No, the piont of running Linux isn't to "get away" from MS. If you're doing that, then your stupid. The point is to use a better product. Right now, MS Office is that better product. Very few people complained that WordPerfect for Linux was less than perfect due to LibWine, so why should they complain if MSs product has an OS layer around it? It can't possibly be more bloated than, say, star office, which has it's own DESKTOP!
This is slightly offtopic, but important none-the-less.
As most of you know, the GeForce2 Ultra Windows drivers took a jump to version 6 (detonator 3.) However, it seems that most of the performance improvements in Detonator 3 are already in the Linux OpenGL driver. That's a bad thing. Since it seems that the Detonator 3 drivers improve performance by 20% or so on average, it means that the benchmarks where Linux was running neck-and-neck with Windows were inaccurate. If benchmarking is done between equal platforms, Detonator3 and Linux OpenGL, it might show that Linux was still 20% or more behind the performance of the Windows drivers. Is anybody aware of such a test? Also can anybody back up the fact that the Linux OGL driver contain the tweeks in Detonator 3. My source was some guy on Slashdot, be he seemed to know what he was talking about.
Actually, the best way to describe it is OpenGL-retained mode. It is a library that sits on top of OpenGL and deals with stuff on the object level instead of the triangle level.
The other 3D drivers get updated more often, because they're Open Source. However, they shouldn't need to be updated more often. Saying, "okay, the driver API is stable as long as your product is Open Source and you update them whenever the API changes" is a crutch. It's a hack to solve a more fundemental problem. Perfect example: The glue layer between the NVIDIA driver and the kernel is Open Source. However, because the driver API has changed in 2.4-test6, that glue layer no longer compiles. If it was another OSS driver that depended on those particular features, those wouldn't compile either. As it is, driver that depend of parts of the driver API that change have to be updated whether or not they are Open Source. Thus, the problem lies not with the driver (which simply uses all the features of the driver API) but in the API itself (which changes the way those features are accessed.)
There WAS competition, the RivaTNT. That offered about the same performance as a Voodoo-whatever. Remember, this isn't NVIDIA's next main chip, it's the dual-voodoo2-SLI of the gaming world. As the benchmarks show, it really does kick ass. It means the difference between playable and unplayble frame-rates at 1600x1200. People will ALWAYS pay for that.
That's fine with me. I'm not particularly happy with 3DFx. The Voodoo3 series was decidedly low quality, they really don't give a hoot about OpenGL beyond the minimum required to run Quake III, and their windows drivers aren't always polished. Matrox is a good company, and they care about OpenGL, but their drivers are less than polished. Neither company offers the power and polish of NVIDIA. Their drivers are stable, fast, and pro-quality. I'm pissed that they don't support BeOS (although if you use BeOS, check out BeNews today. Apparently BeBits is involved in a top-secret project to bring GeForce drivers to BeOS.)
However, in the end, it's what's more imporant to you. I prefer getting the ultimate in performance, and don't mind rebooting to do it. You do mind rebooting and I can understand that.
SDL is pretty limited though. For some games, it's enough, but for a real commercial game it's not that great. (Yes I know real commercial games have used SDL.) For most platforms, SDL is just a wrapper for the native services with less features. That leads to lower performance, and a "least common denomenator" API. What is really needed is for some smart cookie to get a bunch of people together and work on a portable API like DirectX.
Of course, you have the problem that no code is truely portable if it takes full advantage of all the feature of the OS. For example, write anything to take advantage of ALSA, and you're limited to Linux. And it's not just people writing non-protable code. If you code just for POSIX, then you end up with a game that really doesn't take full advantage of the user's machine.
Look, I care more for quality than a peculiar attachment to a particular OS. Right now, OpenGL on all the other consumer cards sucks compared to NVIDIA's. They're the only ones I can use for 3D Studio. They also have the most stable drivers, and I'm benifeting from all their driver work even though I've got an older card. That's real quality.
Actuall according to another guy, DirectX8 is completely designed around/for NVIDIA card. NVIDIA is aweful big in pushing for the feautres they want during conferences, (DirectX is planned by hardware and software vendors talking about what features should be put in.) Given the fact that NVIDIA's cards totally dominate in terms of features, I wouldn't be surprised. I do remember, though, that DirectX6 looked awefully like a TNT.
Yes I'm pissed at NVIDIA for Be thing. But given that the company does so much for it's users in a day where videocard companies are generally down the tube, I forgive them for it.
NVIDIA is still toe-ing the XFree86 market. But they're doing it whole-heartedly, and the drivers have improved quite a bit.
