The CIA and the Times have already done the endangering. It seems like the Times has a lot to answer for in this. How the heck did they get the source of a 'secret' CIA document in the first place? Ultimately it has to be someone in the CIA who is responsible for this foul up. It looks like at the very least they should review their procedures on controlling & releasing electronic versions of their documents. This isn't exactly a novel phenomenon. It also seems unlikely that it's the only pdf file out there from the master spies with this problem. I guess there will be some overtime getting worked in Langley right about now.
This is simply amazing. I think the Times accusations have more to do with covering their own ass than concern for their unfortunate victims.
No, Rambus would have to give a huge chunk of their company to Intel if RDRAM succeeds and Intel ships enough mobo chipsets with RDRAM support. So, since Rambus intends to charge more royalties for the SDRAM & DDR anyway they will probably be quite content to watch RDRAM wither on the vine now. The truth is that RDRAM can be faster for the moment if you have more channels (and you can with a simpler layout). It's been much maligned because reviewers have used a single RIMM, if it could have reached sufficient volume and there was a more objective and rational press reviewing it (ie. not that hotheaded German fool) end users might have seen it's merits. It's death was a self fulfilling prophecy.
Either way, the onset of DDR memory is going to be good for Rambus and might even let them keep 20% of their company that would have gone straight to Intel. Intel is the real loser in this debacle. The irony here is just exquisite. Intel hoses it's mobo strategy trying to stay cozy with Rambus but loses to chipsets with faster SDRAM and the prospect of DDR which turns out to be owned by Rambus all along, and in the end Intel misses out on the carrot it's been chasing after.
You are correct about the evidence supporting this theory, however the appropriate conclusion to draw from this is that we should be far more worried about volcanic eruptions than earth crossing asteroids. They seem much more likely. I'd wager that Krakatoa will grow and errupt again long before an equivalent impact event... if I thought I'd be around to collect.
This was Carl Segan's fear. The danger as he saw it was the ability to destroy the Earth using this technology. Something which even man cannot currently manage.
I think you underestimate the danger these rocks represent. You wouldn't deflect these things onto a city or country, they are continent busters, Earth crushers. These are the kind of objects you cannot get safe from. If one hit the USA there would be no living creatures left on the continent and probably no humans alive anywhere a short time later.
You are also right that smaller rocks represent a danger too, the big question is will it be easier to deflect such a rock at Earth than to use other unconventional weapons. How easy will it be in future to go up there paint a rock black and nudge it at the Earth, perhaps to hit a city of your choice in a few years time.
The probability that any of these is going to hit a few months after we find it is vanishingly small.
We're all here because the chance of anything large hitting us is so remote that no extinction class impact events have happened in MANY millions of years. We can take our time about finding these, maybe a few centuries before we should really consider ourselves tardy.
As for the prospects of deflecting an orbiting body, it's much more likely that we'll have decades of forewarning rather than months for a regular Earth crossing orbit. That should give us more time to plan and more opportunity to act early. We might not have much warning for a cometary impact but unless it's an old one it should be easy to spot and maybe easier to deflect through heating.
As for other deflection schemes, there seem to be several viable alternatives. The solar sail option sounds promising and NASA is about to test the technology. Using anything up there you can find as drive mass for the nuke also seems like something you would consider. With all those earth crossing bodies it might be possible in a few decates to rendevous with one at closest approach to the threat and mine some mass for use as projectile matter for your nuke. It isn't so far fetched when you consider that NEAR is in orbit around such an asteroid.
Given the likely forewarning (not the Hollywood timescale) a NEAR like scouting missions to the threat and other objects followed by a flotilla of missions intended to deflect the threat seems like the most likely scenario.
I doubt we'll ever see this type of mission for thousands or even millions of years. We should probably be more worried about smaller more frequent climate threatening objects, and we should be able to do more about those although they are harder to spot. I guess they're not as sexy as big Earth busters but they're still extremely threatening and vastly more likely to happen. These might be frequent enough to merit a sense of urgency in searching for them.
