Message from JavaScript discussions

October 2018

— [a, b, c] - parralel, until all finished


I am in the process of designing a hardware model to run this entire system, and extremely clustered system of specialized super computers, with the only responsibility being to execute the instruction streams being sent to them, 100% agnostically to what they are or where they come from

— For every CPU core in the system, there is a single instruction stream sent to it over the network

— There is no multitasking, there is no event loop, there is simply data input, computation, and data output

Message permanent page

— Too much mashinery!

— Reduce

— Right now each clustered system consists of only two components, V8, and a Linux kernel with almost all multitasking features completely disabled, except those necessary to operate the networking stack and other needed system services

Message permanent page

— Complexity

— Of what you do

— It is reduced

— It is not complex

— It is only the bare minimum and components needed to run the system at a minimal overhead, it is so elegantly simple, but no one has actually tried it before

Message permanent page