ALU - Trouble in starting

[ollieisamuppet]

I managed to complete all the previous chips without too much trouble, but I can't seem to make a start on the ALU.

I've read a few posts on this forum from people with similar problems, only I've already looked at what was suggested to them, and I'm still clueless.

For example, I (eventually) noticed that a multiplexor might give me some sort of "if" functionality in logic, that might perhaps be useful in branching for zero/negative. I thought I'd work through the skeletal .hdl for the ALU and try to figure it out stage by stage, but I can't even see how I would make the multiplexor work in this context.

I know that for x, I need to get it in, zero and/or negate it based on the zx/nx bits, and then branch similarly depending on the value of "f". From a high-level, I think I can define what it needs to do, but I'm really struggling to understand how to achieve that spec with logic.

In fact, I'm so confused that I'm even struggling to explain how confused I am
I know it's not intended to be a five-minute job, but I feel like I could stare at it for 100 years and not get any further.

Also, I take it I'm correct in thinking that the nx/zx (plus y variant) bits are independent from the other bits, so their logic can be completely separate?

I'd really appreciate it if someone could perhaps put me on the right track. I have a tendency to over-think problems and miss glaringly obvious points, so I'm willing to admit I may have done that here, also!

Thanks

[cadet1620]

Don't think about if / then / else branching structures.  Branching of that sort is a software concept.

In hardware, everything that you might need is computed, and a mux can choose the required result.

For example the code
    if (flag)
        y = f(x)
    else
        y = g(x)

might become the HDL (assuming F and G chips that compute f() and g() )
    F(in=x, out=fOut);
    G(in=x, out=gOut);
    Mux(a=gOut, b=fOut, out=y);

Also, I take it I'm correct in thinking that the nx/zx (plus y variant) bits are independent from the other bits, so their logic can be completely separate?

Yes, two identical circuits.  In my ALU I call the outputs from them xx and yy. xx and yy then go into the function computations.

--Mark


   

[ybakos]

The key thing to understand about the ALU is that it is able to accomplish what it does (the various operations) through a combination of other operations. Take a look at the ALU Worksheet.

[ollieisamuppet]

Thank you both.

With Mark's advice, I was finally able to look at it the right way around. I've just finished the ALU (minus status flags) and it passed the no-stat test, so I'm very pleased