This is the same idea presented in technical terms
on
Cleartype In Depth
·
· Score: 1
I approached this paper hoping that there would be some interesting innovation which the press had missed in the scramble to accuse M$ of stealing the Apple/Woz original idea.
Unfortunately the only point of this paper is to describe the displaced filter for each of the color components to correlate with the displaced color components on the display device, this is exactly what has been announced before. There is really nothing noteworthy about this other than to describe in detail how unoriginal the idea is.
So if you split it horizontally who writes the next version of office suite? Who writes the next version of Windows? Are they allowed to 'collaborate' on standards & software engineering projects? Which of the Baby Bills do Gates and Balmer work for?
This won't work. 7 years out there would probably be one left standing anyway and a factional mess between now and then.
Ellison would relish the confusion, but let's face it, he has a different agenda to practically everyone else involved.
It's NOT fast. The bandwidth is limited by the network in addition to the encoding to & from GLX protocol at each end. If you're fill limited then you probably won't see a difference once you get going. If you have significant geometry to draw in immediate mode or the X server cannot hold all your resources (display lists & textures mainly) in memory then you will see a major performance degredation over a network.
No. The networked layer is only there when rendering indirectly. There is the notion of direct rendering to the hardware in OpenGL implementations where the OpenGL API is implemented but doesn't go through GLX.
That is what the 'D' in DRI stands for *Direct* Rendering Infrastructure. Unfortunately with large geometry loads the dispatch mechanism and kernel support becomes critical and begins to affect performance. This will become more significant with hardware T&L and large geometry loads of the future. Linux will need to improve to attain parity with windows drivers for hardware T&L cards on geometry intensive applications and fast image/texture downloads.
Libraries are generally speaking not GLP, they are LGPL and that means you can link to them WITHOUT open sourcing the code. The notion that on Linux you must open source because of the GPL restrictions is a popular myth.
Unfortunately this study didn't really reflect the real and significant performance differences between OS 3D support.
An ostensibly fill limited benchmark will not reveal any disparity beyond state issues in the hardware. The geometry performance is not going to be significantly different in a software implementation of T&L provided there is at least a moderate effort by the vendor to support instruction sets like SSE or 3DNow on Linux.
What will be really telling in future is the difference between dispatch mechanisms & kernel level support for graphics when we get some heavy duty T&L support and a benchmark which exercises it. This will become significant for many games and is already significant for 'serious' 3D applications.
You are simply not going to get anything near parity unless we can persuade Cox & Torvalds that 3D is important enough to distribute the kinds of kernel hacks needed for efficient 3D support. Right now they are intransigent and have even lambasted 3D experts like Jon Leech in public over kernel support issues. In addition we'll need to remain flexible about driver implementation frameworks we want to support. Something like the DRI is OK now but when you need to send 15 million triangles a second to the graphics card it's not going to cut it.
Unfortunately Linux currently does not give anything near parity with hardware T&L. Even with an excellent driver implementation and the best will in the world it cannot and it may not for some time to come.
Re:The Strange Case Of The Video Card Industry
on
Goodbye, Number Nine
·
· Score: 2
S3 has announced it's intention to sell (and may have already sold) it's graphics business to someone, probably VIA was the speculation last I looked, just check the press, it's official. They may own Diamond but Diamond never designed graphics chips, and infact with the reference designs produced by the chip vendors Diamond never did much of any graphics hardware design. It's is probably a healthier business to be in but it means they simply sell other folks designs with minor tweaks at best, often just software tweaks if that. Also it's only healthier if you're not shackled to a flagging chip vendor who can't design a competitive card for you to move through your sales channels.
S3 made a loss of $98 million in FY98 on declining revenues (halved from FY97), from FY96 profit of $42 million. It's not a pretty picture, and they are bailing on the graphics business.
