/tech/ - Technology★

Doesn't look like the AMD am29k.

>>1004333 (OP)

Recently finished implementing one on an FPGA (with accurate cycle lengths) the state machine for the control unit was a lot of fun build. The instruction set is very small, but it has a complex set of addressing modes. Perhaps I'll do an AoC challenge in 6502.

>>1004333 (OP)

>Let's discuss

Do you have anything to add? Linking to a wikipedia article makes for a very shitty thread.

Haven't touched the family outside of the nes/snes/2600, but you have a couple sbc options nowadays:

https://wdc65xx.com/boards/w65c02sxb-engineering-development-system/

https://wdc65xx.com/boards/w65c816sxb-engineering-development-system/

>>1004767

definitely not the 8051

>>1004767

OY VEY

>>1004767

>the greatest 8051

gnoo

>not MC680XX

en.wikipedia.org/wiki/Motorola_68000

>>1005119

>>1005095

>>1005078

Plebs that don't know how to code efficiently detected

Although that motorola is pretty cute I have to say

>>1004333 (OP)

> the greatest processor of all time

Isn't it literally possible to fry the chip through software by accessing certain undocumented registers through a couple POKEs?

>>1005361

The only examples I could find were the Commodore PET being able to fry its CRT and the 1571 disk drive being able to be fucked by a wrong command. Both use the 6502 in some form, but I couldn't find anything on the CPU being able to self destruct.

In the other hand, system like the Xbox 360 and up have special circuits built into the chips that allow them to brick themselves or block certain firmware from loading. Efuses are a bitch.

>>1005361

>>1005365

These bugs in pet and disk drive are in their kernel (or kernal as Commodore called it), not in the cpu. The cpu has a couple of interesting bugs though, some of which are even useful.

Indirect jump doesn't read next address byte from another page correctly, for example

jmp ($10ff) should read the address from 10ff 1100 but it doesn't add carry to MSB so it's 10ff 1000

inc will store the value twice. First the original value then the incremented one like this

read memory

inc

store original

store incremented memory

from programming perspective it doesn't make any difference but MMIO will see 2 writes and it might be used for something.

and of course the illegal opcodes... Some of them are so useful I use them all the time, for example lax and sax. Load to register a and x at the same time and store both a and x to memory (Anding a x)

sbx to subtract from x while ignoring carry flag

A processor is evaluated in terms of practicality of instructions, number of accumulator records, records of operations, indexes records and CPU usage cycles to perform an operation. For example: in the time of 6502 was the z80 as a direct competitor. While the 6502 was running at 1MHZ, the Z80 was running at 4MHZ. But the Z80 needed 4 clock cycles to read an instruction the 6502 did in 1 cycle. Almost like a RISK but in the 70s. The 6502 allowed the processes to be carried out by the perispherical processors. the Z80 had to do them all! That made the 6502 somehow more powerful than the Z80. The same thing always depends on the ability of the programmer to take advantage of the power of the processor. Today we have microprocessors that run to thousands of MHZs but even so, the programs are STUPIDLY slow. And that is not related to the predator but with the STUPID level of programmers that we have today ... I'm sorry for you assholes! (The problem is that we have more JAVA programmers than Programmers in assembler or C!) (Hold Arjona!)

>>1009290

I can't shake the feeling that its not "just" the programmers. I think x86-like arches have enabled programmers alot for the sake of compatibility and their way of programming.