sk2f+@andrew.cmu.edu (Seth D. Kadesh) (11/29/90)
I've been thinking about this for quite some time now. So..... Lots of people have been talking about fixing the GS. Why should we stop there?! Why do we need to build a GS clone?! We know what philosophy this computer should embody - the things we like about the Apple II family, and don't like about the others. I'm talking real big here, but let's build for the future! Let's build the ultimate (I wanted to say next, but I can't :-) personal computer. I had in mind something powerful, with the home environment in mind, but there's no reason why this work can't be applied to other environments. I'm being vague, I know. I don't have enough knowledge (I'm between a sophmore and a junior in CS here at Carnegie Mellon, with some physics also). But I'm not just talking faster processors, better memory access, or stuff like that. I'm thinking about more radical changes..... 1. HDTV. It will be here soon. What are its specifications? (I think it's better than most current computer resolutions, but I'm not sure). But I don't just want to have the resolution of HDTV, I want to interface with it. I'm not quite sure yet what I mean by interface, but let your imagination roam. Take your favorite SF stories..... Other aspects of the computer would follow, but i have a class to get to, and a problem set to do. I'm not quite sure what I'm proposing here, but all interested parties are quite welcome to expound on this.... I'll post more thoughts as I consolidate them..... -seth - tHe mAd ScienTisT, and other carnations ------------------------ sk2f@andrew.cmu.edu | everything seems so easy this way but I'm going under tmSatCMU@DRYCAS.BITNET | fast, I'm slipping away, am I so crazy - Marillion
$CSD211@LSUVM.BITNET (Mark Orr) (11/29/90)
|I've been thinking about this for quite some time now. So..... |Lots of people have been talking about fixing the GS. Why should we stop |there?! Why do we need to build a GS clone?! We know what philosophy this |computer should embody - the things we like about the Apple II family, and |don't like about the others. I'm talking real big here, but let's build for |the future! Let's build the ultimate (I wanted to say next, but I can't :-) |ersonal computer. I had in mind something powerful, with the home environment |in mind, but there's no reason why this work can't be applied to other |environments. I think that building an Apple II for the future is a great idea. Any computer we designed would have to conform to what we have come to expect from II's (i.e. great expansion possibilities, built in BASIC, a ROM based system monitor and in general a great deal of flexibility) I would direct you to Todd Whitesel Apple IIf paper as a start. But take it with a grain of salt. It seem to be written from the perspective of "what is a realistic interpolation of what Apple could do with the II?" It contains many exceptional ideas but, as I have said, I don't think it goes far enough (i.e. it would please Apple II users to no end...but how would it attract new users) Certainly I think that we need a new processor. The 6502/65816 has clearly had its day. Some people (on this net) are waiting for the ASIC 65816 clone which may run at up to 25 MHz. With a cache, and without the IIgs's bottlenecks, this could provide a dramatic speedup for the II. But let's take a look at some solid facts: (1) The ASIC does no yet exist. And with the recent events involving the Duet (i.e. hardware and software supporters jumping ship),it may never exist. (2) With a cache, the ASIC will, at best, turn 2 MIPS...probably lower. (3) The ASIC, like all other chips - the manufacturer wanting to recoup the R&D, will not be cheap for a while. (4) Thirty-two bit processors capable of much higher performance than the ASIC are available NOW. I have mentioned the AMD 29000 RISC, the VLSI VL86C020 (the ARM - another RISC). Both these are powerful, inexpensive processors (used mainly in embedded applications). Let me add the Inmos T400, the highest performance to cost ratio chip available. The T400 can do 20 MIPS and costs about $20 (of course it does require a cache, so the actual cost is higher. But remember, the ASIC would need a cache too to achieve maximum performance. the T400 is also an embedded RISC processor. (5) A thirty-two bit RISC chip has about 1000 times the marketing value of a 65816 (which enough people know has an 8-bit data bus with a sixteen bit architecture). Look at the trade papers...everyone is saying that the '286 is dead. In what state of conciousness is the 65816 (6) The processors I have mentioned have wider address buses than the 65816. Not only does this mean that more memory may be added (right this is not a factor, since you can get far more memory than you need for the IIgs), but one can map more peripherals to that address space. What say we choose the AMD for our main processor and use the new FutureBus+ (IEEE 896.1), now we have an industrial strength bus that developers will flock to, and it's mapped in its full glory on the address bus. (BTW, FutureBus+ not only has a smokin' name (market wise), but can transfer at over 100 Mb/s) Add to this Super-VGA level graphics (which we can animate like lightning), and an Ensoniq (or two :-) ). We can use a 65816 on the motherboard to handle the I/O (and emulate the gs). We use UNIX as our operating system and X-Windows as our windowing environment. Add to this Direct Slots, which allow full access to the CPUs, video and sound hardware (for that extra measure of cleverness). Like Apple IIs of the past (and