There's no question that support by OEM's et.al. is key to any graphics chip company (which I was calling an IHV). It's their bread & butter. That's why the supply dries up so darned quickly when you fall behind in the performance stakes. They drop you like a ton of bricks when you fall behind, it's totally cutthroat and very non linear.
This failure to comply with Direct X statement is complete nonsense.
The failed to build a fast 3D texture board.
Even the Imaging 128 Series 2 supported Direct X & D3D (it didn't texture but that's a hardware deficiency), the trouble was they never managed to produce an interesting 3D card after 3D arrived. The supported DirectX and Direct 3D in every single card they made in the relevant time period. Direct X support was always there and was never the issue with Number Nine. Features and performance was the problem.
This ignorant revisionist claptrap is rather annoying.
Re:The Strange Case Of The Video Card Industry
on
Goodbye, Number Nine
·
· Score: 2
This isn't that strange. Your #9 analysis is very weak. The Imagine 128 Series2 was still a major card and it was PCI. It wasn't the fast bus which sealed their fate, it was 3D.
S3 were huge back then too (#1 in fact) and they are now out of the graphics card business, is the 4-5 year death spiral a new 'Moores law' for troubled card IHV's? This was another player who couldn't get enough 3D performance soon enough and rev it fast enough. They gave it a good fight but couldn't match the product turnaround and performance hikes of other IHV's.
The key in the card business seems to be win the big OEM deals, get DELL, Gateway et. al. to bundle your cards. #9 were too slow to build fast 3D cards, their ticket to ride was too little too late, and earlier cards couldn't texture. They missed the 3D boat and it killed them. This keeps happening in the PC space. When you miss like that the buyers are so fickle that almost all of your business dries up overnight and when that happens it's difficult to catch up, difficult to retain staff, difficult to hire people and difficult to fund new development. You're on a steep slippery slope.
This is still happening in PC land, IHV's are leap frogging each other regularly. The latest rule seems to be that you must produce a significant performance and/or feature boost every 6 months or lose lot's of ground. You've seen 3Dfx miss, and now their VSA 100 is nice but they're behind on T&L and are a little off the money on fill performance unless you throw a lot of parts at the problem and that sucks power and costs $$. Their hail mary play is to buy Gigapixel and try and come back with something interesting and competitive maybe in a year (maybe spoiling the other Gigapixel liscensees was added incentive). In the meantime they are probably praying that the tbuffer is interesting enough to tide them over.
There are also some really desperate schemes to try and scale performance, for example ATI's time multiplexed dual pipeline card. The demand for more performance is so great that you absolutely must turn a new product or wave your business goodbye overnight. It's easier to lay out a board than it is to design a new chip and if your chip can't scale well then you add a big assed FIFO on the front, add two parts, multiplex the video at the back and you're still in business. ATI mobility chips seem to be doing well (no real competition yet) but they've lost ground on the desktop. I don't see charisma producing enough performance to beat GeForce2 never mind the followon which might even beat charisma out the door in volume, and nVidia are doing a good job of stealing their feature thunder with some...err... marketing.
Matrox,... well I hope they have some super secret design ready to ship real soon or they might hit the same brick wall other victims have.
#9 will probably not be the last to go. Let's hope there are always enough players to keep things interesting and moving at a brisk pace. There's nothing worse than VP's and marketing droids in a monopoly setting prices and development schedules.
It does not use the firmware of the 320 with the unified memory architecture and PROM if that's what you mean. It uses a high end OEM PC mother board with AGP 4x support and a BIOS, SGI has done some work with the manufacturer improving the quality & reliability of the mobo.
Have you even looked at the price of a decent PC? Go configure a decent system at Dell or Gateway and see the price. Unless you're buying a 500MHz Celeron with a 17 inch monitor and crap graphics card you're going to be paying $2-3k easily.
You're not making sense. This system isn't over priced you seem to agree. You're then saying something about other systems which are (in your opinion) over priced, but that doesn't make THIS system over priced.
