Talk:8-bit computing

applications/gadgets

What applications/gadgets still have 8bit processors in this era of 64 bit processors?

Controllers, for example, [1] because such tasks rarely need much CPU-power or memory. Low power-consumption, low design complexity and device size are more important in such cases. In space travel, they stick to more robust and simple hardware also because it is less sensitive to radioactivity. Therefore, you probably won't see any common 64-bit CPUs in space anytime soon. --217.87.87.117 (talk) 19:46, 22 November 2007 (UTC)

Most applications/gadgets still use 8 bit processors. For example, all of the following use 8 bit processors:

• mouse (computing) -- some examples at optical mouse
• computer keyboard#Control processor
• the touch-sensitive wheel in some iPods -- the 8-bit Cypress PSoC [2]
• microwave oven
• the key fob of a remote keyless system

Should I put these usage examples into the article, similar to the way the propane article lists a few usage examples? --68.0.124.33 (talk) 21:40, 26 June 2009 (UTC)

Why 8 bits?

Seems to me, the section fails to answer the question it raises. thus, should either be expanded, moved elsewhere or edited out. cheers. 132.69.253.48 (talk) 02:57, 18 July 2009 (UTC)

IA-8

Why does this term link to this page? Ive not seen one thing on this page that suggests any sort of confusion justifying its inclusion. Simply put: How do you confuse IA-8 with 8-bit (aside from the dash and the number 8.) It seems like a mistake. 74.128.56.194 (talk) 19:21, 28 June 2011 (UTC)

IA-8 was originally an article talking about the 8-bit Intel 8008, Intel 8080, or Intel 8085 processors, by analogy to IA-32 being a term referring to the instruction set of the 32-bit Intel x86 processors.

It was then made into a redirect here; the discussion for that can be found at Wikipedia:Articles for deletion/IA-8, but no rationale was given for the decision. (I would have voted Delete.)

It now redirects to IA8, which is a disambiguation page that points to two Iowa-related pages ("IA" being the US two-letter code for the state of Iowa), and also points here because, presumably, IA-8 redirects to the disambiguation page and sounds like a name for Intel's 8-bit architecture (along the lines of "IA-32" and "IA-64"), speaking of it as "an unofficial backronym for Intel's 8-bit processor architecture", with "8-bit" being linked.

Whether 1) it's a backronym used anywhere other than the imagination of the person who put that entry on that page and 2) if it is (so that it shouldn't just be removed), whether it should link to "8-bit", Intel, or to one of the Intel processor pages, are separate matters. Guy Harris (talk) 02:00, 5 November 2021 (UTC)

I found nothing to indicate that anybody uses "IA-8" or "IA8" to refer to Intel's 8-bit instruction sets, so I removed the link from IA8. Guy Harris (talk) 04:48, 5 November 2021 (UTC)

Reversible Logic

Many 8-bit shift registers use reversible logic for energy efficiency. (https://www.researchgate.net/profile/Rohini-Hongal/publication/329317259_Design_and_Implementation_of_8_Bit_Shift_Register_using_Reversible_Logic/links/5ca312fba6fdccab2f67d3a1/Design-and-Implementation-of-8-Bit-Shift-Register-using-Reversible-Logic.pdf) and (https://ieeexplore.ieee.org/document/5166832). Are there any thoughts on including that in the article? CessnaMan1989 (talk) 00:41, 5 November 2021 (UTC)

The first of those papers is about the general concept of using reversible logic for shift registers; the one they happened to design was an 8-bit register, but they talk about other bit widths as well. Is there some reason why this is relevant to 8-bit computing rather than to, say, shift registers (or barrel shifters, if that's what that first circuit really was)? Guy Harris (talk) 01:28, 5 November 2021 (UTC)

Hmmm...I think it allows for more situations where 8-bit computing becomes more efficient that other alternatives, especially when I think about parallel processing. However, I readily admit those applications are not in the source, and I have to honestly concede that including those applications would be in violation of the prohibition on original research. Shift registers are the main reason I mentioned it. CessnaMan1989 (talk) 01:56, 5 November 2021 (UTC)

But would that first design still "[allow] for more situations where 8-bit computing becomes more efficient that other alternatives" if they produced a 16-bit or 32-bit or 64-bit reversible-logic shift register? Would the 8-bit ones be more efficient than the larger ones? And would using 8-bit shift registers require more shift registers than using the larger ones? If so, would it still be more efficient? Guy Harris (talk) 04:52, 5 November 2021 (UTC)

