For those interested in Computer Architecture...

[ygolo]

I have an urge to do a brain dump about what I've learned about the subject over the years (which turns to be less than I should have learned)

Precise interrups, compiler theory/VLIW, CMP, SMP, benchmarking, the Tomasulo Algorithm, computer aithmetic, design for low power, pipelining and hazzards, Moore's Law, the economics of chip design, and on and on...

I don't want to bore people to tears, so I will take a poll.

Who wants to read my brain dump on the subject (will come in sections, perhaps crossposted to my blog)?

[Jack Flak]

I wouldn't have minded aside from the "Can I ask you a question?" thread. Low.

[ygolo]

Originally Posted by Jack Flak

I wouldn't have minded aside from the "Can I ask you a question?" thread. Low.

[Night]

Originally Posted by Jack Flak

I wouldn't have minded aside from the "Can I ask you a question?" thread. Low.

...who are you, again?


Anyway, ygolo, I'd be more than interested in your data. You seem to have a wealth of positive information available.

[Victor]

Originally Posted by ygolo

I have an urge to do a brain dump about what I've learned about the subject over the years (which turns to be less than I should have learned)

Precise interrups, compiler theory/VLIW, CMP, SMP, benchmarking, the Tomasulo Algorithm, computer aithmetic, design for low power, pipelining and hazzards, Moore's Law, the economics of chip design, and on and on...

I don't want to bore people to tears, so I will take a poll.

Who wants to read my brain dump on the subject (will come in sections, perhaps crossposted to my blog)?

I must admit, I am rather impressed by you, Ygolo.

And I also suspect that my iMac is more intelligent than I.

So I was rather hoping you might put me on a more level footing with this amazing machine sitting so smug in front of me.

[Nocapszy]

Originally Posted by Victor

I must admit, I am rather impressed by you, Ygolo.

And I also suspect that my iMac is more intelligent than I.

Most of our brainspace is wasted on routine maintenance of its vehicle. Steven Hawking is smarter because his body doesn't work, and doesn't require repairs.

[Samuel De Mazarin]

Originally Posted by Nocapszy

Most of our brainspace is wasted on routine maintenance of its vehicle. Steven Hawking is smarter because his body doesn't work, and doesn't require repairs.

That is utter nonsense... sorry... I try not to make absolute statements with such vigor, but that's nonsense...

Hawking was smarter and more driven than normal people well before his body started deteriorating from his disease. And Christopher Reeve didn't suddenly start writing top-notch literature or conducting experiments on spinal injuries just because he lost the use of his body.

And we tend to forget that the able-bodied Einstein was and will always be a far greater influence on the world and its thinking than Hawking... Newton, Einstein... Hawking doesn't quite make it in that highest of echelons... so am I to draw conclusions about his intelligence from his disease?

I think this is a bad induction from an n of 1.

[Samuel De Mazarin]

Originally Posted by ygolo

I have an urge to do a brain dump about what I've learned about the subject over the years (which turns to be less than I should have learned)

Precise interrups, compiler theory/VLIW, CMP, SMP, benchmarking, the Tomasulo Algorithm, computer aithmetic, design for low power, pipelining and hazzards, Moore's Law, the economics of chip design, and on and on...

I don't want to bore people to tears, so I will take a poll.

Who wants to read my brain dump on the subject (will come in sections, perhaps crossposted to my blog)?

Do it... do it now... we are hungry for knowledge.

<envision scrawny little chicks all desperately chirping and straining for food>

[ygolo]

I will start basic, since I don't want to loose people early. Forgive me if this is condescending.

Inside your Personal Computer (inside many other things, like in most modern cars, planes, guidance systems, and even some toaster ovens and washing machines) is a Microprocessor. It is often known now as an instruction processor (this term is used to distinguish this model of computing from newer models that are emerging). I will just refer to it as the processor most of the time.

To simplify things further, I will strip the processor of most of it's modern features. I will first describe a processor that is single issue, without interrupts, without pipelining, and consisting of a single core (don't worry if you don't know what these terms mean yet).

These processors run "programs." Programs are a sequences of instructions issued to processors so that they perform tasks. We can leave it at this basic level for now (on many systems they run an Operating System, which in turn sends other programs to the processor).

Here is what a program may do:

LOAD A, 0
ADD A, 3

This is a program in an abstracted version of assembly language. In this case the program loads register A with 0, then add 3 to it. So now register A has 3 in it.

There are few things I should mention before proceeding:
1) This is not the version of the program that the processor sees (it gets machine code)
2) Most systems with processors will have memory that can be written and read from, and either an Input/Output (I/O) system
or memory mapped I/O. I/O (memory mapped or otherwise) is how the processor communicates with other hardware, like graphics cards, or control circuits in a guidance system.

Use of Memory (could be memory mapped I/O) may look like this to in an assembly language (again very abstracted)

READ B, 0x10(A) # read what is at the memory address specified in register A plus an offset of 0x10 (which is hexadecimal for 16) into register B
ADD B, 3
WRITE 0x10(A), B # write the same memory location with a value three more than what was read

Still, I don't want to digress too far into the land of programming, so lets get back to what the processor gets.

Each of these "instructions" (like the LOAD, ADD, READ, WRITE, etc. above will have its own "Operation Code" (Op.Code for short). This is the binary representation of the instruction to be issued to the processor (varies from processor to processor). In "LOAD A, 0", "A" and "0" are called operands. The opcode, plus the operands constitute the full binary representation of what the processor must execute.

There are lots of complications here (micro-ops, expanding opcodes, addressing modes, privilege levels, segmentation, and many other things). But this descriptions preserves the general idea.

Please let me know if you follow this much. Because, I plan to describe how it is that the processors perform these instructions next.

[Samuel De Mazarin]

Seems straightforward....

but a few quick questions (I'm an utter end-user, so I'm totally ignorant about the inner-workings)...

How is the number of registers of a processor determined? Is it technically infinite and limited only by raw computing power?

How does a processor without 'memory mapping' work? Does it just take in input and spit out output and then 'forget' what happened? And memory mapping sort of stores earlier executed commands in the processor for later reference? I mean, what exactly is memory mapping?

I'm also assuming the very abstract nature of the codes you've described, like adding 3 to register A or some such, will be explained later...

lastly, could you give a VERY brief (I could handle more, but you're going to be working on this, so it's your time you're giving us), one post explanation of how processors relate to circuits, or vice versa?

I really appreciate this... I could look up stuff on the net, but then I wouldn't be able to get answers to questions I have at various points... will be reading as long as you post. Thanks!