I'll deal with the Intel/Linux comments first, I don't think you understand the factors involved in SGI's circumstances. To say Intel smoked SGI shows a profound ignorance, or at best trivialization of events. On the Linux issue, SGI would be in a much worse position without Linux. Linux is not a nail in SGI's coffin. It is infact the reverse. Linux is the last hope for the technical community to have a decent functional operating system and for SGI it represents a viable alternative which in future might scale to support architectures like the Origin and meet it's customers needs.
SGI is at least making a contribution to Linux Open Source, not just riding the wave. SGI's commitment to Linux is not new, this is just the first graphics machine and SGI has been doing a lot of work getting it right, not just slapping a few boards together and using whatever software they can download from the OS community.
Knowing a fair old bit about graphics I can tell you that your PC won't smoke an Onyx2, and certainly not by 50X. Maybe some day, but not yet, you can't get near the pixel fill rate with the antialiasing you need to beat an Onyx2. An Origin2000 has no graphics and the point of that system is it scales well and has memory bandwidth up the wazoo, again you can't get close with a PC.
It's a high end PC with fast PIII 128 MB RAM and GeForce graphics with AGP 4X + very high quality drivers and support. You can't get a better Linux system for significantly less.
It looks blue & gray to me, not purple, (just incase you buy one for the color and are disappointed.
"The Arecibo telescope underwent major upgrades in the 1990s, which dramatically improved its sensitivity and made it feasible to image more distant objects."
It has nothing to do with FSF or RMS. If the developer let's it slide that's his business, if he doesn't then they're breaking the GPL, even if FSF, RMS and Tux the penguin disagrees. FYI they won't see any of the margin above MSRP, that's supply & demand on the distribution side. i.e. probably the retailer sticking it to you. Any IHV will only see a small fraction of the MSRP and with OEM's making most of the boards it's probably a VERY small fraction.
Oh, and lest we all forget Linux is still insignificant in terms of card volumes right now.
You don't mistakenly lift code from someone else's work and forget that it was GPL'd. How can a commercial developer claim to have done this by mistake? At the very least, even if they intended to replace the code they've derived their work at a fundamental level from someone elses copyright code.
Leaving this driver downloadable without releasing the module source code should not be acceptable. It is in breach of the GPL, the honorable thing to do is pull the drivers or release the code. In fact now that the binary is out there, releasing the code isn't optional, it's a requirement under the GPL unless the original author waives the right(if they can). That's exactly the point of the GPL. Using the GPL'd code as a stopgap until you replace it is also against the spirit and letter of the license. It would be almost impossible for nVidia to demonstrate that their replacement code was not derived from or influenced by the design of the original work. They can't exactly claim to have the reverse engineered the replacement code now that the original is written into their work.
Release the code! It shouldn't be a request, it should be an unequivocal demand.
There's another issue here. Namely commercial developers trawling through open source code, seeing how it's done then implementing it in a proprietary binary, even if they don't directly lift the code. Is this acceptable reverse engineering or is it theft? Clearly nVidia has crossed the line with a few lines of code but what about the rest of the code involved?
You wrote:
___
Sun and SGI may not offer source either, but at least their stuff is FAST and HIGH QUALITY. If SGI ships me a binary IRIX driver for (let's say) an Infinite Reality Engine, I'm pretty damn sure that it'll work and not crash my system. Do you really place the same level of faith in nVidia? Coding for a system they most likely don't understand well? Sure, kid.
___
I Reply:
These drivers from nVidia are written by SGI, and the GeForce drivers are FAST and are High Quality, although they are currently only Beta. Get them and try them before shooting your mouth off. You won't find a faster OpenGL implementation on Linux anywhere.
That said, even SGI implementations have been known to have bugs, so the real issue is support. Who is going to support you in the absence of a contract and no OS community. You're going to want to have purchased a system from a vendor who supports Linux if that's your bag. But you don't know what nVidia is going to offer until the real drivers MR, in the mean time try them and marvel at their performance, know that you have the fastest Linux graphics system available.
