IRC log started Thu Nov 25 00:00:01 1999 -:- SignOff smkl: #TUNES (blargh) [msg(TUNES)] permlog 1999.1125 -:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes -:- SignOff air: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep) -:- SignOff yoo: #TUNES (Leaving) -:- haiwu_ [bbs@develop.21cn.com] has joined #tunes -:- haiwu_ [bbs@develop.21cn.com] has left #tunes [] -:- SignOff eihrul: #TUNES (Ping timeout for eihrul[usr5-ppp29.lvdi.net]) -:- eihrul [lee@usr5-ppp29.lvdi.net] has joined #tunes -:- SignOff eihrul: #TUNES (Ping timeout for eihrul[usr5-ppp29.lvdi.net]) -:- eihrul [lee@usr5-ppp29.lvdi.net] has joined #tunes -:- SignOff eihrul: #TUNES (Ping timeout for eihrul[usr5-ppp29.lvdi.net]) -:- Fufie [stig@tunnel-44-23.vpn.uib.no] has joined #tunes !services1.openprojects.net!! PANIC! buffer = :doogie QUIT :going home -:- Downix [down@d-gnaps-221.ici.net] has joined #tunes hello all Need help with an OS, again Anyone awake? what is your problem? 07:00am I found an OS coded in 68k ASM, I don't understand 68K ASM, is there any way to turn it into C? (I know there has to be a way to) you could try to find an emulator I have a 68K machine to run it on but I want to study HOW it works haven't seen any asm2c's lately.. you could turn it to C but the c code would suck I found several, for OS/390, Alpha, VAX, even x86 smkl: Even if the C code sucked, I could study it easier. read the comments No comments sounds like a good os :P Damned good. Too bad Commodore is an idiot Well, was * Downix/#tunes still can't believe they didn't comment the damned thing Anyways comments are easy to strip Likely they were stripped except there are a few comments left are you sure it was written in asm from scratch and is not generated asm? Fufie: I don't know, it could be generated. I am suprised I even found a copy of it in the first place 07:10am -:- smoke` [tw026024@zaalf24.twi.tudelft.nl] has joined #tunes hi hey hey! what is up? the Languages.html page has a few pointers on decompiling Downix: the sky Fare: Oh? I have a decompiled source I'm trying to figure out. 07:40am either it was stripped of comments (likely) or didn't have any (also likely knowing the origin) Downix: what program/language/purpose/foo ? AmigaOS, ASM, running an Amiga computer * Downix/#tunes is very curious how the hell this OS handles multiple CPU's Also it's a bit of a curiosity to see how the hell it's so damned flexible while it's so damned fast 1) sound software design on top of sound hardware design 2) no paranoia * Downix/#tunes nods I'm trying to achieve the same with my own hardware design Our hardware is good, but we need an equally good OS design F-CPU? F-CPU is http://f-cpu.tux.org/ or just plain MIPS? MIPS if you don't want linux, contact core@suntech.fr... his clementine is only on PCs at the time being, but the design is portable and maybe external alliances will help speed up the release of the code under GPL... 07:50am -:- AlonzoTG [Alonzo@216-164-137-81.s335.tnt4.lnhva.md.dialup.rcn.com] has joined #tunes Fare: I've talked to him, it sounds good. But w/o the documentation, I can't work with my OS guy here on it I need to see if it a) can be ported and b) ease of porting And for that I have Al here to help om 08:00am -:- binEng [Anders@j141.ryd.student.liu.se] has joined #tunes Is bespin.cx down?? binEng: no 08:20am So how come I've been unable to connect for the last day or so? Downix: do you have cvs access to see by yourself? I have CVS, but I don't have access to core's site I don't think Downix: ask him for it I would if he doesn't ping out whenever he logs on I'll email him asking phone him, that's surer! Sure, call France just out of the blue... hold on, what's the phone #? The fact is, I can't even get my mail here. bespin won't answer. 08:30am * Downix/#tunes looks up suntech's web site nope, only a snail mail address -:- SignOff Downix: #TUNES ([BX] Everybody was Kung Fu fighting!) 08:40am -:- Downix [down@d-gnaps-190.ici.net] has joined #tunes -:- SignOff AlonzoTG: #TUNES (Read error to AlonzoTG[216-164-137-81.s335.tnt4.lnhva.md.dialup.rcn.com]: Connection reset by peer) Fare, what do you know of Clementine? * Downix/#tunes is trying to get as much dat as possible on it 09:00am Downix: modular design, due to be GPLed RSN for a long time, but apparently the upper management is engaged in releasing it as GPL when it's "ready" (i.e. a shell with cp and mv) something like 7 people are now working on it full time * Downix/#tunes nods I know this all information I have is either second hand from core, or from peering at the source 09:20am * Downix/#tunes nods they use some kind of IDL to describe their module interfaces, with which they are not fully satisfied, but they have nothing else hmm * Downix/#tunes has the AmigaOS's source, maybe that could give some pointers since it runs in a similar concept, modules for most primary functions Downix: I suppose the AmigaOS source code is not redistributable? Fare: Right. It is not even open-source Die! Die! Die! evil proprietary software. I agree Fortunately AmigaOS is easy to decompile -:- SignOff smoke`: #TUNES (BitchX-75+Deb1an -- just do it.) 09:30am which is why proprietary licenses are a disadvantage to honest people, but not to dishonest competition I agree Of course licencing my companies architecture will be open and availible, limited only to quality of course we're