Fujitsu is credited for creating Fugaku, the world’s strongest supercomputer. Earlier this month, the Japanese IT providers firm started providing high-performance computing as a service (HPCaaS), which is able to enable corporations to benefit from the 442 petaflops of computing energy that Fugaku offers. Petaflop is a measure of a supercomputer’s efficiency, and what number of calculations it might carry out per second. In easy phrases, a human being performing one calculation per second for 31,688,765 years would match what a pc system operating at 1PFLOP can do in a second. In an interview, Fujitsu’s chief know-how officer Vivek Mahajan stated HPC as a service has an enormous future in India, as do much more highly effective future applied sciences equivalent to quantum computing. Edited excerpts:
What’s the state of superior computing in India in the present day?
There’s lots of want for computing in India, however we imagine that in future, HPCs are what each finish person, such as you or me, can be utilizing. They’ve that want in the present day, however they will’t use it since you or I, nor many firms, should buy a supercomputer.
The best way we see that taking place is by computing as a service. HPC can be provided by way of the cloud, and also you as a person ought to solely must care about what your drawback is, not whether or not it wants quantum or non-quantum, and so forth.
Do you already see Indian companies turning to HPCs as a service? What is going to or not it’s used for?
Properly, it’s (Fugaku) not out there in India proper now. However I see lots of utilization for this. As an illustration, (it’s helpful) if you wish to put your cash in retail, shares, and so forth., and also you wish to mannequin that to seek out the most effective worth.
It’s a quite simple drawback, and you may resolve to run two cycles (supercomputers are priced based mostly on time spent to run a simulation on them) in a supercomputer as an finish client.
We simulated covid on Fugaku to calculate the best angle for sporting a masks to minimise the influence of the virus and the space folks needs to be at from one another.
Now we have additionally used it for tsunami predictions, which saves lives. We are able to do the identical in India for cyclones or floods. As well as, visitors, logistics and so forth. are actual issues in the present day.
However there are firms in India coping with such issues with out HPCs; why would we want HPCs?
I wouldn’t say that they’re fixing it (in the most effective method). How tough is it to unravel the covid drawback? I don’t assume that modelling is straightforward to do on regular computing platforms.
They’re in all probability utilizing on-premise HPCs for climate forecasting, however (with HPC as a service) every state can do their very own flood predictions, and so forth., and so they don’t must have an HPC.
Supply: Live Mint