I'm not talking about paradigm shifts like quantum computing, but rather just a new CPU architecture that isn't based on the x86 or other existing systems, or at least not one of the commonly used systems.
For reference, here's my current design idea, which is pretty bizarre, I think. https://medium.com/technology-past-present-and-future/design-idea-the-1802%C2%B2-afb85607cf0a
It's your time to find new gap and strong literature review to initiate theories on adding more CPU.
The way in which the registers are aggregated will be slightly different, but the primary difference is simply that it is based on the 1802 architecture. However, you say that RISC has 64K second rank registers. That might be the case, but one thing that still seems to be unique to the 1802 and its general family is that any of its scratchpad registers can be used as the program counter.
RISC uses a dedicated PC register, while the 1802 uses a pointer to a given scratchpad register. That seems to be one difference between this design and RISC.
Since one core can only execute one thread at a time, there is no point in having more than one PC register - what would it be for?
You're commenting on something which was pivotal to the way the original 1802 was designed. Perhaps it would be helpful to familiarize yourself with the original processor on which I am basing this design.
I'm afraid that comment fools not.
So, the first thing is that my real question was whether there's room for a new design. The link was just a thought on what kind of design that might be. So the question wasn't really answered here.
But as importantly, I think that if you're going to criticize my design idea, which is a perfectly acceptable thing to do and I am more than open to criticism, then you should familiarize yourself with the 1802.
By the way, RISC machines form the internals of all CISC machines.
That's not correct. While modern processors are largely CISC on RISC, there are CISC designs that are at the fundamental hardware level.
Moreover, if I had all the answers, I wouldn't be asking the RG community if there's room in the CPU world, would I?
Then focus on the question that I asked, which is whether there's room for a new CPU design, rather than criticizing this design, which is based on a CPU that you know nothing about.
Again, I am all for constructive criticism, but responding to such criticism should not have to include first explaining how an existing CPU, with a long history of use, functions.
But I will explain that the scratchpad register and PC pointer is important, because it is how one executes routines on the 1802 and its general family.
And actually, Peter Breuer, I do explain this aspect in my article on the 1802 itself. https://medium.com/technology-past-present-and-future/remembering-the-rca-180x-d0c258fa67f3
It's covered in the article. But let me ask you a question. How do you expect someone to call some routine on the 1802 without utilizing multiple scratchpad registers?
There is no stack pointer on the 1802, Peter Breuer. There is no CALL-RET pair in the 1802 instruction set. With all due respect, it is arrogant to try to criticize someone's idea of extending an existing technology when you are unfamiliar with the original technology.
- Badges
- Science topic
- Similar topics
- Computer Science