keeping the chipset closed, so you have to buy from us. (have to make our money somehow) IP has corrupted the very heart of world economy More like patents have Sealed off IP is one thing, you can always re-do it all in an IP-free world, not only wouldn't you need close design to sell your design services, but no one would want to purchase/manufacture/distribute/finance a closed design right * Fare/#Tunes considers shipping costs for a cybertool -- 50% of the knife's price I would love to have a completely open design but I know I cannot, yet * Downix/#tunes has a plan on releasing his design after 4 design generations have passed 09:40am why 4 and not 3 or 5? because that's the number of primitive elements ? before 3 I would be tied to an agreement with my technology partners and 5 because why wait any longer than I have to? Even then I couldn't release 1 piece of technology because of licencing issues stupid closed licence downix: put on Belle & Sebatian's "If you're feeling sinister" and you'll get in better mood :) Fufie: Hrm? Fare: Primitive elements are more like 24 seperate small elements. you sound so depressed and sinister: "stupid closed license".. lighten up :) Well, I am a bit upset over that -:- ree [jwm@207.98.153.227] has joined #tunes but I love my technology hey all hey ree hey down we all do.. :( We'll see how things go heh, after a year I still haven't taken the time to figure out the tunes project at least the open licence to the reference design plus complete open-source drivers and open specifications is a step in the right direction sounds better yes 09:50am so how is the current development going? ree: Beats the hell out of me, I'm a hardware guy ahh =) what sort of hardware if I may ask? I'm making an architecture that will be availible for licence and a chipset to power it more akin to the AS/400 or Amiga than to a PC ahh do you focus on any specific aspect? like distributed shared memory, etc my hardware? yeah synchronized function, unified memory system, integrated functions, very scalable cool unified memory system =) how are you currently developing it? Right now mostly re-drafting how to put everything together simplifying the design to drop cost Trying to drop from our original $200 a chipset to our final goal of $65 a chipset ahh, it's a group effort I take it company? * Downix/#tunes nods what will the size of the hardware typically be? We're trying to get a new name tho, our last submission to the state was rejected 10:00am huge market for extremely small and quiet systems 2 chips, total of about 1.8 mil gates low-heat output ahh, so you won't be manufacturing entire systems? ahhh crap kde password prompt came up screensaver password feature doesn't work heh =) ree: Only reference and development systems. We'll be licencing the design to other companies to produce the system Avoid the problems Amiga, Atari, and Macintosh have had ahh Plus the problem of being an "unknown" company generating a name for itself if we licence it to one big company or several smaller ones, we'll get equal marketing I've always been interested in hardware design and concepts just never got into it you need a few people around you to start something like that has to be a group effort hmm would you be willing to actually develop your own systems if a company doesn't license it hoping that won't happen of course 10:10am Of course We have a contingency plan in place for that we'd purchase a small, nearly extinct brand name and market it throught that ahh well, it sounds very interesting Take advantage of another companies money troubles I hope all goes well for you I do as well BTW, 90% of the work is one man's wow He developed a system for the nearly extinct company years ago but they went bankrupt before patenting it smart guy ouch didn't he still manage to patent it? He did, yes doesn't cost too much for a initial patent filing ahh but the original design is obsolete now but it served as a begining as do most obsolete works =) right Oh well, their loss, our gain and if we can't licence it, their gain. 8) heh =) We're likely going to buy it anyways I hate to see an underdog beaten like that ironically, the reason it's nearly worthless is that they sold off all of their assets, people are paying attention to the assets, not to what people remember, the name yeah brand name and advertising are worth much more than product quality industry is a joke imo =) I agree 100% but why not bring quality AND a name? well, I'd first keep quiet about your good quality My partners don't want me to just buy out the company now and re-incorporate ourselves under their name just focus on the name So that's why I came up with the buying it seperately good quality seems to turn people off ahh They'll see our cost and jump at it so you're going to front the capital yourself? No, we need VC's ahh 10:20am but to buy the name and make a seperate company, I'll take a loan from my partnership company and buy the name myself who will take care of the issues of getting software developers for "your" hardware I take it the licensees No, I've been cutting deals with software developers already or will you be very public about your hardware design? ahh One of the points of allowing "Quality" licencees is so that software written for one of our licences will run on all yeah We will be public that's great I want to see our