNIX and TCP/IP Network Security”.

Opinions expressed are solely those of the author and do not necessarily represent anyone else.

Sign up for our insideHPC Newsletter