Slashdot Mirror
The CIA and the Times have already done the endangering. It seems like the Times has a lot to answer for in this. How the heck did they get the source of a 'secret' CIA document in the first place? Ultimately it has to be someone in the CIA who is responsible for this foul up. It looks like at the very least they should review their procedures on controlling & releasing electronic versions of their documents. This isn't exactly a novel phenomenon. It also seems unlikely that it's the only pdf file out there from the master spies with this problem. I guess there will be some overtime getting worked in Langley right about now.
This is simply amazing. I think the Times accusations have more to do with covering their own ass than concern for their unfortunate victims.
No, Rambus would have to give a huge chunk of their company to Intel if RDRAM succeeds and Intel ships enough mobo chipsets with RDRAM support. So, since Rambus intends to charge more royalties for the SDRAM & DDR anyway they will probably be quite content to watch RDRAM wither on the vine now. The truth is that RDRAM can be faster for the moment if you have more channels (and you can with a simpler layout). It's been much maligned because reviewers have used a single RIMM, if it could have reached sufficient volume and there was a more objective and rational press reviewing it (ie. not that hotheaded German fool) end users might have seen it's merits. It's death was a self fulfilling prophecy.
Either way, the onset of DDR memory is going to be good for Rambus and might even let them keep 20% of their company that would have gone straight to Intel. Intel is the real loser in this debacle. The irony here is just exquisite. Intel hoses it's mobo strategy trying to stay cozy with Rambus but loses to chipsets with faster SDRAM and the prospect of DDR which turns out to be owned by Rambus all along, and in the end Intel misses out on the carrot it's been chasing after.
You are correct about the evidence supporting this theory, however the appropriate conclusion to draw from this is that we should be far more worried about volcanic eruptions than earth crossing asteroids. They seem much more likely. I'd wager that Krakatoa will grow and errupt again long before an equivalent impact event... if I thought I'd be around to collect.
This was Carl Segan's fear. The danger as he saw it was the ability to destroy the Earth using this technology. Something which even man cannot currently manage.
I think you underestimate the danger these rocks represent. You wouldn't deflect these things onto a city or country, they are continent busters, Earth crushers. These are the kind of objects you cannot get safe from. If one hit the USA there would be no living creatures left on the continent and probably no humans alive anywhere a short time later.
You are also right that smaller rocks represent a danger too, the big question is will it be easier to deflect such a rock at Earth than to use other unconventional weapons. How easy will it be in future to go up there paint a rock black and nudge it at the Earth, perhaps to hit a city of your choice in a few years time.
The probability that any of these is going to hit a few months after we find it is vanishingly small.
We're all here because the chance of anything large hitting us is so remote that no extinction class impact events have happened in MANY millions of years. We can take our time about finding these, maybe a few centuries before we should really consider ourselves tardy.
As for the prospects of deflecting an orbiting body, it's much more likely that we'll have decades of forewarning rather than months for a regular Earth crossing orbit. That should give us more time to plan and more opportunity to act early. We might not have much warning for a cometary impact but unless it's an old one it should be easy to spot and maybe easier to deflect through heating.
As for other deflection schemes, there seem to be several viable alternatives. The solar sail option sounds promising and NASA is about to test the technology. Using anything up there you can find as drive mass for the nuke also seems like something you would consider. With all those earth crossing bodies it might be possible in a few decates to rendevous with one at closest approach to the threat and mine some mass for use as projectile matter for your nuke. It isn't so far fetched when you consider that NEAR is in orbit around such an asteroid.
Given the likely forewarning (not the Hollywood timescale) a NEAR like scouting missions to the threat and other objects followed by a flotilla of missions intended to deflect the threat seems like the most likely scenario.
I doubt we'll ever see this type of mission for thousands or even millions of years. We should probably be more worried about smaller more frequent climate threatening objects, and we should be able to do more about those although they are harder to spot. I guess they're not as sexy as big Earth busters but they're still extremely threatening and vastly more likely to happen. These might be frequent enough to merit a sense of urgency in searching for them.