chipset in every possible niche I want to get rid of large computers and pcs 10:30am There will always be a need for a large computer really? but te chipset I have designed will reduce that need a lot if quantum computing ever becomes a reality I really doubt we'll need large systems Even if it does, people will combine quantum computers together to make quantum supercomputers since silicon CAN match a quantum computers performance hopefully by then though we'll realize how little meaning the computer industry for the personal consumer has I agree there yeah, but there won't be a need for quantum supercomputers There shouldn't be I don't see quantum computers as being within the reach of most people for a VERY long time by then we'll have mapped all human genomes and probably come to our senses about technology and life But there's more out there sure Personally, I don't concern myself with that What I am concerned with is that PC technology is STUPID! I really think we should be more concerened with what's here than out there =) hah, of course it is and it's about 20 years too old it should've been a mishap in some lab at ibm more than that IBM didn't make the PC architecture the original PC architecture was made at Digital in 1963 then quietly forgotten well, they basically started the marketing and use of it Yup but you can blame Compaq for making it the #1 architecture well you sort of have to blame us if Compaq hadn't cloned it, likely Macintosh or Amiga (more likely Amiga) would have won well, true we're the people who made them who they were even the most intelligent people in the field bought pc's Right now I'm thinking that my arch has a chance it can also be set up as a plug-in card so you can have one inside of your PC yeah, I think that is a great starting point -:- Cadet [manowar@cdu4-202.gu.net] has joined #tunes I mean, you stand to make enough money from just that alone to really market and develop your product more Our 3D will not be as good as nVidia's new chipset, but we can output to TV's, HDTV's, built in sound and I/O * Downix/#tunes nods -:- SignOff Cadet: #TUNES (Leaving) you just really need to hit the software developers Yup yeah multimedia is the way to go A lot of game companies like the design Makes development simpler for them 10:40am BeOS has stuck in the market since it's very multimedia friendly, even with the lack of hardware support Yup yeah, game consoles Amiga for the same reason, even tho it's 12-years out of date I really think more people should make more of an effort to produce a good game console Game consoles that they can make the software for on a standard PC with a plug-in card. 8) * Downix/#tunes nods, Well, this will be a good one even if it just happens to last a few months they still stand to make a profit Right not that I play games =) really think they're a waste of time * Downix/#tunes admits that he does They are If we can get 1 major PC game maker, we've got an in however, at that same time, they're the number one market for making a system widely accepted yeah * Downix/#tunes nods maybe some sport game maker Plus games made for the game console would run on the normal desktop there are many of those like the CD-32 so are you doing the major engineering work for it now? more like we're doing the cost-cutting work now ahh !irq:*! anyone know where services and friends went? Right now our chipset would cost $200 to produce too much We're getting it down to $65 Had to dump a few things what kind of stuff does your chipset do? like a PCI controller built-in isn't PCI in its end of life as far as design is concerned? -:- eihrul [lee@usr5-ppp31.lvdi.net] has joined #tunes In the end all it will do is contain the CPU (MIPS-based), a 3D system (nothing too fancy, all simple RISC), video display system (everything except the final screen draw section) sound system (digital output) and I/O controller (USB and Firewire). Plus there will be a local bus off of the chipset for furthur expantion PCI-X is not, and that's the PCI controller we dumped PCI-X is next-gen PCI !larne:*! services program probably died we're moving functions to the local bus section, off of the chipset so you can build a system with them, but you don't need them for things like a game console !irq:*! well can you start it up again? !larne:*! well no what's in PCI-X? !irq:*! why? PCI-X is the next-gen PCI standard, to begin service this summer 133Mhz PCI, backwards compatable with current PCI !larne:*! because doing that generally requries that you have access to the machine services runs on also, isn't LCD a completely different thing to drive than VGA video? !irq:*! what machine does it run on? !larne:*! varley 10:50am !irq:*! thats what I thought. and let me guess, only lilo and perhaps netgod have access to it? does PCI-X remove the need for AGP? LCD is, that is why final draw is not in the chipset !larne:*! and crom PCI-X does, yes !irq:*! tombstone was nice enough to search his buffer for the quit message: !