Except that so very little of programming these days is about algorithms. Rather, it is about elegantly solving businesses problems and to know one's way around huge frameworks.
And that attitude illustrates exactly what leads to big, crap and unreliable systems. Coders who do not understand that even huge frameworks... especially big frameworks.... are just big algorithms. Which does not need to be explained to a true wizard. Foo, you are no wizard;)
I wasn't talking about people writing drivers. We certainly do not do that by decompiling code, though running the driver to see what it outputs to the card is fair game.
I was talking about the highly theoretical possibility, which to put it bluntly, I regard as a pile of hooey, that NVidia refuses to open their hardware because the driver is doing something for which they might have to pay patent fees. Did I say hooey? I'll say it again. Hooey. Whoever advanced that theory, please do not insult my intelligence. Shipping binary drivers in no way protects against discovery of patent infringement.
My personal theory about why NVidia does not open its hardware specs is that they simply have their corporate head shoved so far up their collective ass it goes around twice.
Broadcom is a highly profitable and successful company. You could argue what they're doing is financially beneficial.
The argument "we must be doing it right because we're making money" works great right up until the point you realize somebody did it better and passed you by.
I think you kind of missed the point. Whoosh. The GP trotted out the tired old claim that NVidia doesn't open the source because it would reveal patent infringement. Which is a load of hooey because any interested patent owner can trivially decompile it.
And sorry, if it looks like a big ugly mess of assembler to you, that's your problem.
People keep spouting these things without actually taking into account that NVIDIA most likely has all sorts of contracts and license agreements they just can't break.
That sounds like a load of selfserving bullshit to me.
The big deal is the army of patent trolls who will jump out of the bushes with baseball bats the minute NVIDIA or any other major GPU manufacturer publishes a full set of register-level specifications.
What a load of hooey. If it was worth anything they would have decompiled the NVidia drivers long ago. I don't know about you, but that would take me 5 minutes. Troll.
Either NV hands over the goods, or we have nothing
Crawl back into your cesspool, idiot. Like GP said, we need register specs. That's it. NVidia no doubt has a lot of years invested in their driver, but they can keep it. We have way more man years available than they do.
I'm favorably impressed with Intel's latest GPU efforts, but cheapo Radeon cards and embedded still clean their clock and obviously high end cards are completely in another league.
Look, you're rambling. You keep switching between driver performance and card performance. The point is, the open source Radeon driver performance is pretty impressive, contrary to the FUD going around. Sure, it could be better, but it's already not bad, admit it.
Fill rate is just not one of the things that varies a lot depending on driver quality. Triangle setup is. That's why I talked about triangles, and mentioned Phong just so you know the driver is actually doing some work. But apparently you don't know that and you just want to dump on everything.
As far as your Sandy bridge graphics goes.... have you even thought about the latency difference between onboard GDDR5 and Sandy Bridge DDR3 over the memory bus? Right, I suspected as much. And that's just the beginning.
And by the way, fill rate is not an issue with this cheapo Radeon. Anisotrophic filtering bothers it a little at 1920x1200, but not enough so that I turn it off. I can imagine that crappy engines with tons of overdraw or really rambling shaders could be an issue, but I don't do stupid things like that and whoever does deserves to suffer for it.
(Phong shaded triangles do not use textures. I suspect that despite the big number of them you get per second, your actual framerate is fairly low).
For your information, Phong shading is quite demanding because of the exponentials involved. No my framerate is not low, it is capped at 60 FPS. Very impressive for this class of card. And at least a single layer of textures does not seem to bother it. See, I'm not throwing crap at the card, I'm using it the way it was meant to be used, that makes quite a difference. Now please crawl back into your hole fudster, and don't come out again until you are armed with some facts.
I'm getting 75 million Phong shaded triangles/second at 1920x1200 out of a 6450 running the Xorg Radeon driver. What's not good about that? Note: that's a fanless $50 card.
Slashdot Mirror
All of computer science is just an exercise in caching, don't you know.
Except that so very little of programming these days is about algorithms. Rather, it is about elegantly solving businesses problems and to know one's way around huge frameworks.
And that attitude illustrates exactly what leads to big, crap and unreliable systems. Coders who do not understand that even huge frameworks... especially big frameworks.... are just big algorithms. Which does not need to be explained to a true wizard. Foo, you are no wizard ;)
Maybe "Zunetab" was already taken.
Enough of a taste to scare the crap out of OEMs before the Windows 8 launch...
Looks to me like Microsoft is all set to scare the crap out of themselves.
I wasn't talking about people writing drivers. We certainly do not do that by decompiling code, though running the driver to see what it outputs to the card is fair game.