I approached this paper hoping that there would be some interesting innovation which the press had missed in the scramble to accuse M$ of stealing the Apple/Woz original idea.
Unfortunately the only point of this paper is to describe the displaced filter for each of the color components to correlate with the displaced color components on the display device, this is exactly what has been announced before. There is really nothing noteworthy about this other than to describe in detail how unoriginal the idea is.
So if you split it horizontally who writes the next version of office suite? Who writes the next version of Windows? Are they allowed to 'collaborate' on standards & software engineering projects? Which of the Baby Bills do Gates and Balmer work for?
This won't work. 7 years out there would probably be one left standing anyway and a factional mess between now and then.
Ellison would relish the confusion, but let's face it, he has a different agenda to practically everyone else involved.
It's NOT fast. The bandwidth is limited by the network in addition to the encoding to & from GLX protocol at each end. If you're fill limited then you probably won't see a difference once you get going. If you have significant geometry to draw in immediate mode or the X server cannot hold all your resources (display lists & textures mainly) in memory then you will see a major performance degredation over a network.
No.
The networked layer is only there when rendering indirectly. There is the notion of direct rendering to the hardware in OpenGL implementations where the OpenGL API is implemented but doesn't go through GLX.
That is what the 'D' in DRI stands for *Direct* Rendering Infrastructure. Unfortunately with large geometry loads the dispatch mechanism and kernel support becomes critical and begins to affect performance. This will become more significant with hardware T&L and large geometry loads of the future. Linux will need to improve to attain parity with windows drivers for hardware T&L cards on geometry intensive applications and fast image/texture downloads.
Libraries are generally speaking not GLP, they are LGPL and that means you can link to them WITHOUT open sourcing the code. The notion that on Linux you must open source because of the GPL restrictions is a popular myth.
Unfortunately this study didn't really reflect the real and significant performance differences between OS 3D support.
An ostensibly fill limited benchmark will not reveal any disparity beyond state issues in the hardware. The geometry performance is not going to be significantly different in a software implementation of T&L provided there is at least a moderate effort by the vendor to support instruction sets like SSE or 3DNow on Linux.
What will be really telling in future is the difference between dispatch mechanisms & kernel level support for graphics when we get some heavy duty T&L support and a benchmark which exercises it. This will become significant for many games and is already significant for 'serious' 3D applications.
You are simply not going to get anything near parity unless we can persuade Cox & Torvalds that 3D is important enough to distribute the kinds of kernel hacks needed for efficient 3D support. Right now they are intransigent and have even lambasted 3D experts like Jon Leech in public over kernel support issues. In addition we'll need to remain flexible about driver implementation frameworks we want to support. Something like the DRI is OK now but when you need to send 15 million triangles a second to the graphics card it's not going to cut it.
Unfortunately Linux currently does not give anything near parity with hardware T&L. Even with an excellent driver implementation and the best will in the world it cannot and it may not for some time to come.
S3 has announced it's intention to sell (and may have already sold) it's graphics business to someone, probably VIA was the speculation last I looked, just check the press, it's official. They may own Diamond but Diamond never designed graphics chips, and infact with the reference designs produced by the chip vendors Diamond never did much of any graphics hardware design. It's is probably a healthier business to be in but it means they simply sell other folks designs with minor tweaks at best, often just software tweaks if that. Also it's only healthier if you're not shackled to a flagging chip vendor who can't design a competitive card for you to move through your sales channels.
S3 made a loss of $98 million in FY98 on declining revenues (halved from FY97), from FY96 profit of $42 million. It's not a pretty picture, and they are bailing on the graphics business.
There's no question that support by OEM's et.al. is key to any graphics chip company (which I was calling an IHV). It's their bread & butter. That's why the supply dries up so darned quickly when you fall behind in the performance stakes. They drop you like a ton of bricks when you fall behind, it's totally cutthroat and very non linear.
