Generic CPU discussion

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6585
Credit: 310894417
RAC: 96934

Thanks for sharing that,

Thanks for sharing that, that's a very good article. What it demonstrates is that for a given task there is an irreducuble algorithmic complexity, with the choice to put that into the instruction set architecture and/or the implementing hardware (compilation aside). But then if you change the task .....

I never really understood RISC vs CISC when it first came up - yes, I am that old - so now I'm pleased to hear it now doesn't matter. A non-question for a non-problem. But it is right in saying that the CPU will nearly always beat DRAM response, the speed of light being what it is : you can consider circuitry as fancy waveguides if you like and proximal beats distal nearly always. Now if only we had an ASIC for every eventuality. :-)

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

mikey
mikey
Joined: 22 Jan 05
Posts: 12639
Credit: 1839025349
RAC: 5518

Mike Hewson wrote: Thanks

Mike Hewson wrote:

Thanks for sharing that, that's a very good article. What it demonstrates is that for a given task there is an irreducuble algorithmic complexity, with the choice to put that into the instruction set architecture and/or the implementing hardware (compilation aside). But then if you change the task .....

I never really understood RISC vs CISC when it first came up - yes, I am that old - so now I'm pleased to hear it now doesn't matter. A non-question for a non-problem. But it is right in saying that the CPU will nearly always beat DRAM response, the speed of light being what it is : you can consider circuitry as fancy waveguides if you like and proximal beats distal nearly always. Now if only we had an ASIC for every eventuality. :-)

Cheers, Mike.

What would be cool is an on the fly re-programmable ASIC, then each need could be addressed by the programmers and handled by the device. Probably too expensive to make and then repair when it breaks but cool anyway.

Ian&Steve C.
Ian&Steve C.
Joined: 19 Jan 20
Posts: 3927
Credit: 45668302642
RAC: 64027086

mikey wrote: Mike Hewson

mikey wrote:

Mike Hewson wrote:

Thanks for sharing that, that's a very good article. What it demonstrates is that for a given task there is an irreducuble algorithmic complexity, with the choice to put that into the instruction set architecture and/or the implementing hardware (compilation aside). But then if you change the task .....

I never really understood RISC vs CISC when it first came up - yes, I am that old - so now I'm pleased to hear it now doesn't matter. A non-question for a non-problem. But it is right in saying that the CPU will nearly always beat DRAM response, the speed of light being what it is : you can consider circuitry as fancy waveguides if you like and proximal beats distal nearly always. Now if only we had an ASIC for every eventuality. :-)

Cheers, Mike.

What would be cool is an on the fly re-programmable ASIC, then each need could be addressed by the programmers and handled by the device. Probably too expensive to make and then repair when it breaks but cool anyway.



1. if it was reprogrammable/reconfigurable, it would no longer be an 'ASIC'
2. what you're describing is exactly what a FPGA is.

_________________________________________________________________________

mikey
mikey
Joined: 22 Jan 05
Posts: 12639
Credit: 1839025349
RAC: 5518

Ian&Steve C. wrote: mikey

Ian&Steve C. wrote:

mikey wrote:

Mike Hewson wrote:

Thanks for sharing that, that's a very good article. What it demonstrates is that for a given task there is an irreducuble algorithmic complexity, with the choice to put that into the instruction set architecture and/or the implementing hardware (compilation aside). But then if you change the task .....

I never really understood RISC vs CISC when it first came up - yes, I am that old - so now I'm pleased to hear it now doesn't matter. A non-question for a non-problem. But it is right in saying that the CPU will nearly always beat DRAM response, the speed of light being what it is : you can consider circuitry as fancy waveguides if you like and proximal beats distal nearly always. Now if only we had an ASIC for every eventuality. :-)

Cheers, Mike.

What would be cool is an on the fly re-programmable ASIC, then each need could be addressed by the programmers and handled by the device. Probably too expensive to make and then repair when it breaks but cool anyway.



1. if it was reprogrammable/reconfigurable, it would no longer be an 'ASIC'
2. what you're describing is exactly what a FPGA is.

I did not know that, interesting.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.