I was talking about the highly theoretical possibility, which to put it bluntly, I regard as a pile of hooey, that NVidia refuses to open their hardware because the driver is doing something for which they might have to pay patent fees. Did I say hooey? I'll say it again. Hooey. Whoever advanced that theory, please do not insult my intelligence. Shipping binary drivers in no way protects against discovery of patent infringement.
My personal theory about why NVidia does not open its hardware specs is that they simply have their corporate head shoved so far up their collective ass it goes around twice.
"Fair" is newspeak for "cartel".
Broadcom is a highly profitable and successful company. You could argue what they're doing is financially beneficial.
The argument "we must be doing it right because we're making money" works great right up until the point you realize somebody did it better and passed you by.
What you suggest would be tantamount to illegal market control. Of course Microsoft would never do such a thing.
And you also suggest that NVidia is a Microsoft whore. That may be. I don't need their hardware.
I think you kind of missed the point. Whoosh. The GP trotted out the tired old claim that NVidia doesn't open the source because it would reveal patent infringement. Which is a load of hooey because any interested patent owner can trivially decompile it.
And sorry, if it looks like a big ugly mess of assembler to you, that's your problem.
Haiku, what?
Consider how silly it is not to publish register information so you can write software for it. It is silly
It's stupid beyond belief. A firing offense in my opinion.
People keep spouting these things without actually taking into account that NVIDIA most likely has all sorts of contracts and license agreements they just can't break.
That sounds like a load of selfserving bullshit to me.
The big deal is the army of patent trolls who will jump out of the bushes with baseball bats the minute NVIDIA or any other major GPU manufacturer publishes a full set of register-level specifications.
What a load of hooey. If it was worth anything they would have decompiled the NVidia drivers long ago. I don't know about you, but that would take me 5 minutes. Troll.
you wont ever be in a position where you have the required brain capacity to write a graphics driver
I am, and you're a moron.
Either NV hands over the goods, or we have nothing
Crawl back into your cesspool, idiot. Like GP said, we need register specs. That's it. NVidia no doubt has a lot of years invested in their driver, but they can keep it. We have way more man years available than they do.
that's why my next computers will have Intel GPUs
I'm favorably impressed with Intel's latest GPU efforts, but cheapo Radeon cards and embedded still clean their clock and obviously high end cards are completely in another league.
for whatever reason, if their higher ups don't want to be telling the world how their drivers work, and so they won't.
That's nothing a good firing wouldn't fix.
Not really. What is needed is hardware that can be documented. Nice of nVidia to confess they are "not it". Spares having to consider them.
I've been NVidia-free for quite some time, and happier for it.
NVidia should open their register specs like AMD or else fuck themselves. I believe that is what Linus meant.
What packages properly handle the translation from AD permissions to the Linux level of permissions?
Samba4 is what you're looking for.
Look, you're rambling. You keep switching between driver performance and card performance. The point is, the open source Radeon driver performance is pretty impressive, contrary to the FUD going around. Sure, it could be better, but it's already not bad, admit it.
Fill rate is just not one of the things that varies a lot depending on driver quality. Triangle setup is. That's why I talked about triangles, and mentioned Phong just so you know the driver is actually doing some work. But apparently you don't know that and you just want to dump on everything.
As far as your Sandy bridge graphics goes.... have you even thought about the latency difference between onboard GDDR5 and Sandy Bridge DDR3 over the memory bus? Right, I suspected as much. And that's just the beginning.
And by the way, fill rate is not an issue with this cheapo Radeon. Anisotrophic filtering bothers it a little at 1920x1200, but not enough so that I turn it off. I can imagine that crappy engines with tons of overdraw or really rambling shaders could be an issue, but I don't do stupid things like that and whoever does deserves to suffer for it.
Libreoffice cares, that's who. This boneheaded move by Microsoft will be good for at least doubling the downloads.
(Phong shaded triangles do not use textures. I suspect that despite the big number of them you get per second, your actual framerate is fairly low).
For your information, Phong shading is quite demanding because of the exponentials involved. No my framerate is not low, it is capped at 60 FPS. Very impressive for this class of card. And at least a single layer of textures does not seem to bother it. See, I'm not throwing crap at the card, I'm using it the way it was meant to be used, that makes quite a difference. Now please crawl back into your hole fudster, and don't come out again until you are armed with some facts.
What is good about facts? Oh, never mind, I see you don't let any such thing stand in the way of a perfectly good rant.
I'm getting 75 million Phong shaded triangles/second at 1920x1200 out of a 6450 running the Xorg Radeon driver. What's not good about that? Note: that's a fanless $50 card.