taking into account the mistakes of Steve Jobs) we don't want to build everyting in, even if we do have all those slots. I/O should include serial ports, MIDI, SCSI, and inputs for keyboard, mouse and two joystick (we will lose ADB naturally) A full implementation of MIDI should be on the board because the MIDI adapters that attach to the gs's serial ports are not a full MIDI set. A full MIDI set would include MIDI in, MIDI out, MIDI thru, and fourth port (which I cannot remember). There are compelling reasons for NOT including SCSI, like leaving room for third parties like CVTech to build a high speed SCSI (or SCSI-2) card. But since were using UNIX, we'll probably need it. GS compatibility would be maintained in a manner similar to the UNIX //e emulators, except with hardware support...this way we needn't copy or rebuild the GS ROMs...just provide a hardware interface for it. The RISC would still be used to handle video and sound and to run the software-interface. Todd Whitesel expressed the valid concern that, with the addition of the RISC, the GS side would become less important. In time, I hope it would. It would take time for applications to be developed. Users would still use GS software to start. |1. HDTV. It will be here soon. What are its specifications? (I think it's | better than most current computer resolutions, but I'm not sure). But I | don't just want to have the resolution of HDTV, I want to interface with it. | I'm not quite sure yet what I mean by interface, but let your imagination | roam. Take your favorite SF stories..... Problem...HDTV is yet to be defined in the USA (as far as i've heard). Besides VGA and Super-VGA level resolutions are higher than HDTV (1148x524 I think). Video overlay cards could be built for FutureBus+ by third parties. We don't want to build everything in. Just what is considered a competitive standard (i.e. resolution, sound, I/O) plus the high performance bus and CPU. |-seth | - tHe mAd ScienTisT, and other carnations |------------------------ |sk2f@andrew.cmu.edu | everything seems so easy this way but I'm going under |tmSatCMU@DRYCAS.BITNET | fast, I'm slipping away, am I so crazy - Marillion ---------------------------------- | MARK A. ORR | | $CSD211 @ LSUVM.SNCC.LSU.EDU | | @ LSUVM.BITNET | ----------------------------------
toddpw@nntp-server.caltech.edu (Todd P. Whitesel) (12/01/90)
$CSD211@LSUVM.BITNET (Mark Orr) writes: >Apple IIf paper as a start. But take it with a grain of salt. It seem to be >written from the perspective of "what is a realistic interpolation of what >Apple could do with the II?" It contains many exceptional ideas but, as I have >said, I don't think it goes far enough (i.e. it would please Apple II users to >no end...but how would it attract new users) Damn right it was a realistic extrapolation. I originally wrote the paper and mailed the second version to Sculley. He forwarded it to Birss, who wrote me a nice letter back that said essentially nothing (sources tell me he read it thoroughly before passing it on). The //f paper's primary purpose was to attempt to convince Apple that they could do it. So now Apple's working with ARM to make a new generation of notebook machines. My idea was limited, but that was so it could be realistic enough to make. Your idea is superior as an idea, but where are you going to get the funding? Todd Whitesel toddpw @ tybalt.caltech.edu
weird@netcom.UUCP (Richard Ney) (12/01/90)
the guys would Kill me if this got out... Hear from ya soon! :) BAAUGfest - Information Plan BAAUGfest - [WHO] An organization of Apple ][ users, primarily from Northern California, are concerned about the severe lack of development and exposure of the Apple ][ computer. BAAUGFest - [WHAT] This organization of end users of the Apple ][ Computer, have resolved that the machine is well deserving of its own show. The Fest will consist of products that have been or are being, developed for the Apple ][. BAAUGFest - [WHEN] The dates have been scheduled for Mid-February. This time was selected so that it would not interfere with the schedule of the southern California AppleFest in December and the East Coast Fest in the spring. BAAUGFest - [WHY] With the reality of "AppleFest" being moved to southern California so that it could be held in conjunction with the Macintosh show, it may be assumed that: a... the convention developers feel that the Apple ][ cannot support a show of its own and that it is a crippled and antiquated system. b... they apparently intend to eventually phase out this show in the future. This is clearly incorrect and unacceptable. We believe that with advertising alone the dynamic capability of the Apple ][ will become apparent to the first time home computer buyer, small businesses and other interested parties. Factually, there is NOTHING that any other personal computing system can do, including Macintosh and MS-DOS based systems, that an Apple ][ cannot do. If basic design developments are made to the ][ by Apple Inc it would serve to further enhance its appeal to this available market. The keyword is exposure. BAAUGFest - [WHERE] Our Planning and Operations Sections are currently searching for a location suitable for the Fest. Suggestions are more than welcome and appreciated. Some of the qualifications that must be met are: Within the immediate San Francisco/Oakland area for ease of access. Approximately 4,000 square feet are needed for exhibitions. The cost, based on at least 4 days access, must be reasonable. COMMENTS: There will be no additional charge for keynote addresses, seminars and etcetra. We