As for the time it took to release, remember this. The specs for XFree86 weren't really set that early in the game. The driver ABI changed up until right before it's release. Also, an OpenGL driver is a complete implementation of OpenGL. Not only does it bang interrupts, but totally implements the entire pipeline. Functionally, it does just as much as Mesa, but is slightly easier to code because most functions don't need emultors. It's a wonder they got it out in the time they did. As a preempt to any OSS comments, remember, this isn't just a driver, it's an OpenGL implementation. It's also the highest quality GL implemenation on a consumer level card. It is simply too much to ask them to OSS it. They'd be giving away all the tweeks to the GL pipeline, not just register level info. Matrox is really hurting for a good OpenGL ICD. Is it really fair to ask NVIDIA to give them one?
Weird, I had no problem back in the Riva128 days. Of course, take a look back then. The main reason there were no incompatibility problems with Voodoo back then was because all games were designed with 3DFx in mind. Of course 3DFx had no problems because everyone tested on their card. Of course, take a look at what NVIDIA did back then. The brought DirectX6 support and OpenGL ICD's in, putting a big hurt on 3DFx's Glide monoply. Of course take a look at the now. NVIDIA drivers have been more or less rock solid since 3.x, they have great OpenGL support, and they release high quality products. Compare this to 3DFx that released the totally flaky Voodoo3 3500, that still doesn't have good OpenGL support, and still is in the back of the pack in terms of features. I'm suprised you didn't go all the way back to the NV1 and the load of crap that was. Of course, NVIDIA has made a huge turnaround. The Riva128 was their first real competitive card (and only their second commercial design.) With the TNT they experianced some growing pains early on, but ever since the driver issues cleared up (I got a TNT several months after they came out, and I've had no problems) and so far, the only problems they've had is the fact that their boards are so demanding of current that lower quality motherboards have problems.
About the Microsoft thing. What is wrong with writing code, protecting it, and making a profit? It is those kinds of feelings that get people to think of OSS people as communist. Not that YOU release your software for free, but don't want to give other people THEIR freedom to release THEIR software how THEY choose.
How are they screwing over the OSS community? The used some code. I'm sure a lot of people have used some code. They keep their drivers closed. I've actuall gained new insight into that one. An OpenGL driver is a full implementation of OpenGL. Not just a driver that bangs interrupts, but something that handles everything from glVertex on down. NVIDIA has THE highest quality OpenGL ICD in consumer space. In Windows-land Matrox and ATI are both struggling to get high quality OpenGL drivers. (Read the interview with Matrox's OpenGL guy in this month's MaximumPC.) Asking someone to give that away is simply too much.
Significantly, I mean you can go out and buy 128MB of RAM for that. Also, V5600 won't be out for awhile, and the price should go down as DDR RAM prices go down. (A big chunk of the increased price is the 230MHz DDR RAM)
I'm saying that quality and price ARE the only IMPROTANT factors. So is service, and support. NVIDIA aces all of these. I'm just saying that it is not in one's best interest to go support companies who really aren't as high quality over something like Open Source DRIVERS. You're free to do it, but just realize that by doing so you have no right to complain that computer companies these days don't care about product quality, and care even less for their users.
As for alternatives, you could buy ATI and live with crappy drivers, or buy Matrox and live with (comparativly) crappy performance and bad OpenGL support. Desicions, desicions.
Maybe I'm wrong, but...
If quality is what counts (you said they make quality products) then why is there a new card every 6
months or less? Also, why new drivers nearly every week?
There are two ways this could be viewed:
>>>>>>>>>>>>>>
Quality counts, but so does speed. Unlike most companies, NVIDIA manages to keep one from being exclusive of the other. NVIDIA cards are all high quality, but they are released often so that NVIDIA stays on top of the industry. Unlike 3DFx they are not resting on their asses waiting for everyone to catch up. As for new drivers, why is a new Linux kernel released every few weeks? Does that mean that the kernel isn't high quality? Of course not. NVIDIA continually improves the quality of their drivers. There are always tweeks you can implement, and NVIDIA is taking already high quality drivers and making them better. You're complaining about that?
It would seem to me that the wisest choice would be to release a solid card with solid drivers, rather
than a card plagued with incompatibility issues and drivers that constantly improve performance on
each release.
>>>>>>>>
Where is there any indication that NVIDIA's cards aren't solid? The NVIDIA drivers are rock solid. (Have you actually used the Windows ones?) They're all fast, stable, and very compatible. NVIDIA is one of the only cards a workstation 3D user could seriously consider using with apps like 3D Studio. As for compatiblity issues, they're in your head. There have always been problems with crappy hardware and NVIDIA's boards. There were problems with NVIDIA's boards and the early Super 7 ones. There were problems using them in lower-quality boards because of the large amount of current they use. The Via KX133 chip had problems with them. Of course, they were also incompatible with the Thunderbird Athlon's, so it seems that the fault lies with VIA. NVIDIA cards are big, and take a lot of current. If your motherboard manufactuer isn't building stuff to spec, then how can you possibly blame NVIDIA for compatiblity issues?
