it on every reference. Do you disallow it? Well, then do you want unaligned reads to work on regular memory? If so, now you have some memory that is magic unalign-able and some that isn't. Is that a separate failure mode? What do you do about it?
-
-
Replying to @rygorous @cmuratori
Do you make the memory word-adressable only? Machines used to be like that, and it turns out that manipulating anything small with that is a giant pain in the ass and requires massive amounts of code.
2 replies 0 retweets 0 likes -
Replying to @rygorous
This is going to be very hard to reference... I guess I could cut and paste :)
1 reply 0 retweets 0 likes -
Replying to @cmuratori @rygorous
But unfortunately I don't think we're talking about the same thing? It sounds like you're talking about actually having individual memory operands for every place you would have a register.
1 reply 0 retweets 0 likes -
Replying to @cmuratori @rygorous
I am talking about having memory-backed register file. It still may be that the things you're mentioning are a problem, but I'm not sure I understand exactly why they would be a problem from the (Twitter-limited) explanation.
2 replies 0 retweets 0 likes -
Replying to @cmuratori @rygorous
Another way to say it would be, I am asking why we don't memory map registers the way we memory map files, for the same reasons (so the _processor_ knows where they would go when they need to be swapped out).
2 replies 0 retweets 0 likes -
Replying to @cmuratori @rygorous
Maybe you can't _actually_ ever make that work in any kind of efficient way that isn't worse than just using explicit save and restore on thread switches, but I'm not sure I get why.
1 reply 0 retweets 0 likes -
Replying to @cmuratori @rygorous
In much the same way that "self modifying code" is no longer a thing, "writing to the register memory without calling it a register" could also not be a thing, right - we do that sort of thing all the time.
3 replies 0 retweets 0 likes -
-
Replying to @cmuratori @rygorous
So I'm just looking for a more in depth understanding of whether it's really the right idea to have n "phantom" addresses that don't actually map to anything as the core way that instruction streams are written...
1 reply 0 retweets 0 likes
... instead of some other potential scheme that still reflects the fact that these are special, just not necessarily "phantom".
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.