are encouraging developers to display the potency of their products and the ability of the Apple ][. Each exhibitor will have the opportunity to take the stage and demonstrate their product(s). There have been committments by many of the most prominant Apple ][ developers in the general and K-12 areas. If you or your organization are interested in additional information regarding BAAUGfest, space rental or getting on our mailing list, contact the: Bay Area Apple ][ Users Group 750 La Playa, #522 San Francisco, CA 94121-3227 Attn: Ernest Moore -- =====[ The Bay Area Apple ][ Users Group ]===== Our specific goal is the restoration of the Northern California Apple ][ classic Fest. Our purpose is to enhance the reality of the value of the Apple ][ computer in two primary sectors: <1> to the computing public that should be aware of Apple ][ potential. <2> to those involved in decision making at Apple Inc. *** We realize that there are other personal computers on the market that are popular, more powerful and faster. Yet we also realize that the Apple ][ is the perfect computer for the cross section of people that have no need for the BEST or the fastest that is available. The Apple ][gs is a far cry from being antiquated technology. With general upgrades it could once again become the standard system for use in the home, in the private business arena and in the classroom. *** We believe that the evangelistic process attempted by others has not produced adequate results. Our goal is to SHOW the value of the Apple ][ and let the computing public judge the capacity of this computer on it's merit alone. This cannot be accomplished by letters and verbal pleas to deaf ears. It must be accomplished by exposure to the potential buying market. -- Even though only being a Beta Draft, you can see what is going on here. We hope you can acquire the Duet Card for I'm working on getting the Die Hard II end Credits (since done on a Custom Program) and when people walk into our Fest, they'll REALLY BE Surprised to see such things possible on a GS! Apple II Forever! (If you can track that GS down on Comp.sys.apple2 for $500 or less, let me know... someone needs a computer but is broke... I heard there was a GS there for $500, but since I'm stuck on the Net being Blind... Thanks!) Be Seeing You. -The Prisoner (Prisoner@Darkside.com) -- I got this from a friend's account. Enjoy. (DO NOT respond to me about this. I'm just posting it!!!) (Weird aka Rick Ney) A ZZZ
jerry@polygen.uucp (Jerry Sheckel) (12/04/90)
weird@netcom.UUCP (Richard Ney) writes: > >Factually, there is NOTHING that any other personal computing system can do, >including Macintosh and MS-DOS based systems, that an Apple ][ cannot do. > This ain't so. Even if you look only at the processors, the 680x0 and 80x86 processors of the Macs and PC's have capabilities that the 6502 (and 65816) simply lacks. This isn't bashing; this isn't flaming; this is simple fact. Then there's graphics capabilities, memory capacity, disk capacity, etc. Sure, almost anything is POSSIBLE with an Apple ][, but many things are just not PRACTICAL. Running an OS with isolated 32-bit memory spaces for each process, for example, is IMPOSSIBLE, because the processor has no support for that. Running a serious CAD or number-crunching program is POSSIBLE, but definitely not PRACTICAL. And please, don't start flaming me with statements like, "Ah, but with BlowoutBrickWall Technologies' plug-in 75MHz RISC processor card, my IIgs can outrun a Cray!". The comments I made above are all about systems running in their respective native modes. > > (Weird aka Rick Ney) > -- +-------------------+----------------------+---------------------------------+ | JERRY J. SHEKHEL | POLYGEN CORPORATION | When I was young, I had to walk | | Drummers do it... | Waltham, MA USA | to school and back every day -- | | ... In rhythm! | (617) 890-2175 | 20 miles, uphill both ways. | +-------------------+----------------------+---------------------------------+ | ...! [ princeton mit-eddie bu sunne ] !polygen!jerry | | jerry@polygen.com | +----------------------------------------------------------------------------+
$CSD211@LSUVM.BITNET (Mark Orr) (12/05/90)
|weird@netcom.UUCP (Richard Ney) writes: |> |>Factually, there is NOTHING that any other personal computing system can do, |>including Macintosh and MS-DOS based systems, that an Apple ][ cannot do. |> | |This ain't so. Even if you look only at the processors, the 680x0 and 80x86 |processors of the Macs and PC's have capabilities that the 6502 (and 65816) |simply lacks. This isn't bashing; this isn't flaming; this is simple fact. Yeah...Like I said, What other personal computer in '90 uses a processor without multiply or divide instructions. Sure you can routines to do that, but that wastes memory and time. |Then there's graphics capabilities, memory capacity, disk capacity, etc. |Sure, almost anything is POSSIBLE with an Apple ][, but many things are |just not PRACTICAL. Running an OS with isolated 32-bit memory spaces for |each process, for example, is IMPOSSIBLE, because the processor has no |support for that. Running a serious CAD or number-crunching program is |POSSIBLE, but definitely not PRACTICAL. Just like people were talking about Mathematica, Stellar-7, etc. Sure the new Stellar-7 could be ported to the GS...but it wouldn't be as good as the PC version. Someone stated "a good assembly language programmer could probably make a version even faster than a '286." Why should we force software companies to use our assembly language. Mathematica wasn't done in assembly language, and it's extremely inconvenient for software companies to revert to low level languages and make machine-specific versions - that just isn't practical anymore. This is the reason that Sierra uses SCI (their own interpreter), even though Stellar-7 dosen't use it. They can make a program that can be used on all machines. You can bet that they aren't hacking out programs on PC's, Macs Amigas, GSs or Atari STs. Besides, Sierra said that if accelerators like the Transwarp (or ZipGSX) became popular enough, they'd complete SCI. But for now the GS is too slow to run it. (Remember, any programs they made would have to be compatible with both ROM01, and ROM03. No sense making two versions of SCI.) |And please, don't start flaming me with statements like, "Ah, but with |BlowoutBrickWall Technologies' plug-in 75MHz RISC processor card, my |IIgs can outrun a Cray!". The comments I made above are all about systems |running in their respective native modes. Yeah, that BlowoutBrickWall card will do it every time. ;-) There just no substitute for hardware. |+-------------------+----------------------+---------------------------------+ || JERRY J. SHEKHEL | POLYGEN CORPORATION | When I was young, I had to walk | || Drummers do it... | Waltham, MA USA | to school and back every day -- | || ... In rhythm! | (617) 890-2175 | 20 miles, uphill both ways. | |+-------------------+----------------------+---------------------------------+ || ...! [ princeton mit-eddie bu sunne ] !polygen!jerry | || jerry@polygen.com | |+----------------------------------------------------------------------------+ ---------------------------------- | MARK A. ORR | | $CSD211 @ LSUVM.SNCC.LSU.EDU | | @ LSUVM.BITNET | ----------------------------------
sk2f+@andrew.cmu.edu (Seth D. Kadesh) (12/05/90)
You guys all kind of missed the original intent of my post, though Mark Orr was kind of close. I'll try again soon, after I write this compiler (in Scheme) for my CS class. Obligatory piece of news: Apple Computer bought a 30% share of ARM. They're talking about cheap, fast, small, notebook computers. I'm not sure of the exact share percentage. The article is in the latest Macweek. -seth - tHe mAd ScienTisT, and other carnations ------------------------ sk2f@andrew.cmu.edu | everything seems so easy this way but I'm going under tmSatCMU@DRYCAS.BITNET | fast, I'm slipping away, am I so crazy - Marillion
m.tiernan@pro-angmar.UUCP (Michael Tiernan) (12/06/90)
In-Reply-To: message from $CSD211@LSUVM.BITNET Mr Shekhel is dead right, the basic problem with the Apple II line is/has-been/will be that it lacks a few required instructions at the processor level to allow it to be a competative force. This has been the problem since way back in the days of the Apple ][+ when they decided that no one will ever need an inerrupt line so they took them out of the card slots. I'm not too sure that I agree that it's largest lack of abilities revolves around the concept of multiply and devide, yea it takes memory to do that with software but memory is cheaper than microcode. The Apple II line has a history of lack of support of the concept of interrupts. ProDOS is really the first OS that actively supports it. The UCSD p-system allowed it but didn't support it. ProDOS does provide you with methods of hooking your routines in using a very nice clean manner but then at EVERY interrupt, ProDOS runs a little ditty before handing off control to your code. (I grant you, they did it for the right reasons, to return the machine to a known state before you get control) but hell, that's not their judgement to make beyond the absolute minimum. Aside from all that, the II line lacked a non-preempt style of instruction that would allow you to run (cleanly) multiple tasks. Even the oldest IBM-PC had an instruction to increment a memory word that blocked all attempts at prememption until it completed. Using this, you could signal tasks into and out of a run/ready state. We ain't got it. HOLD IT! Before the flames get turned on, I am not saying I should be able to run a uucp transfer while I recompile the great American program in the background, but even little things like clean print spoolers, alarm clocks and the such are hindered in their abilities without such a thing. Oh well, time to move on. Thanks for the use of the soapbox, "NEXT!" << MCT >> GEnie : M.Tiernan AppleLinkPE : M Tiernan or BCS Mike Internet : pro-angmar!m.tiernan@alphalpha.com UUCP : ...!uunet!alphalpha!pro-angmar!m.tiernan "God isn't dead, he's only missing in action." - Phil Ochs
jb10320@uxa.cso.uiuc.edu (Desdinova) (12/07/90)
In article <10895.apple.net@pro-angmar> m.tiernan@pro-angmar.UUCP (Michael Tiernan) writes: [ mentions Apple's apparent early disregard for interrupts ] >ProDOS does provide you with methods of hooking your routines in using a very >nice clean manner but then at EVERY interrupt, ProDOS runs a little ditty >before handing off control to your code. (I grant you, they did it for the >right reasons, to return the machine to a known state before you get control) >but hell, that's not their judgement to make beyond the absolute minimum. With the IIgs and GS/OS, interrupts are actively used for all sorts of system tasks. The heartbeat timer, appletalk and serial ports, keyboard, etc. all generate interrupts and GS/OS depends on them. What you're probably referring to is vectored vs. single line interrupts. Tis' true the 6502/65816 doesn't have vectored interrupts, but adding them would be a tradeoff against system reconfigurability- in a PC you can only have a certain number of interrupt generating devices- in a II there is no limit. >Aside from all that, the II line lacked a non-preempt style of instruction >that would allow you to run (cleanly) multiple tasks. Even the oldest IBM-PC >had an instruction to increment a memory word that blocked all attempts at >prememption until it completed. Using this, you could signal tasks into and >out of a run/ready state. We ain't got it. HOLD IT! Before the flames get >turned on, I am not saying I should be able to run a uucp transfer while I >recompile the great American program in the background, but even little things >like clean print spoolers, alarm clocks and the such are hindered in their >abilities without such a thing. The 6502 does indeed lack a Test-And-Set instruction, but the 65c02 and 65816 both have it. TSB or TRB are all that's needed for mutual exclusion checking in a preemptive environment. (Note this doesn't hold true for multiple processor systems, but there aren't any, so I'm not worried.) Note that the 6502 DOES have an indivisible INC memory location instruction, but INC alone is not good enough for mutex. >Thanks for the use of the soapbox, "NEXT!" Thank you! -- Jawaid Bazyar | Being is Mathematics Senior/Computer Engineering | Love is Chemistry jb10320@uxa.cso.uiuc.edu | Sex is Physics Apple II Forever! | Babies are engineering
$CSD211@LSUVM.BITNET (Mark Orr) (12/07/90)
|From: Michael Tiernan <pro-angmar!m.tiernan@ALPHALPHA.COM> |Mr Shekhel is dead right, the basic problem with the Apple II line |is/has-been/will be that it lacks a few required instructions at the processor |level to allow it to be a competative force. This has been the problem since |way back in the days of the Apple ][+ when they decided that no one will ever |need an inerrupt line so they took them out of the card slots. I'm not too |sure that I agree that it's largest lack of abilities revolves around the |concept of multiply and devide, yea it takes memory to do that with software |but memory is cheaper than microcode. The Apple II line has a history of lack I didn't mean to say that hardware multiply and divide were all that is wrong (or even the crux of what is wrong) with the 6502/65816. I merely offered it as a typical example of what the 65xxx series lacks. But still, you can't deny that the 65s are lacking in math capabilities. Memory may be cheaper than microcode, but it is also far slower. Too bad the 65s don't have an associated FPU. (of course you can use any on the market, but...) |of support of the concept of interrupts. ProDOS is really the first OS that |actively supports it. The UCSD p-system allowed it but didn't support it. |ProDOS does provide you with methods of hooking your routines in using a very |nice clean manner but then at EVERY interrupt, ProDOS runs a little ditty |before handing off control to your code. (I grant you, they did it for the |right reasons, to return the machine to a known state before you get control) |but hell, that's not their judgement to make beyond the absolute minimum. So very true. Interrupt handling in the OSs leaves something to be desired. |Aside from all that, the II line lacked a non-preempt style of instruction |that would allow you to run (cleanly) multiple tasks. Even the oldest IBM-PC |had an instruction to increment a memory word that blocked all attempts at |prememption until it completed. Using this, you could signal tasks into and |out of a run/ready state. We ain't got it. HOLD IT! Before the flames get |turned on, I am not saying I should be able to run a uucp transfer while I |recompile the great American program in the background, but even little things |like clean print spoolers, alarm clocks and the such are hindered in their |abilities without such a thing. Yes, I can see that that would be nice. This is what concerns me about the ASIC. If you're going to go to the trouble, why not correct the '816s deficiencies. Instead of building a '816 clone, why not build an enhanced '816 more along the lines of the Hitachi 64180 or the 68HC11 or one of those other well groomed device controllers. I can't believe that ASIC is cloning the '816 just for a fast GS (that's much too risky, it might not be there by the time they're done :( ) WDC makes good money on the '816 in device control applications (medical devices, I think)...but the '816 just dosen't have the hardware features of the HD64180 or 68HC11 (even though these are "smaller" chips, i.e. address space) |<< MCT >> | |GEnie : M.Tiernan |AppleLinkPE : M Tiernan or BCS Mike |Internet : pro-angmar!m.tiernan@alphalpha.com |UUCP : ...!uunet!alphalpha!pro-angmar!m.tiernan | |"God isn't dead, he's only missing in action." | - Phil Ochs ---------------------------------- | MARK A. ORR | | $CSD211 @ LSUVM.SNCC.LSU.EDU | | @ LSUVM.BITNET | ----------------------------------
gwyn@smoke.brl.mil (Doug Gwyn) (12/07/90)
In article <10895.apple.net@pro-angmar> m.tiernan@pro-angmar.UUCP (Michael Tiernan) writes:
-Aside from all that, the II line lacked a non-preempt style of instruction
-that would allow you to run (cleanly) multiple tasks. Even the oldest IBM-PC
-had an instruction to increment a memory word that blocked all attempts at
-prememption until it completed. Using this, you could signal tasks into and
-out of a run/ready state. We ain't got it. HOLD IT! Before the flames get
-turned on, I am not saying I should be able to run a uucp transfer while I
-recompile the great American program in the background, but even little things
-like clean print spoolers, alarm clocks and the such are hindered in their
-abilities without such a thing.
