@stman @thatguyoverthere @vidak imagine you have a bunch of functional ASICs on a multilane bus. the MOVE instruction just tells the computer to move some data between each functional unit to perform the computations. it's definitely more hardware-centric and may better support parallelism.

@stman @thatguyoverthere @vidak @50htz What you said makes sense to me. The MOVE32INT is a specific design of a TTA, so you could imagine something grander. The manual says it can handle massive parallel MIMD and operates on 4 data transports in parallel. Yes, still, MOVE is a Harvard architecture.

@stman @thatguyoverthere @vidak @50htz I think the MOVE32INT was designed to be compatible with an Abstract C Machine. It's a clever compiler. You can see their positioning when they state lack of virtual addressing, MMU, protection support or kernel mode as disadvantages :)

@stman @thatguyoverthere @vidak @50htz Like I've said before, Abstract State Machines ought to offer the modeling capability needed to formally describe your conceptual architecture. I really need to implement ASMs in Ada/SPARK but it's one of the many on my TODO list that doesn't seem to get shorter. :sadness:

@stman @theruran @50htz @thatguyoverthere

just a friendly reminder that we have a forum that supports longer messages a little better than the fediverse—and doesn’t bury them in big chains of notifications either!

i may even set up a mailing list, now that i have a proper mail server set up @ germination.systems/mail

i enjoy reading your messages stman—but, if i may be permitted to make a gentle suggestion (and please let me know if this is out of order), you should join an instance with a longer post character length than 500, that way you may be done with your communication in a single post!

I am sorry I am not able to be more helpful. I do try and follow along, but a lot of the time I dont have anything to really contribute other than maybe a question or casual comment because the discussion is more detailed than I am really able to be and I am fairly limited in time (especially over the summer months when I have a lot of outdoors work to do when I'm not on the clock).

The general idea of the OSI model is that you can focus on one layer and its immediate interfaces. The reality is that it's really hard to do anything new because abstractions leak up and down the model and layers get combined in any attempt then need to be separated again to be standardized, e.g. quic and http3
@theruran @thatguyoverthere @vidak

Sign in to participate in the conversation
Cathode Church

A place for trans makers, coders, tinkerers and dreamers.