Of course. But from the sound of your question, it doesn’t sound like you realize that is not what is being benchmarked. In fact there is now way to even compare application code performance between the U.S and China, because there are no set standards by which to measure, nor is that code even available for review. As far as anyone knows, their application code could far exceed ours, so with all respect your question seems completely off base. Or more likely, a diversion attempt to justify our extremely sensitive U.S government code being put into the foreign born Linux clone of Unix, then given to the rest of the world for free.
My question was aimed at your original assertion that Linux was to blame for China being in the top spot.
Not really sure what your claim is? What did NASA supposedly “give away”? SMP scheduling? Their MPI implementation? Modern supercomputers are built to use fairly standard software components. Not much “secret sauce” here that the Chinese couldn’t do themselves if they needed to.
You almost make sense. Data management belongs to the OS, and to,the app; data manipulation mainly to the app (yes, I know there are certain optimized shell functions, but they’re a PITA for all but simple files). The key is that a bottleneck in either one will kill performance, see also Amdahl’s Law. So why solve half the problem for the Chicoms? (Not to mention Lenovo itself as a security risk.)