How many years it would take our laptops to be as fast as the fastest super computer in 2000

Question

How many years it would take our laptops to be as fast as the fastest super computer in 2000 according to Moore's law?

Answer 1

Let's suppose that by "fastest supercomputer in 2000", and as your metric of speed, you choose the High Performance LINPACK benchmark, as used in the TOP500 list. Depending on when "in 2000" means, you're either selecting as your metric of performance:

• 2.4 TFLOP/s, in June 2000
• 4.9 TFLOP/s, in November 2000

Setting aside all of the issues involved in using the LINPACK benchmark -- namely, that it no longer reflects useful performance metrics for today's HPC systems as they relate to actual application performance -- the assumption of Moore's Law growth is also...well, not terribly useful, since it relates to the growth in the number of transistors in integrated circuits (more or less, CPUs, which consist of integrated circuit(s)).

Thermal limitations are going to limit frequency scaling, so that's probably going to be more or less flat (maybe even go down) while number of cores increases. At 4 FLOP/s/cycle, assuming a middle-of-the-road 2.5 GHz processor yields 10 GFLOP/s/core. So, in your laptop, you'd need roughly 250-500 cores @ 2.5 GHz operating at their theoretical peak performance (which also doesn't happen). Efficiency (ratio of actual to peak performance) varies a lot for the High Performance LINPACK benchmark; it depends on a lot of different factors (memory bandwidth, interconnects, etc.). A pessimistic efficiency would be something like 20%; wildly optimistic would be something like 99%. Taking that into account, anywhere from 250-2500 cores @ 2.5 GHz. I'm not sure there's enough room for that number of cores in a laptop using current form factors, which leads me to believe it would be difficult to forecast. For desktops, Intel Knight's Landing will use 72 Atom cores, and they might run around 2.0 GHz, guesstimating from the Atom specs; you'd need a slew of those to reach your target. A Parallella board is 66 cores, and their literature claims it reaches 90 GFLOPs. I don't think the technology for this number of cores is mature yet, so...maybe one to two decades? It's far beyond the reach of current technology, and I'm not even sure it would be useful for everyday use.