This failure to comply with Direct X statement is complete nonsense.
The failed to build a fast 3D texture board.
Even the Imaging 128 Series 2 supported Direct X & D3D (it didn't texture but that's a hardware deficiency), the trouble was they never managed to produce an interesting 3D card after 3D arrived. The supported DirectX and Direct 3D in every single card they made in the relevant time period. Direct X support was always there and was never the issue with Number Nine. Features and performance was the problem.
This ignorant revisionist claptrap is rather annoying.
This isn't that strange. Your #9 analysis is very weak. The Imagine 128 Series2 was still a major card and it was PCI. It wasn't the fast bus which sealed their fate, it was 3D.
...err... marketing.
S3 were huge back then too (#1 in fact) and they are now out of the graphics card business, is the 4-5 year death spiral a new 'Moores law' for troubled card IHV's? This was another player who couldn't get enough 3D performance soon enough and rev it fast enough. They gave it a good fight but couldn't match the product turnaround and performance hikes of other IHV's.
The key in the card business seems to be win the big OEM deals, get DELL, Gateway et. al. to bundle your cards. #9 were too slow to build fast 3D cards, their ticket to ride was too little too late, and earlier cards couldn't texture. They missed the 3D boat and it killed them. This keeps happening in the PC space. When you miss like that the buyers are so fickle that almost all of your business dries up overnight and when that happens it's difficult to catch up, difficult to retain staff, difficult to hire people and difficult to fund new development. You're on a steep slippery slope.
This is still happening in PC land, IHV's are leap frogging each other regularly. The latest rule seems to be that you must produce a significant performance and/or feature boost every 6 months or lose lot's of ground. You've seen 3Dfx miss, and now their VSA 100 is nice but they're behind on T&L and are a little off the money on fill performance unless you throw a lot of parts at the problem and that sucks power and costs $$. Their hail mary play is to buy Gigapixel and try and come back with something interesting and competitive maybe in a year (maybe spoiling the other Gigapixel liscensees was added incentive). In the meantime they are probably praying that the tbuffer is interesting enough to tide them over.
There are also some really desperate schemes to try and scale performance, for example ATI's time multiplexed dual pipeline card. The demand for more performance is so great that you absolutely must turn a new product or wave your business goodbye overnight. It's easier to lay out a board than it is to design a new chip and if your chip can't scale well then you add a big assed FIFO on the front, add two parts, multiplex the video at the back and you're still in business. ATI mobility chips seem to be doing well (no real competition yet) but they've lost ground on the desktop. I don't see charisma producing enough performance to beat GeForce2 never mind the followon which might even beat charisma out the door in volume, and nVidia are doing a good job of stealing their feature thunder with some
Matrox,... well I hope they have some super secret design ready to ship real soon or they might hit the same brick wall other victims have.
#9 will probably not be the last to go. Let's hope there are always enough players to keep things interesting and moving at a brisk pace. There's nothing worse than VP's and marketing droids in a monopoly setting prices and development schedules.
It does not use the firmware of the 320 with the unified memory architecture and PROM if that's what you mean. It uses a high end OEM PC mother board with AGP 4x support and a BIOS, SGI has done some work with the manufacturer improving the quality & reliability of the mobo.
Have you even looked at the price of a decent PC? Go configure a decent system at Dell or Gateway and see the price. Unless you're buying a 500MHz Celeron with a 17 inch monitor and crap graphics card you're going to be paying $2-3k easily.
You're not making sense. This system isn't over priced you seem to agree. You're then saying something about other systems which are (in your opinion) over priced, but that doesn't make THIS system over priced.
No, it's blue, see the web site or an optician. I too have seen one.
