Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I pointed this out above, but this is really a course for CS people that glosses over a lot of computer engineering topics that you need to, say, design your own processor. It's definitely a tremendous accomplishment, but let's keep our pants on here.


Such as? Does it skip over K-maps or something?


If you control-f, there's squabbling at length below about the fact that they don't cover k-maps, k-maps aren't important, etc.

Or you're being sarcastic. I'd hate to assume that, but someone did go through and downvote all of my posts.


I was looking for more of a list of topics it omits that are absolutely required to implement a functioning processor. A sufficiently simple little register-based RISC CPU with memory-mapped IO, no interrupts, no caches or TLBs, and so on is a functioning (if gimped) computer.


Sorry about that, you can see I'm kind of getting clobbered below.

I don't think the problem is that the end result isn't a computer (it certainly sounds like it is), but that the computer only runs in the provided simulator, and is written in a custom HDL designed to make this project relatively simple. The simulator itself ignores a bunch of complexities around timing that a commercial one (like ModelSim) would consider.

Personally I haven't done this class, but I'd be curious to know whether the students design the control unit and data path themselves. I know that was a giant pain in the ass when I did it for a gimped RISC processor (as you described).


Probably you're getting clobbered because you made lots of negative comments about something that's really cool, and because your criticism (that it falls short of a complete degree program in computer engineering) misses the point of the thing.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: