From: dhaynie@jersey.net (Dave Haynie) Newsgroups: comp.sys.amiga.misc,comp.sys.amiga.hardware,comp.sys.amiga.advocacy Subject: Re: Problems with PIOS (was Re: State of PIOS) Date: Mon, 15 Sep 1997 05:53:48 GMT Organization: PIOS Computer Lines: 333 Message-ID: <341cbdba.907698030@news.jersey.net> References: <1472.183T953T12723469@accesscomm.net> <5ulqfo$62v$1@coranto.ucs.mun.ca> <5vbomv$6kl$1@coranto.ucs.mun.ca> NNTP-Posting-Host: swip130.jersey.net X-Newsreader: Forte Free Agent 1.1/32.230 Path: 195.95.96.10!news.unisource.be!planetinternet.be!newsgate.unisource.nl!newsgate.cistron.nl!het.net!Cabal.CESspool!bofh.vszbr.cz!nntprelay.mathworks.com!news-dc-10.sprintlink.net!news-pull.sprintlink.net!news-in-east.sprintlink.net!news-peer.sprintlink.net!news.sprintlink.net!Sprint!uunet!in4.uu.net!209.66.0.12!news.jersey.net!not-for-mail Xref: 195.95.96.10 comp.sys.amiga.misc:153279 comp.sys.amiga.hardware:151367 comp.sys.amiga.advocacy:181710 On 12 Sep 1997 15:56:15 GMT, gnoel@morgan.ucs.mun.ca (George Noel) wrote: >In article , >Dr. Peter Kittel wrote: >> We from PIOS, however, feel that we do something with our >>PIOS ONE hardware which is practically what would have been the next >>Amiga hardware from Commodore, like it or not! Yes, there was also >I highly doubt it and I can't see how you would be promoting that. Sure >all of the staff from PIOS would have more inside information then I do >but do you seriously think Commodore would have went in the same direction >if they had stayed alive? It's this simple: I designed the PIOS One. I was designing the post-A3000 architecture at Commodore, pretty much single handedly. If you took a few gate arrays out of the first implementation of the "Acutiator" architecture, and replaced them with off-the-shelf system chips (unavailable back in 1991/92 when I started the architecture), you would have a very similar system. Sure, we still had Amiga chips. I had worked out a system interface chip that could let us build AAA modules for different goals, and while a AA module was possible, it's unlikely we would have done one, since cheap SVGA chips would work better. The AmigaOS 4.0 would have some support for basic AAA features, and it could also drive some basic bitmapped modes in any SVGA-type PCI chips it detected. More advanced graphics modes, in both cases, would run through RTG drivers. Later on, Hombre looked like our likely long-term goal for graphics module; unlike AAA, it would have been cutting edge by its release, if C= had stayed on-track. >They were heading towards Hewlett Packard with their HP-PA RISC which >would have ruled out PPC and CHRP/PPCP. You're mistaken. PA-RISC was used as the embedded processor architecture for the Hombre chipset. The systems group had quite a few reasons why PA-RISC would not have made a good RISC architecture for the general purpose systems. Top on that list is the fact that HP just plain doesn't sell their more aggressive PA-RISC chips. >They probably would have went PCI We were definitely going PCI. In '91, the new architecture used a new bus called the Amiga Modular Interconnect (AMI bus for short). This was optimized for custom chips (Zorro II and III, like ISA, EISA, and most of the "expansion buses", were supposed to be something you could build without access to custom logic), chip-to-chip interconnect, with a CPU independent protocol (we had 68K, we knew it was going away). >and probably would have did the CPU daughtercard thing with the ram on it >etc. I had outlined the basic architecture of an '040/'060 module and an 88110 module. PowerPC didn't yet exist, PA-RISC, again, wasn't interesting. > and they might have adopted some of the newer ports/busses There really wasn't anything much on the horizon. The orginal system chip design drove a flexible bus that could handle both 650x/6800 and ISA-type peripherals, with some built-in expandability. The second version was much more weighted to running an ISA-like bus, but not for user-expansion, just as an X-bus kind of thing, totally replacing the 650x bus in the Amiga. We had to get rid of the 8520s anyway, so I had suggested a chip integrating three 8520 cores but using the standard bus. Pretty much anything else you'd want in chips ran the ISA protocols. Which was fine for non-PCI stuff anyway, you ran a 16-bit bus at 8MHz, whereas the 650x bus was 8-bit at 716kHz. > but I >doubt we would have seen the same hardware configuration as the PIOS One. You would be quite wrong, of course. Yes, we had to make more of our own parts, but if anything, many would have been worse than what you can get off-the-shelf today, simply because of the fact that this was a big project for Commodore. Not that we lived long enough to do more than bits and pieces of design and simulation -- even at the start, management wasn't sold on any future planning, so I was doing most of this as a skunkworks project. Which was the rule in late '91. The CPU module for the A4000 began life as an undergound project to build "the cheapest '040 module known to man", as Berlin put it, to add to the "A2200" (also known as the A1000jr), a scaled-down A3000 down wrong (ECS chips in the Spring of '92). >There are some still out there thinking the Amiga shouldn't have gone the >direction of the Mac again, Other than "down the drain", the Amiga hasn't been heading anywhere near the "direction of the Mac" lately. Going to the PowerPC has nothing to do with the Mac, directly. Yes, it's the Mac that's driving volume on the PowerPC, and if that's the only success story, maybe the PPC won't be valid target for the AmigaOS in awhile. But really, where else would they go? Given that 90% of the Amiga market was in cheap systems, and yet, there's a demand for high-end systems, where could you turn? Except x86, which as ugly as it is, wouldn't be a shock, considering Gateway's strengths. >>Hombre, but I've seen some docs where this was scheduled to become >>either a games console, not necessarily much Amiga-like, or an own > >True but some great features. Hombre was a brand new chip architecture. The two-chip set, perhaps with a few cheap jellybeans around them, and of course enough VRAM, would have formed the basis for the next generation CD32-type thing. They didn't plan anything like an AmigaOS prot to PA-RISC, but rather a simple graphics kernel with 3D library. We would use the Hombre on PCI (it has PCI built-in) for Amiga systems, via RTG. >>gfx card with just an own kind of gfx chip instead of some from the >>gfx card makers. >Yes, which could have possibly been made to work on higher end Amigas >also. I think Dave has said he was working on a system to allow this or >was thinking about it anyways. Hombre came along a bit after I started working on Acutiator, but I think Ed Hepler saw the same "SUCCESS" as I did, written all over PCI. > AAA would have still served its purpose and last I heard, HP was >interested in using them in set-top-boxes.\ Commodore and HP talked about set tops, but it was AA (eg, AGA) where the interest was. They had talked about hooking up with Motorola and building an integrate AA and CPU (maybe '030, maybe CPU32) for set tops and, naturally, cheaper A1200s. >>likely that Commodore would have gone hardware-wise in exactly this >>direction where the PIOS ONE now develops. >No. Yes. When I had finished the PIOS One, it wasn't really until I had the PC board in my hands when it occurred to me how much it was like the Acutiator. When I was working on Acutiator, the cheaper SparcStations came out, and they showed off their motherboard in the EE Times, about the size of a normal US sheet of paper (11x8.5"), so I wanted that too for the Acutiator motherboard. The PIOS One is about 12" x 10". >They might have taken more advantage of off-the-shelf graphics chips >etc. also but I wouldn't have seen them going in the same direction. Guess that's why they pay ME the big buck (well, I can dream)... >>> Unfortunately Dave is wrapped right up there in >>>it. I wouldn't have this attitude had their company provided some kind of >>>support to the Amiga community directly >>What should we do? Sell 68k stuff? For a new company, this would mean >It would have helped keep a bit of loyalty. We're still thinking about stuff to directly support the Amiga. Doing a 68K system isn't really an option. You don't want to know how many moderate speed PPCs you can get for the price of an '060, or even an '040. Not to mention that, to go to a modern hardware system, we would need an AmigaOS source licence, since everything's dependent on hardware you basically can't get anymore. What we can do at PIOS will expand in the coming months, now that I'm working full time, and we have some actual money to spend. The PIOS One is flexible enough, certainly, to let us follow any reasonable lead that Gateway/Amiga Inc. makes, even if it's not PPC. >> working on next-generation Amiga, as we think it should >I think that is a false claim, I am telling you, unequivocally, that the architecture for the A5000 was very much the same idea as the PIOS One. You can argue all you want, but you can't change that fact. The differences were strictly based on the times: I couldn't get parts for the Acutiator architecture off the shelf yet. I certainly had hopes that for many functions, our first chipset would be the only one we needed until something better was available off-the-shelf. Other things, like SCSI, were absolutely dependent on PCI chips. AAA would have cost much, much more than off-the-shelf PCI chips, it would have been killed by those, eventually, if Hombre didn't get there first. > and I also think pOS being referred to as the next-generation AmigaOS is a false claim I can't say, I don't know pOS. Clearly, it was designed enough as an Amiga-like OS to coexist with the AmigaOS. According to our software gurus, they've even made some of the same mistakes. If Amiga-like is good enough, maybe you'll be happy with pOS. When we first spoke to Pro-DAD, there was very little hope of any real AmigaOS being done for any RISC processor. While the hope has improved, I don't recall hearing any official announcement yet. > also as well as BeOS being the successor of AmigaOS. The BeOS is certainly NOT AmigaOS, any more than NeXTStep is UNIX or Solaris. But it certainly is a modern operating system designed to solve many of the same problems that the AmigaOS was designed to solve. It also addresses most, if not all, of the AmigaOS's current shortcomings. And it does incorporate some of the best ideas from the AmigaOS. It's also the only one on the list with the potential for immediate large scale support. You may or may not like it. We're not trying to shove it down anyone's throat, but after the VisCorp chimera vanished into smoke last August, the BeOS was the only game in town, and it was a perfectly good one. So we support it. As I've said many times, we'll be first in-line to support a real AmigaOS for the PowerPC. We'll be very interested in supporting the AmigaOS for any other RISC chip, if that's the only option. We specifically did not support any OS that couldn't run on any other "Amiga-industry" system moving to PowerPC -- we're not, for example, writing our own OS and claming that to be the next AmigaOS. >>look like and as probably Commodore would have also built it. > >Commodore's might have LOOKED the same (case wise) but that's about it. No, that's the only part you can guarantee is different -- C= wouldn't put a PIOS logo for an air intake. >>In which way is cloning with official licenses "stealing"? >In this case I am referring to the Amiga market. Last I heard, you had no >Amiga license. You can only steal something that exists. The Amiga systems market has been all but nonexistant. I'm certain nothing PIOS has done changed the 100 or so A4000Ts that QuikPak sells every month or two. >> as we don't yet have some >> AmigaOS of any kind for our hardware. > >Exactly and that would at least be one step back in the right direction. And we're still waiting for it. WE went to VIScorp to talk about doing an AmigaOS for the PowerPC, back in May '96. We offered to assist on the port, or do it ourselves. They blew it, not us. After that, we had the choice of sitting around and waiting for the next guy on a white horse, or doing something. Since we had some dream about running a successful computer company, someday, we decided to do something. I can understand that guys who just sit around and criticize don't get this part of the story. You're still just waiting. >> but we're getting close as much as possible >>to the real Amiga. > >pOS and Siamese is not good enough in my opinion. Then why aren't YOU out there doing something better? I mean, if it's so easy and all. >>>Ok, back to the point way above, I chose the Amiga for a reason and its >>>purposes suit my need for the next while, if I went with the PIOS One, I >>>will lose a lot of that functionality. >>So you use majorly the functionality of the custom chips? Weird. >In a lot of cases, yes and parts of the OS aided by those custom chips, >yeah. That's the copper only -- every other part of the Amiga chipset is standard these days. And some even do have display list coprocessors, though I don't know how they compare (quite a few of the super-high-end graphics cards, the new Number Nine "Ticket to Ride" chip, etc). Video has become standard on many PCI cards in the last few years, with overscan, often with multiline interlace filters, etc. Some make it an option, but add capture as well, not a bad tradeoff considering that most folks who need video out also need video in. So what are you doing so much with the copper? >>It's more typical for Amiga folks to see such openness as an advantage. >In some ways yes.. but the PIOS One offers more openess then even Amiga >users are used to. There still is the question whether this is better or >not for productivity. Amiga users will appreciate it much more than Mac users simply because they've been burned by proprietary systems much more than Mac users have been, yet. At least if they really think about it. For example, Commodore's death. Because of the fact that the hardware was proprietary, the AmigaOS only ran on it. So the AmigaOS was of absolutely no value to a company not also interested in taking over the hardware works. And that, given that the hardware was a good two generations behind the times when C= went under, and not easily retargeted to new stuff, has been the problem to date. Current Amigas are at least five generations behind, while current Macs are only about two -- and since they can use PCI devices driectly, and do have faster CPUs to start with, much of that's hidden to the user, because it's not always a performance issue. Also, Apple isn't dead yet, and they may manage to get up to date on system hardware -- all the CHRP machines are, even if Apple's cutting them off.. >>Perhaps your error in perception is that Dave did *not* change his >>attitude? >Well of course I cannot fully tell that, you are right. I do not know Dave >personally, nor have I worked with him. But looking at him from a >difference it seems like it would almost be like a change from drinking >hard alcohol to going to drinking beer. :) Nope, I was drinking beer all along. Only, years back, that was a problem. I could pay way too much for stale beer from Europe, or drink the swill from Budweiser and Miller. So I had a group to make me my own custom beer, and while that tasted good for awhile, they just couldn't keep up. I noticed they'd take a long time to make a new batch, and usually it wasn't much different than the previous batch. But at the same time, a number of micro and craft breweries went into business (well, more like 6,000, but that's another story), and they also made custom beer. It started out a bit rough, but gradually they made it better than my own guys, charged less for it, and kept up with the same innovations. So I could get the same old lager from my guys, by now even new batches tasted a bit stale, and the couldn't make enough of it. Or I could pick from a wide variety of porters, ales, IPAs, stouts, bocks, octoberfests, pilsners, lambics, bitters, weizens, dopplebocks, etc. >> He always just chose the way which was appropriate for the >>time, and times *did* change. Think over it. >Sure, times did change and yes, the OS has become more important but there >is still a need and a market for the hardware aiding the OS so the OS can >get the advantage. The hardware does aid the OS. It aids everyone's OS these days, and most of the hardware is doing much more for them than it's doing for us, on the Amiga, these days. The OS is the advantage -- there is nothing you can do to make Windows anywhere near as fast as AmigaOS, even on the same hardware. And yet, every Windows machine still for sale is faster than every Amiga ever made. Do the math. The think about RISC, where we get another advantage. Dave Haynie Vice President, Technology PIOS Computer haynie@pios.de "...no RISC, no fun"