FelineGrace@cup.portal.com (Dana B Bourgeois) (04/05/91)
Good 'Ol -MB- wants ta know why CBM hasn't improved the chips yet. And Randall tells him he's too impatient. Well, we all know the *real* reason, don't we? It's because -MB- hasn't COMPLAINED enough yet. CBM isn't gonna get serious about improving the old design until -MB- makes them uncomfortable enough. They'll just tweak it here and there to patronize him when he gets up on his soapbox so he'll leave 'em alone until his blood pressure builds up the next time. So come on -MB-, let 'em have it! Bitch! Moan! Complain! Turn up the Heat! Push that design, lift that test suite. You gotta make the 'chippies' at CBM nauseaous to the point they work 24 hours a day and get the new chip set done. Fini. Finished. Completed. So get going -MB- and Stick It To CBM but be Serious this time, OK? Stay on that soapbox until the new chip set is done, (or until an unruly mob of crazed, tired, over-worked, nauseaous former CBM chip designers hunts you down and rips you into Little Tiny Pieces.) Tongue firmly in cheek I remain, Dana Bourgeois @ cup.portal.com
cree@elec.canterbury.ac.nz (Michael Cree) (04/05/91)
It seems to me that being able to protect unallocated memory in
machines with an MMU should be no problem. Here`s roughly how I
would see such a an idea being implemented:-
First of all, the EXEC recognises which CPU is installed (68000, etc),
so it should know about the presence of an MMU. If the MMU is present
then link into the AllocMem() call code to mark which memory is used
by tasks and which memory is not. This much would be really quite
simple. Obviously there are still little points like making sure the
8 bytes of unused (or free) memory is not protected so that the EXEC
can maintain its FreeMem list, etc.
Now, such an implementation would only protect against accesses to
unused memory space. Of course, a task referenceing another task's
memory space would still be possible whether it is "valid" or not.
Furthermore, task resource tracking is by no means possible under
such a sceme (read CBM employee' comments in previous posts to see
why). Note also that I havn't addressed what would happen if an
illegal memory access happens other than to say it probably implies
a GURU.
However, I do beleive such a scheme would provide some extra
protection against runaway tasks. Certainly not total protection, but
a little more than is presently available at little extra coding.
Michael Cree "Gold there is
cree@elec.canterbury.ac.nz and rubies in abundance;
but lips that speak knowledge
are a rare jewel." Prov 20:15xgr39@CCVAX.IASTATE.EDU (Marc Barrett) (04/06/91)
In article <40924@cup.portal.com>, FelineGrace@cup.portal.com (Dana B Bourgeois) writes: >Good 'Ol -MB- wants ta know why CBM hasn't improved the chips yet. And >Randall tells him he's too impatient. Well, we all know the *real* >reason, don't we? It's because -MB- hasn't COMPLAINED enough yet. CBM >isn't gonna get serious about improving the old design until -MB- makes >them uncomfortable enough. They'll just tweak it here and there to >patronize him when he gets up on his soapbox so he'll leave 'em alone >until his blood pressure builds up the next time. So come on -MB-, >let 'em have it! Bitch! Moan! Complain! Turn up the Heat! Push that >design, lift that test suite. You gotta make the 'chippies' at CBM >nauseaous to the point they work 24 hours a day and get the new chip set >done. Fini. Finished. Completed. So get going -MB- and Stick It To >CBM but be Serious this time, OK? Stay on that soapbox until the new >chip set is done, (or until an unruly mob of crazed, tired, over-worked, >nauseaous former CBM chip designers hunts you down and rips you into >Little Tiny Pieces.) I can tell that you totally misunderstand me. My intention is not to pressure the 'chippies' at Commodore to work harder until the design is done. With R&D investments so low, they are probably far too overworked as it is. Rather, my intention is to pressure the Commodore management to fork over more investment capital for R&D, so that Commodore can hire more chippies, relieve the pressure on the ones who are already working, and give them all the best engineering hardware development equipment that is available. Commodore has the capital, but their management just doesn't use it. For the fiscal year 1990, Commodore's profits were down, but they still had well over $90 Million that they could have used to increase R&D had they wanted to. For the fiscal year of 1991, Commodore should have even more investment capital available. Therefore, there are no excuses for the R&D investments per year to be on the level of 2% of net sales. What I would like to know is, what dire evils would come to Commodore's engineering department if I had my way? My flames are not directed at anyone in the engineering department, as it is my oppinion that they are all some of the best engineers in the computer industry today. Rather, my flames are directed only at Commodore's management, which does not give them the funds they need to get their projects done quickly. Unfortunately, Commodore's management does not read CSA, and the engineers who do read it have to put up with the brunt of my flames. I can really hardly blame them for not liking me, though I think if they stepped back and took a look at what I am trying to accomplish, they would see that my intention is not to make them feel bad but to give them long-term job security and make them all filthy rich, (even at the possible expense of my own personal future). For this reason, I intend to start trying to more directly influence the heart of the problem -- the management -- though I cannot yet say precisely how. > >Tongue firmly in cheek I remain, >Dana Bourgeois @ cup.portal.com -MB- ---------------------------------------------------------- / Marc Barrett | BITNET: XGR39@ISUVAX.BITNET / / ISU COM S Student | Internet: XGR39@CCVAX.IASTATE.EDU / ----------------------------------------------------------
