sutin@astrovax.UUCP (02/10/87)
Where does one look for information about the various microprocessors, such
as comparisons, etc, from a hardware standpoint? The old workhorse Z-80 is
no longer good enough, either in word size or address space. I was
considering the TI family of 99000 chips because they have a single chip
version, a fast version(25Mhz), come on a 40 pin dip, address 256k without
memory mapping, and have a sixteen bit word. I don't know anything about
the instruction set. I loved the LSI-11 instruction set, but DEC won't
give the time of day about it, and the address space is too small anyway.
I think at one time Byte magazine contained information of any sort, but
those days are now gone. There must be some other place, especially with
all the recent RISC micros that must be out now that have simple instruction
sets, lots of registers, large word sizes, and big busses: all things that
make assembling programs easier. Boy, is this erratic. Am I making any
sense?
Brian Sutin
Department of Astrophysical Sciences/Princeton University
{ princeton, ... }!astrovax!sutin
sutin@astrovax.princeton.edu
sutin%astrovax.princeton.edu@puicgl.bitnetgrr@cbmvax.UUCP (02/11/87)
In article <840@astrovax.UUCP> sutin@astrovax.UUCP (Brian Sutin) writes: >Where does one look for information about the various microprocessors, such >as comparisons, etc, from a hardware standpoint? Various electronics industry trade magazines publish occasional comparisons and comparitive articles. Unfortunatly, this stuff tends to die off by the time the devices are in common use. Check if you university library has back issues of such magazines. There is also a moderatly useful, but out of date Osbourne book on 16-bit microprocessors. > I was >considering the TI family of 99000 chips because they have a single chip >version, a fast version(25Mhz), come on a 40 pin dip, address 256k without >memory mapping, and have a sixteen bit word. I don't know anything about >the instruction set. I've worked with the 990* chips. Simply put, not worth the effort. They are an outgrowth of TI minicomputer systems that were initially designed to show what nifty things you could do with bit-slice MSI TTL devices. While they have some interesting features, they are, for all practical purposes, obsolete. As far as I know, only TI-propriatary software is readily available and there is no unix port... If you want to play, look at NS32000, MC680X0, and Intel 8086 chips. If you want to probe the future, look at the Fairchild Clipper, MIPS chips and the Intel 386 stuff. Intel development can be PC based, National has some nice (almost affordable) single board systems. > I loved the LSI-11 instruction set, but DEC won't >give the time of day about it, and the address space is too small anyway. The 68000 and 32000 both carry the PDP-11 torch farther down the road. >Brian Sutin Department of Astrophysical Sciences/Princeton University -- George Robbins - now working for, uucp: {ihnp4|seismo|rutgers}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@seismo.css.GOV Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
daveh@cbmvax.UUCP (02/11/87)
> > If you want to play, look at NS32000, MC680X0, and Intel 8086 chips. If > you want to probe the future, look at the Fairchild Clipper, MIPS chips > and the Intel 386 stuff. Intel development can be PC based, National > has some nice (almost affordable) single board systems. If you're interested in modern architectures, look into the INMOS Transputer series. These are 16 and 32 bit processors that employ an number of RISC architectural ideas. And they're designed to allow simple implementation of multiprocessor systems. Each Transputer processor has several high speed (10 Mbits/second, DMAed) serial links for inter-processor communication. The software design of the chip can take advantage of this link. You basically design separate processes to run on a Transputer, in a way similar to what you might do on a 68000. When following the Transputer process model, though, individual processes can run on individual Transputers or all on one, all transparently. Some other aspects of the system are that all Transputers run off a single TTL level 5MHz clock. The faster clock speeds are internally generated, as are memory control signals, to allow operation as fast as 10 MIPS (some of the new ones are even faster). There's also memory (2K-4K typically) built into each Transputer. The Transputer isn't as mainstream as a 680x0, 80x86, or 32xxx, and the cost is still in the high range, though as a single chip part its probably priced comparably to the 80386 or the Fairchild Clipper. Most any system, UNIX, VAX/VMS, even IBM PC can be used for development, and development boards are available as either stand-alone or IBM PC resident. I've got no connections to INMOS, other than an appreciation of the Transputer. -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Dave Haynie {caip,ihnp4,allegra,seismo}!cbmvax!daveh "You can keep my things, they've come to take me home" -Peter Gabriel ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
ian@loral.UUCP (02/12/87)
There is a good article on the new high speed super-microprocessors
in the Feb. 1, 1987 issue of Computer Design, by Ken Marrin, pgs. 57-76.
This article compares some of the features of the microprocessors from
Motorola, National Semiconductor, Intel, MIPS and AT&T.
Another excellent source of information is the IEEE Micro magazine,
published by the IEEE Computer Society. Various issues have detailed
articles on the National, Motorola and Intel 386 chips.
If you are looking for information on single chip computers you will have
to survey the vendors yourself. The magazines and journals usually only
cover the "sexy" new microprocessors.
Ian Kaplan
Loral Dataflow Group
Loral Instrumentation
(619) 560-5888 x4812
USENET: {ucbvax,decvax,ihnp4}!sdcsvax!loral!ian
ARPA: sdcc6!loral!ian@UCSD
USPS: 8401 Aero Dr. San Diego, CA 92123andy@batcomputer.UUCP (02/12/87)
daveh@cbmvax.cbm.UUCP (Dave Haynie) writes: >If you're interested in modern architectures, look into the INMOS Transputer >series. These are 16 and 32 bit processors that employ an number of >RISC architectural ideas. ... >I've got no connections to INMOS, other than an appreciation of the >Transputer. Its a simply elegant processor at the instruction set level. Don't look to a Transputer for gobs of registers, though. It's only got an instruction pointer, a workspace pointer (essentially a stack pointer -- local variables are offsets from the workspace pointer), and a stack of registers that are 3 deep known oddly enough as aReg, bReg, and cReg. Because they are maintained as a stack, you can't randomly address them. eg, addition: ldc 1 ldc 2 add leaves 3 in the A register. It wins hands down (so far :^) as a hardware multi-tasker. It could use a smaller time-slice, though. Andy ps: It does stash a number of other "processor temporaries" and processor registers in on-chip RAM.... -- Andy Pfiffer andy@tcgould.tn.cornell.edu Cornell Theory Center / Cornell U. cornell!batcomputer!andy Home of the first usable T-Series (607) 255-8686 "...that's the way a Transputer works, right?" Systems Group