***! !services1.openprojects.net! PANIC! buffer = :doogie QUIT :going home !larne:*! oh no where does "final draw" begin? !larne:*! that's a Bad Thing !irq:*! you're telling me external ASIC that the system needs !larne:*! services is my next project after I finish on ircd, so I'll have a look at it hey eihrul =) !larne:*! the version we run here is an ancient P09 version, afaik turns the digital audio channels into the analog needed, turns the video into something displayable yes, but I mean, what's still in the VDU? !irq:*! yeah.. sounds like a misaligned pointer or something because obviously a quit message like that shouldn't have done anything bad ree: okay, so where's your compression algorithm? ree: you're in a fervently free-availability-of-sources channel... :) hah RGB control, pixel map, color palette, pretty much everything including actually drawing the pixels, but final draw (lining it up, changing the screen refresh rates) is in the ASIC I know =) !larne:*! in the mean time, a cronjob to restart services would be nice. it goes down quite a lot !irq:*! yeah well I don't think that's going to get done unless we CALL lilo and _beg_ for it !larne:*! heh. well we could run it on leguin !irq:*! being that its some lame American-only holiday, he's probably busy !irq:*! Do you have recent dbs? !larne:*! I don't have any DBs, we'd have to get them off varley !larne:*! plus threaten netgod until be agress to furnish me with services source !irq:*! thought you didn't have access to varley? well? !larne:*! I don't :> I mean we'll have to talk to lilo and stuff so that way you change the ASIC if you're displaying to TV or to VGA or to LCD or to HDTV can't release it or there is a master ASIC to do all, but it's more expensive !irq:*! well then running buggy code on leguin wont be any different than running it on varley !larne:*! well no, but a) I can hack on it b) I can restart it !irq:*! oh yeah. or, they could just give you access to varley ree: gah, still don't see what its going to get you by keeping it closed :P !larne:*! the varley owners don't like new people, I did already ask lilo look at where gif went.... !irq:*! well damn ok, so what does the VDU output? sequence of digital RRGGBB pixels? Right with special stuff to program clocks, etc? right can you cope with vertical/horiz vs horiz/vert scanning ? !irq:*! I bet netgod has been sitting here the whole time reading this conversation and just not saying anything we have a method of doing that (and interlace, etc) !larne:*! no, he would be shouting at me by now * Downix/#tunes nods what software do you use for CAD? A system developed by our hardware guy CCAD do you control things at gate level? transistor level? mask level? Well, that's what I call it all of the above it's interesting you can develop on all levels at once I typically stick to the mask-level stuff tho 1.8 mil transistors is a bit too many for me to deal with * Fare/#Tunes remembers a PhD thesis about a set of macros atop LE_LISP to transform it into a multi-abstraction-level CAD system * Downix/#tunes nods Fare: are there any good GC books outlining specific implementations like mostly-copying, etc, that are hard to find elsewhere? 11:00am do you know that PhD thesis? The guy (Jacques Chambon) works in french CNET, doing software now. eihrul: there's the Jones&Li. and planar's thesis afraid not also his papers on the caml gc cool (adapted from the Lang&foo GC) * Downix/#tunes nods hmm Downix: which is cool? what's people's point of view on ExOS Fare: palar's thesis ExOS? what's that? ExoS? yeah, the exokernel system from mit Fare: mit's Xok based OS aegis/exos you know that planar was also crowned best hacker at the ICFP contest? cool Fare: hmm, found a nice looking list of papers while searching for jones and li on google XOK, ok, I know what you mean then yeah I really like their concept Good concept Fare: was the only thing i found, actually otherwise. yawn hah, yawn? this is a no-kernel polarized channel pretty much * Downix/#tunes nods http://caml.inria.fr/icfp99-contest/ eihrul: what did you find? yeah, tunes is no-kernel I just didn't find many papers on the idea of a non-kernel one sec, getting url ree: the tunes glossary :-/ Fare: www.cs.technion.ac.il/~erez/gc-seminar.html Too bad Tunes is so slow in development, but it's spawning off a lot of other OS's eihrul: did you subscribe to the GC list? down, slow really isn't the word for it =) Fare: i went to iecc and couldn't find where to subscribe Fare: i read the whole archive though, or most of it ree: Ok, makes a glacier seem like a participant in the Indianapolia 500 hah eihrul: ask for help to majordomo@iecc.com that's a closer description alright ree: Anyways, I think our hardware design is definately a winner Downix: where do I apply for a board? Fare: hehe, we're not near silicon yet. Once we finish the cost reduction, we have to FPGA it Fare: oh, they use md... downix, that's great I look forward to seeing it in my microwave and on my desktop hehe 11:10am What use in the microwave would we be? buying an ARM w/ USB controller would be smarter online browsing I need a mips processor, nothing less while waiting for your burrito to cool? * Downix/#tunes is using MIPS hot pocket I know =) MIPS is a cool CPU * Fare/#Tunes regrets the ARM much better name at least than x86 The ARM was a good idea, but they're stagnant now Downix: how much transistors are there in good FPGAs? doesn't the FPGA constrain the architecture in different ways than normal silicon? Fare: It does. But we can prove concept Fare: We're using million transistor FPGA's, availible from Xilinx and how much slower than real is it? all I know about FPGA are that they are necessarily slower than their silicon counterparts