— That is what the system does
You can make 2 million functions run as if they were only a single function, and experience time travel
— I do not even use the term time travel lightly, this is a technical term in this context
— Who will split then? if not human programmer🤤 i mean, how it will detect that this function should wait for the result of that function?
Message permanent page
— That cant be resolved automaticly
— So you have to provide the API
— It absolutely can be resolved automatically, and transparently, and without the users involvement or even their knowledge that it is happening
— It is very easy to detect that a future calculation relies on a past calculation, so that you can decide which calculations are allowed to run simultaneously, and which calculations must wait
— Yeah, very easy
— Super easey
— Then you can split each unit of work, package it into an individual instruction token and data token, and transmit them through the instructions stream, with each one making it to its very own CPU core in the cluster
— Do you have such "detector" already or it's in plans?