drichard@jarthur.Claremont.EDU (David Richards) (04/06/91)
> Commodore has the capital, but their management just doesn't use >it. For the fiscal year 1990, Commodore's profits were down, but they >still had well over $90 Million that they could have used to increase >R&D had they wanted to. For the fiscal year of 1991, Commodore should >have even more investment capital available. First of all, I can't believe you have any experience managing a multi- national corporation, so I don't think you are in any position to intelligently comment on Commodore's use of capital. I know that I am not. I would prefer to have slower improvments than a bankrupt company. Remeber that Commodore bought Amiga, after the firm that was trying to innovate it went belly up and died! > Unfortunately, Commodore's management does not read CSA, and the >engineers who do read it have to put up with the brunt of my flames. >I can really hardly blame them for not liking me, though I think if >they stepped back and took a look at what I am trying to accomplish, >they would see that my intention is not to make them feel bad but >to give them long-term job security and make them all filthy rich, >(even at the possible expense of my own personal future). For this (Oh Please! aren't we being a little egocentric and melodramatic?) >reason, I intend to start trying to more directly influence the >heart of the problem -- the management -- though I cannot yet say >precisely how. Well I do read CSA and I am tired of reading complaints. Why don't you take all of your complaints and write them in a letter to the people you are ostensibly trying to influence. You wouldn't get nearly as much spleen out of it, but you wouldn't be annoying the people you're "trying to help" Finally, problems aren't solved by complaints, but by ideas. If you have a few ( and I'm talking about something a little more specific than spend more money on R&D) I'm quite sure Commodore will pay you for them. I guess what I am saying is put up or shut up! Dave drichards@hmcvax.claremont.edu
bruce@zuhause.MN.ORG (Bruce Albrecht) (04/08/91)
In article <1991Apr5.173422.6215@news.iastate.edu> xgr39@CCVAX.IASTATE.EDU (Marc Barrett) writes: > What I would like to know is, what dire evils would come to >Commodore's engineering department if I had my way? My flames are >not directed at anyone in the engineering department, as it is my >oppinion that they are all some of the best engineers in the computer >industry today. Rather, my flames are directed only at Commodore's >management, which does not give them the funds they need to get their >projects done quickly. First, the new Amigas would cost a lot more. For example, Dave Haynie says that the twice as fast chips would require faster DRAM than is typically available commercially (and ZIPs already are twice as expensive as Mac/IBM SIMMs) because of the timing issues. Second, the market for the high end machines is so much smaller than the low end machines because CBM has a low end customer base world-wide, and a very small customer base in the workstation area. Unless CBM has reasonable expectations of expanding in the high-end market, they should concentrate on what they can dominate, and provide an upgrade path that provides reasonable performance. -- bruce@zuhause.mn.org
daveh@cbmvax.commodore.com (Dave Haynie) (04/16/91)
In article <bruce.8214@zuhause.MN.ORG> bruce@zuhause.MN.ORG (Bruce Albrecht) writes: >In article <1991Apr5.173422.6215@news.iastate.edu> xgr39@CCVAX.IASTATE.EDU (Marc Barrett) writes: >(and ZIPs already are twice as expensive as Mac/IBM SIMMs) When I say "expensive", I'm thinking from the C= viewpoint, not the average Joe on the street. When you buy them in quantities, DIP, ZIP, or SIMM are all around the same price. Last I looked, actually, SIMM were the most expensive. When you buy memory, though, you have to consider who you're buying from. You are certainly not buying it from NEC or Toshiba, probably from Fred's Computer Jungle or somesuch place. If you're buying a part that's available everywhere, Fred can't charge you alot for it. And its easier for Fred to get in the first place. If you're buying a relatively unusual part (static column SIMMs are likely just as unusual as static column ZIPs, anyway, from Fred's point of view), you get charged more. So, back to the original discussion. A memory part that really costs us 2x, assuming it even exists in production volumes, could very well wind up costing you 4x or more, and you might not even be able to get it. So there are very good arguments, at present, against any system requirement for 50ns DRAM. -- Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: hazy BIX: hazy "That's me in the corner, that's me in the spotlight" -R.E.M.