At the risk of explaining what everyone here already gets: I like how the bottom picture could literally be a diagram of a 4-way deadlock where the buses are threads of execution and the roundabout exits are object locks.
Whereas in top picture there are no passengers and I’m fairly sure the buses aren’t moving. Which probably says something about the state of the documentation
hang on, that first image is everything on a single thread…
Each bus is a timeslice
Nah, those are double deckers.
You sure they’re not dwords on a single bus?
Making the word bigger doesn’t mean there’s more threads per, it just reduces the number of calls to complete some kinds of ops (and something about RAMing busses into each other and… cache locations? idk I’m sleepy)
The double deckers are an example of a ready-made solution in the original language’s standard library, the lower one is getting multithreading working through the C ABI bindings, using some 3rd party solution, all while multithreading a lot of other tasks in the application.
Yeah, I can live with that. Well done!
That’s because they’re using Little Endian in the first picture, and Big Endian in the second picture.
No wonder they’ve got unpatched race conditions if they used the wrong flag at compile time
The tasks are all awaiting each other
“Save us, SuperScheduler! Doctor resource-locking has us trapped!”
Are you a bot?
Would you believe me if I said no?
Also, why?
Because you’ve got a lot of comments simply describing the images of the posts you comment on. I’ve seen obvious bot accounts with that behavior.
💀
Serial parralelization…?
All I wanted was to call ‘read’ while I listen on a socket!
Where did it all go so wrong!?
That’s why they stack the buses instead of linking them.
You didn’t read the documentation correctly. You’ve switched road side.
Locky, not lock free!