(BTW, the design shown in the first paper looks more like a barrel shifter than a shift register to me - it has an 8-bit data input and a 3-bit "shift amount" input.) Guy Harris (talk) 05:00, 5 November 2021 (UTC)

@Guy Harris: It doesn't look like it's barrel shifting to me because the logic looks sequential and the bits are being shifted "place-by-place" even though the "shift amount" input is 3 bits. If it were barrel shifting, wouldn't the bits at the end "circle back" to the beginning of the register? Fig. 14 looks sequential to me. I think it's a multi-bit shift register. CessnaMan1989 (talk) 18:33, 8 November 2021 (UTC)

@CessnaMan1989: So where do the circuits in the first paper depend on the sequence of past inputs, that being what sequential logic says in the first paragraph is what makes a device "sequential logic"?

And "barrel shifter" doesn't mean "device that does circular shifting", so there's no reason why the bits would have to circle back in order for the device to be a barrel shifter. It means "device that takes a word and a bit count and shifts the word by that number of bits using combinatorial logic", and multiplexors look like combinatorial logic - as the barrel shifter article says, "A barrel shifter is often implemented as a cascade of parallel 2×1 multiplexers." Figures 9 and 10 of the first paper look like cascades of parallel 2x1 multiplexers to me. Guy Harris (talk) 19:07, 8 November 2021 (UTC)

@Guy Harris: Barrel shifters are, by definition, bit-rotating. Bit-rotation is what distinguishes a barrel shifter from a shift-register. In a shift-register, the last bit leaves the register and will be lost unless there is some sort of serial output. The logic at Fig. 14 looks to me like the last bit is leaving the register, which is why I think Fig. 14 shows sequential logic. Now, since this logic is reversible, perhaps this kind of bidirectional shift-register could be implemented to function like a barrel shifter in many situations, given how short intermediate-steps can be, but it isn't working like a barrel shifter at Fig. 14. Just because barrel shifters are often implemented with 2x1 multiplexers doesn't mean that shift registers aren't also ever implemented with 2x1 multiplexers. Look at the diagram and values at Fig. 14. CessnaMan1989 (talk) 19:50, 8 November 2021 (UTC)

(Speaking of shifting, a shift was missing above. :-)) @CessnaMan1989: "Barrel shifters are, by definition, bit-rotating. Bit-rotation is what distinguishes a barrel shifter from a shift-register." It appears that the term "barrel shifter" is used for two different purposes:

• the first external link on barrel shifter is to this page, which has a circuit that does not appear to do circular shifts;
• the second external link is to this page, which appears to consider circular shifting to be an essential characteristic of a barrel shifter. It also seems to have a "barrel shifter" as a combination of multiplexors and, in some cases, what appear to be latches, hence the reference to "clock cycles".

So there are at least two different characteristics of a shifting device:

• Is it combinatorial, so that you feed it an input to shift and, if the shift amount isn't constant, a shift amount, and, some number of gate delays later, the shifted output appears, without storing any data, or is it sequential, so that it stores the input and shifts it, with the circuits that store the bits wired up so that the bits are transferred from one to the other to do the shifting?
• Is the shift that it does circular?

I tend to think of a "shift register" as a sequential device, so that a device that doesn't store data isn't a shift register; shift register appears to agree. There appear to be others who use the term differently; I'm not sure whether the Xilinx application note speaks of "shift registers" because some of the circuits described there happen to include an output register made of what I'm guessing are flip-flops, into which the output of the shifting circuit are stored, but that's different from the sort of shift registers depicted on the shift register page, where the flip-flops participate in the shifting process rather than just storing the output of the shifting process done by multiplexors. I don't think of a shifting circuit plus a register as being a "shift register". Guy Harris (talk) 20:56, 8 November 2021 (UTC)

@Guy Harris: You know, it's interesting because I've only ever seen the definition from the second source for "barrel shifter", and I work in the field. A shift register is definitely a sequential device; it has to be sequential. I always have thought of a shifting circuit plus a register as being a "shift register." To be fair, I learned throughout school and work on Xilinx. Pretty much everything I learned about FPGA's, I learned with Xilinx. How many engineering faux pas has this semantic confusion caused? CessnaMan1989 (talk) 17:39, 9 November 2021 (UTC)

Quantum Computing

Does anybody know if 8-Qubit registers are used in Quantum Computing? CessnaMan1989 (talk) 00:43, 5 November 2021 (UTC)

FPGA4Student Blacklist?

Is the blocking of FPGA4Student, a site with relevant information on 8-bit computing, intentional? Why??? I thought it was a useful example of 8-bit computing for training. CessnaMan1989 (talk) 01:55, 10 November 2021 (UTC)