What an idiot.
toddpw@nntp-server.caltech.edu (Todd P. Whitesel) (12/07/90)
WDC's microcontroller versions of the 65c02 and 65816 have true vectored interrupts for EVERYTHING... they are extremely nice microcontrollers and I can see why they are popular. The problem with interrupts in the Apple II was that Woz's original IRQ handler used a zero page location, and backwards compatibility issues became important _before_ he had a chance to fix it. Add that to the fact that Apple didn't settle on an interrupt handler protocol until Prodos -- there were memory state issues to deal with by then and things were just messy. The GS and GS/OS do a far better job, but the real solution is true vectored interrupts. The 65816 provides support for externally vectored interrupts (in fact, all you need is an SRAM and a PAL) but Apple hasn't done anything about it. It's tempting to do something like this as an add-on -- it wouldn't be pretty but it could be done. It might be easier to add to the Zip or the TWGS. Todd Whitesel toddpw @ tybalt.caltech.edu
rhyde@ucrmath.ucr.edu (randy hyde) (12/08/90)
The 6502 indeed supports several "test and set" instructions. You've probably known them as LSR and ASL in past incarnations. Both of these instructions can be used to perform an non-interruptable test and set operation. Of course, anyone who has taken a operating systems class in the past 15 years or so *SHOULD* realize that (1) an atomic (non-interruptable) instruction doesn't always do the job. I.e., TSB would not work if you have two processors accessing memory because only memory cycles, not instructions, separate accesses to the semaphores. I ran into this problem when working with a MILL card (6809 coprocessor) on a lowly Apple II (not even a "+"). The solution? (2) Lamport's Bakery algorithm (or Dekker's algorithm for only two processes) works just fine. You don't need uninterruptable instructions to prevent multiple processes from entering a critical region. As for the complaints about the Apple II not supporting multitasking: It most certainly does. I've written a couple of multitasking OSes for the Apple II (mostly as an academic exercise). What people really want is a multitasking DOS 3.3, ProDOS, or GS/OS. *THAT* is probably not possible. But multitasking certainly is; I've done it. While it's true that the 6502 doesn't support memory management (on chip, it can be added) and other neat features, people were running multiprogramming operating systems on CPUs much me primitive that the 6502 25 years ago. To address the other questions of 6502 shortcomings: keep in mind that the 65xxx family is a family of controller chips, not general purpose CPUs. It doesn't support MUL and DIV or (heaven forbid!) floating pont because the applications (i.e., Nintendo) where the chip was destined don't require it. They need something cheap. MUL, DIV, and FP cost money! Apple II (and especially Apple //gs) users think their machines drive the CPU development. DEAD WRONG! Apple has only sold a couple hundred thousand GS machines. Compare that to the millions of Nintendoes sold. If you want to run UNIX, buy a Next (if you're a student) or an AMIGA (if you're not a student). The 65xxx family just doesn't have enough get up and go to support UNIX. Better yet, scoot on over to the comp.sys.nsc32k group and order a PC532 board (32532 CPU). Now that's a great UNIX chip. *** RAndy Hyde O-)
alfter@uns-helios.nevada.edu (SCOTT ALFTER) (12/08/90)
In article <10395@ucrmath.ucr.edu> rhyde@ucrmath.ucr.edu (randy hyde) writes: >As for the complaints about the Apple II not supporting multitasking: >It most certainly does. I've written a couple of multitasking OSes for >the Apple II (mostly as an academic exercise). What people really want >is a multitasking DOS 3.3, ProDOS, or GS/OS. *THAT* is probably not >possible. But multitasking certainly is; I've done it. While it's >true that the 6502 doesn't support memory management (on chip, it can be >added) and other neat features, people were running multiprogramming >operating systems on CPUs much me primitive that the 6502 25 years ago. You've gotten the II to multitask? That sounds interesting. I've thought one way to multitask would be to grab an instruction from each process and execute it, but that would be painfully slow--basically, a 6502 interpreter running on a 6502! How did you go about multitasking the II? I'd be interested (and other people would also doubtless be interested)--if for nothing else than just to see how it can be done. ----------------------------------------------------------------------------- Scott Alfter _/_ / v \ Apple II: Internet: alfter@uns-helios.nevada.edu ( ( the power to be your best! GEnie: S.ALFTER \_^_/
AABENSON@MTUS5.BITNET (12/08/90)
