The main point of IBM was that by using RAM and Disk storage you could perform computations of the google system that were not as onerous as was put out in the google paper. That said, if you had an extra 10+ qubits (which is essentially what this new work does) you are back in the range where the supercomputer will be taking more time than is reasonable. For a quick eyeball imagine what adding 10+ qubits on to this kind of plot would do (from ibm post)
https://www.ibm.com/blogs/research/wp-content/uploads/2019/1...
Oh, and you need a lot of storage, just a couple of more qubits and it would not have been able to fit in the 250 PiB on Summit, so the IBM argument would have been void if google had 55 qubits in their work.
Can not agree with this more, if people want to plot something that is linear please use a perceptually linear colormap! Just a one second glance at the Mona Lisa in rainbow/Jet is enough to make you gouge your eyes out.
https://peterjamesthomas.com/2017/09/15/hurricanes-and-data-...
For a more technical description the information behind the newer matplotlib defaults, particularly the scipy talk, is great.
https://bids.github.io/colormap/
And for those that do not like the matplotlib options, colorcet provides a wider range of alternatives that are not trash (unlike Jet)
https://colorcet.holoviz.org/index.html
Cognitive. Things like having to strip down abstractions and "write it like Fortran". The fact that it can deal with numpy arrays no problem and can actually deal with more common python objects like dicts means that there is less overhead
I had considered Numba in the past but it just seemed not worth the overhead. A few talks from this year show that they have really expanded the library, to the point where much of the scientific python stack use it instead of Cython. It can target things like ARM devices and is more flexible in the types it can take (dicts!)
For reference
https://www.youtube.com/watch?v=cR8E70GTO8c
and
https://www.youtube.com/watch?v=6oXedk2tGfk
Oh, and you need a lot of storage, just a couple of more qubits and it would not have been able to fit in the 250 PiB on Summit, so the IBM argument would have been void if google had 55 qubits in their work.