|Re: 6502 -- What about the 6809? TMS9900?|
Message #24 Posted by Garth Wilson on 23 Apr 2004, 11:33 p.m.,
in response to message #19 by Paul Brogger
> Does the 6809 live on in any form?
The 6800 family has progressed a lot-- not so much in speed as the 65 family has, but in microcontrollers with a lot of nice features. (See the 68HC11 family.)
> (For example, if I remember right, certain features of the 6502's indirect addressing applied only to certain registers and not to others, right?)
<end quote> Some of the indexed indirect and indirect indexed instructions' operands are only one byte, meaning that first you refer to an address in zero page. Others like JMP(addr) (an indirect) or JMP(addr,X) (an indexed indirect) have 16-bit operands to start with any address in the 64KB space. The 6502's zero page (addresses 0000-00FF) are basically 256 processor registers, and they offer more flexibility and can can be accessed faster than the rest of the memory map. It's part of what gives the 6502 a surprisingly high power-to-complexity ratio.
The to follow the 6800 to 6809 progression in the 65 family would be kind of like going from the 6502 to the 65816. Actually the '816 takes the progression a lot further. It has 16-bit accumulator and index registers, and a 16MB address space. The "zero page" is now called "direct page" because it can be moved around to anywhere in the first 64KB of address space instead of being stuck in page 0. The '816 has extra instructions and features that make it more suitable for relocatable code, multitasking, and other things, without really giving up the simplicity.
The TI 8-bitters were all very poor in performance as far as I know. Even the TMS370 has a minimum interrupt latency that's almost ten times as slow as the 65c02's-- 3 microseconds for the TMS370, versus 0.35 for the 65c02, both running at 20MHz. (Actually, with a certain trick, the 65c02's minimum is only a little over .05us (50 nanoseconds) at 20MHz, but you can't do that for every interrupt.
The 1802 someone else referred to had certain advantages in its time which made it a good choice in some spacecraft and low-power applications. It was one of the first CMOS processors and rad-hard. It had enough onboard registers that for some tiny applications you didn't have to use any external RAM. It was slow as molasses in January though.
Edited: 24 Apr 2004, 2:08 a.m.