Just as a small correction, in the IIgs (running P16 or GS/OS), there IS a little memory word increment thingy-doodad -- It's called the "Busy Flag". It gets incremented when stuff becomes busy, and decremented when stuff becomes unbusy. Also, I think that it would not be too terribly difficult to run a multitasking OS on the IIgs -- although I would HIGHLY recommend more memory and a speed accelerator. For instance, you'd need to be able to schedule interrupts at any time you want -- no problem, that's why nobody ever uses all 16 voices of the Ensoniq chip anyway. By adjusting the frequency (pitch) that it plays a flat (no-sound) waveform at, you can control how quickly it gets done, and if desired (it would be desired.), how often it (the Ensoniq 5503 or whatever number) generates interrupts. Of course, a major problem with this would be that many many programs which use sould would suddenly become incompatible. Also, you couldn't really run a couple desktop programs at once, but it WOULD be good for things like alarm clocks, print-spoolers, or even background modem transfers. What should really be done is for somebody to write this "thing", and then create some sort of shell that would run under it -- Then, you'd boot up GS/OS to this shell, and go from there... Otherwise you're talking about rewriting the whole darn operating system... Maybe somebody wants to do that, but it's certainly not me (by myself). I think a small group of people (or even a single person) could write this sub-OS (if I may call it that) without too much difficulty. Of course, it would have to be incredibly flexible and come with a bunch of nifty stuff to encourage others. I guess that's all my ideas for now. Please please please PLEASE comment on them (at length would be nice)... Also, email would be really really REALLY super-duper nice, since we don't get much net news, and half of what we DO get gets lost because this system (MTUS5) has been losing data lately (crashing I think). So if you can, please email to my internet address below. Thanks! - Andrew A. Benson Internet: "aabenson@balance.cs.mtu.edu" Bitnet: "AABENSON@MTUS5.BITNET" (Internet address is preferred)
unknown@ucscb.UCSC.EDU (The Unknown User) (12/08/90)
In article <10395@ucrmath.ucr.edu> rhyde@ucrmath.ucr.edu (randy hyde) writes: >Apple II (not even a "+"). The solution? (2) Lamport's Bakery algorithm >(or Dekker's algorithm for only two processes) works just fine. You I dunno, maybe it's just my lack of sleep, but it's neat to see these algorithms I read/study in computer classes discussed here! I like the dining philosophers problem too. N-1 people at a table, and N chopsticks (one on each side of a person). >possible. But multitasking certainly is; I've done it. While it's >true that the 6502 doesn't support memory management (on chip, it can be >added) and other neat features, people were running multiprogramming >operating systems on CPUs much me primitive that the 6502 25 years ago. The 68851 (PMMU) car would still be neat... >drive the CPU development. DEAD WRONG! Apple has only sold a couple >hundred thousand GS machines. Compare that to the millions of Nintendoes >sold. Is "10" a "couple"? It seems that the commonly accepted value is at least a million GSes out of over 5 million Apple //s total. >If you want to run UNIX, buy a Next (if you're a student) or an AMIGA >(if you're not a student). The 65xxx family just doesn't have enough >get up and go to support UNIX. Better yet, scoot on over to the >comp.sys.nsc32k group and order a PC532 board (32532 CPU). Now that's >a great UNIX chip. A great UNIX chip? Please explain in mail... Well, the person who is going to hopefully make the 68851 card has also compiled part of UNIX on a GS and I believe it worked... (although I honestly don't know how they could tell if they'd not written the low-level GS specific routines necessary). Anyway, you think the GS with the 25 megahertz 65816 and a 68851 wouldn't have the "Get up and go" to support UNIX? -- /Apple II(GS) Forever! unknown@ucscb.ucsc.edu MAIL ME FOR INFO ABOUT CHEAP CDs\ |WRITE TO ORIGIN ABOUT ULTIMA VI //e and IIGS! Mail me for addresses, & info. | \ "Dammit Bev, is it you inside or is it the clown?" -IT by Stephen King /
$CSD211@LSUVM.BITNET (Mark Orr) (12/09/90)
|From: randy hyde <ucrmath!rhyde@UCSD.EDU> <stuff deleted> |To address the other questions of 6502 shortcomings: keep in mind that |the 65xxx family is a family of controller chips, not general purpose |CPUs. It doesn't support MUL and DIV or (heaven forbid!) floating pont |because the applications (i.e., Nintendo) where the chip was destined |don't require it. They need something cheap. MUL, DIV, and FP cost |money! Apple II (and especially Apple //gs) users think their machines |drive the CPU development. DEAD WRONG! Apple has only sold a couple |hundred thousand GS machines. Compare that to the millions of Nintendoes |sold. Apple has sold more than a million IIgs's. The IIgs was for a while (late 86 early 87, just after it was released) a very hot selling item (despite short- ages of the Ensoniq chip). The Apple II series was not the only one to use the 6502...remember the Atari 400/600/800/1200/XL's and XE's (probably a couple of million of those), and the Commodore PETs and VIC-20s (a million or so more), and the BBC micro (more millions), the Rockwell KIM/SYM/AIM (a few hundred thousand), not to mention other sundry no-names. I really doubt that machines like the Nintendo were the "intended market." To say that MOSTEK didn't have computers in mind when they put together the 6502 is just silly. Compared to the chips of the day (8080s, 6800s) the 6502 looks deluxe. Now the chip you also mentioned, the 6809 (used in Stellation Two's MILL and VitaMILL) was a really nice chip...too bad it was wasted on a veg-o-matic like the Radio Shack Color Computer. Yeah, MUL, DIV and FP instructions cost money...But show me another personal computer (built now) based on a CPU without divide and multiply. To say that these instructions have no use on a computer (even a Nintendo) is goofy. Try doing rapid coordinate transforms without 'em. |If you want to run UNIX, buy a Next (if you're a student) or an AMIGA |(if you're not a student). The 65xxx family just doesn't have enough |get up and go to support UNIX. Better yet, scoot on over to the |comp.sys.nsc32k group and order a PC532 board (32532 CPU). Now that's |a great UNIX chip. |*** RAndy Hyde O-) Any of you seen the new Amiga 3000UX. Comes with UNIX system V release 4, X windows and Open Look...it easily outperforms the old NeXT (pre '040s), and it has all the AMIGA custom hardware (so much for Jobs' "mainframe power on two chips). UNIX could be run on an Apple II...with a coprocessor card. Say, if a card with a Inmos T400 or AMD 29000 were added, along with higher-resolution graphics. (the old Saybrook and PDQ II coprocessors (68000s) ran versions of UNIX, but of course they're long gone). If we formed a company to build a renegade II, with the coprocessor built in...Hey, who knows? ---------------------------------- | MARK A. ORR | "Eww...I'm so mad, I'm gonna get in my | $CSD211 @ LSUVM.SNCC.LSU.EDU | Lincoln Town Car and lean on the gas | @ LSUVM.BITNET | 'till I get to West Virginiaaaaa" ---------------------------------- - Larry Bud Melman
gwyn@smoke.brl.mil (Doug Gwyn) (12/09/90)
In article <9012090428.AA16099@apple.com> $CSD211@LSUVM.BITNET (Mark Orr) writes: >Yeah, MUL, DIV and FP instructions cost money...But show me another personal >computer (built now) based on a CPU without divide and multiply. To say that >these instructions have no use on a computer (even a Nintendo) is goofy. Try >doing rapid coordinate transforms without 'em. Some CDC mainframes did not have "divide" instructions, yet nobody accused them of being slow. Multiplication is of some use but division is seldom used in most applications. Floating point is of even less general use, and in fact floating-point accelerators are frequently an extra-cost option, so that only those who need them pay the price. Note that a 68881 card is available for the Apple II family; I have one but it is of noticeable benefit only for a few specialized applications. >UNIX could be run on an Apple II...with a coprocessor card. That would be rather pointless. In fact, UNIX (circa 7th Edition) could run on a stock IIGS. However, it is doubtful that it would be worth the porting effort. Talk about manufacturing your own enhanced Apple IIs is silly. Hardware is not important; software is. Nobody is likely to market any software for your deviant machine, so you'd have to roll your own. Not even most of the computer industry giants were able to survive without third-party commercial software for their systems.
rhyde@ucrmath.ucr.edu (randy hyde) (12/11/90)
I now have read two replies which claim apple has sold over 1 million Apple //gs systems. I have never seen this in print. Could someone give me a reference to support this figure? I would like to believe you, but it sounds like too much wishful thinking to me. Although 1 million computers is small potatoes today, I still think it's a large enough market to attain the "critical mass" required by various developers to support a machine. That's why I have trouble believing this figure. *** Randy Hyde O-)
mvk@pawl.rpi.edu (Michael V. Kent) (12/12/90)
In article <10426@ucrmath.ucr.edu> rhyde@ucrmath.ucr.edu (randy hyde) writes: >I now have read two replies which claim apple has sold over 1 million >Apple //gs systems. I have never seen this in print. Could someone >give me a reference to support this figure? I would like to believe I really don't have the answer you're looking for, but maybe I can be of some help. Early in 1989, while waiting for someone at the library, I picked up an IBM magazine for a quick look. In it, they published their top ten selling machines for 1988. Number 1 was the IBM PS/2 model something selling 350,000 machines. Number 2 was the Apple IIGS selling 340,000 machines. That was for calendar year 1988. Perhaps someone can look up the numbers for 1986, 1987, 1989, and 1990. I'd be very VERY surprised if they didn't add up to 660,000 machines. The GS was a pretty hot item until a little over a year ago. Michael Kent mvk@pawl.rpi.edu
awillis@pro-angmar.UUCP (Albert Willis) (12/14/90)
In-Reply-To: message from gwyn@smoke.brl.mil Could somebody please inform Free Tools Association that they can't continue to create breath taking 3-d demos, since the IIgs doesn't have multiply and divide instructions in silicon? Gee, you would think that people would understand that doing 3-d graphics on the IIgs is impossible without these instructions. Right? ;-) Albert Willis INET: pro-angmar!awillis@alphalpha.com | America Online: BCS Al UUCP:..!uunet!alphalpha!pro-angmar!awillis | GEnie: A.Willis