I'll deal with the Intel/Linux comments first, I don't think you understand the factors involved in SGI's circumstances. To say Intel smoked SGI shows a profound ignorance, or at best trivialization of events. On the Linux issue, SGI would be in a much worse position without Linux. Linux is not a nail in SGI's coffin. It is infact the reverse. Linux is the last hope for the technical community to have a decent functional operating system and for SGI it represents a viable alternative which in future might scale to support architectures like the Origin and meet it's customers needs.
SGI is at least making a contribution to Linux Open Source, not just riding the wave. SGI's commitment to Linux is not new, this is just the first graphics machine and SGI has been doing a lot of work getting it right, not just slapping a few boards together and using whatever software they can download from the OS community.
Knowing a fair old bit about graphics I can tell you that your PC won't smoke an Onyx2, and certainly not by 50X. Maybe some day, but not yet, you can't get near the pixel fill rate with the antialiasing you need to beat an Onyx2.
An Origin2000 has no graphics and the point of that system is it scales well and has memory bandwidth up the wazoo, again you can't get close with a PC.
SGI does not have the legal option of releasing the driver source for these boards.
How can you say this is over priced?
It's a high end PC with fast PIII 128 MB RAM and GeForce graphics with AGP 4X + very high quality drivers and support. You can't get a better Linux system for significantly less.
It looks blue & gray to me, not purple, (just incase you buy one for the color and are disappointed.
If you follow the JPL links you can read:
"The Arecibo telescope underwent major upgrades in the 1990s, which dramatically improved its sensitivity and made it feasible to image more distant objects."
It has nothing to do with FSF or RMS. If the developer let's it slide that's his business, if he doesn't then they're breaking the GPL, even if FSF, RMS and Tux the penguin disagrees.
FYI they won't see any of the margin above MSRP, that's supply & demand on the distribution side. i.e. probably the retailer sticking it to you. Any IHV will only see a small fraction of the MSRP and with OEM's making most of the boards it's probably a VERY small fraction.
Oh, and lest we all forget Linux is still insignificant in terms of card volumes right now.
You don't mistakenly lift code from someone else's work and forget that it was GPL'd. How can a commercial developer claim to have done this by mistake? At the very least, even if they intended to replace the code they've derived their work at a fundamental level from someone elses copyright code.
Leaving this driver downloadable without releasing the module source code should not be acceptable. It is in breach of the GPL, the honorable thing to do is pull the drivers or release the code. In fact now that the binary is out there, releasing the code isn't optional, it's a requirement under the GPL unless the original author waives the right(if they can). That's exactly the point of the GPL. Using the GPL'd code as a stopgap until you replace it is also against the spirit and letter of the license. It would be almost impossible for nVidia to demonstrate that their replacement code was not derived from or influenced by the design of the original work. They can't exactly claim to have the reverse engineered the replacement code now that the original is written into their work.
Release the code! It shouldn't be a request, it should be an unequivocal demand.
There's another issue here. Namely commercial developers trawling through open source code, seeing how it's done then implementing it in a proprietary binary, even if they don't directly lift the code. Is this acceptable reverse engineering or is it theft? Clearly nVidia has crossed the line with a few lines of code but what about the rest of the code involved?
You wrote:
___
Sun and SGI may not offer source either, but at least their stuff is FAST and HIGH QUALITY. If SGI ships me a binary IRIX driver for (let's say) an Infinite Reality Engine, I'm pretty damn sure that it'll work and not crash my system. Do you really place the same level of faith in nVidia? Coding for a system they most likely don't understand well? Sure, kid.
___
I Reply:
These drivers from nVidia are written by SGI, and the GeForce drivers are FAST and are High Quality, although they are currently only Beta. Get them and try them before shooting your mouth off. You won't find a faster OpenGL implementation on Linux anywhere.
That said, even SGI implementations have been known to have bugs, so the real issue is support. Who is going to support you in the absence of a contract and no OS community. You're going to want to have purchased a system from a vendor who supports Linux if that's your bag. But you don't know what nVidia is going to offer until the real drivers MR, in the mean time try them and marvel at their performance, know that you have the fastest Linux graphics system available.