3DFx doesn't use AGP direct texture execution. This means that textures must be copied into texture RAM first before they can be used. In cards that support "real" AGP, textures can be used directly from system memory. It also lacks fast-writes and sidebanding, and essentially uses AGP as a 66MHz PCI slot.
AGP has no problem with multiple processors, it just can't handle multple devices. That's why Voodoo5 is on one card. However, AGP has no problem with multiple processors, as evidenced by the ATI Rage Fury MAXX, which uses multple Rage 128 Pro chips. It seems that 3DFx didn't engineer support for AGP texture execution into the VSA-100 chips. To 3DFx's credit, it really doesn't matter because in current games AGP makes less than a 1% differences. However, future games are going to be limited both by the lack of AGP texturing, and by the fact that even the 128MB Voodoo5 6000 only has 32MB of RAM (because each proc needs it's own copy of the textures.)
Actually, it won't be 10x more powerful than the PS2. More like twice as powerful. That's nothing. In the same situation, the original Playstation whooped Nintendo's ass even though the N64 was more than 3X faster, and featured stuff like texture filtering and full screen anti-aliasing (which PSX doesn't support.) Of course, Nintendo actually knew the console industry, and knew how to deliver a simple, stable, easy to use product. Microsoft is bringing too much PC garbage into the XBox to be able to do that. XBox will not be simple, stable, nor easy to use. It will require updates, and upgrades, and patches. Nothing against MS, it's just that it doesn't understand the industry. None of this takes into account the fact that Playstation 2 is coming out more than a year earlier. As history shows us, that's a killer. PSX whopped Nintendo largely due to the fact that N64 came out more than a year later. Lastly, Sony is a lot bigger than Microsoft, and they have more developer support and consumer mindshare. MS is going to be eaten for lunch.
Slashdot Mirror
Like all the security issues with BeOS? Or the fact that Chorus OS is just so highly insecure they decided to use it for nation-wide telecom?
No, the piont of running Linux isn't to "get away" from MS. If you're doing that, then your stupid. The point is to use a better product. Right now, MS Office is that better product. Very few people complained that WordPerfect for Linux was less than perfect due to LibWine, so why should they complain if MSs product has an OS layer around it? It can't possibly be more bloated than, say, star office, which has it's own DESKTOP!
What's up calling Windows bloated? My Win98 machine is a hell of a lot less bloated then my Linux +GNOME+KDE+X+Mozilla machine.
This is slightly offtopic, but important none-the-less.
As most of you know, the GeForce2 Ultra Windows drivers took a jump to version 6 (detonator 3.) However, it seems that most of the performance improvements in Detonator 3 are already in the Linux OpenGL driver. That's a bad thing. Since it seems that the Detonator 3 drivers improve performance by 20% or so on average, it means that the benchmarks where Linux was running neck-and-neck with Windows were inaccurate. If benchmarking is done between equal platforms, Detonator3 and Linux OpenGL, it might show that Linux was still 20% or more behind the performance of the Windows drivers. Is anybody aware of such a test? Also can anybody back up the fact that the Linux OGL driver contain the tweeks in Detonator 3. My source was some guy on Slashdot, be he seemed to know what he was talking about.
Actually, the best way to describe it is OpenGL-retained mode. It is a library that sits on top of OpenGL and deals with stuff on the object level instead of the triangle level.
Some addendums to the other guy:
If you're on Windows, check out www.strata3d.com
for a decent free, OpenGL accelerated modeler.
Also, check out www.gamedev.net. They've got some good message boards for OpenGL, and some good tutorials.
There are some good tutorials at http://nehe.gamedev.net/opengl.asp.
The other 3D drivers get updated more often, because they're Open Source. However, they shouldn't need to be updated more often. Saying, "okay, the driver API is stable as long as your product is Open Source and you update them whenever the API changes" is a crutch. It's a hack to solve a more fundemental problem. Perfect example: The glue layer between the NVIDIA driver and the kernel is Open Source. However, because the driver API has changed in 2.4-test6, that glue layer no longer compiles. If it was another OSS driver that depended on those particular features, those wouldn't compile either. As it is, driver that depend of parts of the driver API that change have to be updated whether or not they are Open Source. Thus, the problem lies not with the driver (which simply uses all the features of the driver API) but in the API itself (which changes the way those features are accessed.)
There WAS competition, the RivaTNT. That offered about the same performance as a Voodoo-whatever. Remember, this isn't NVIDIA's next main chip, it's the dual-voodoo2-SLI of the gaming world. As the benchmarks show, it really does kick ass. It means the difference between playable and unplayble frame-rates at 1600x1200. People will ALWAYS pay for that.
I don't think they can. MS is a tiny company compared to Sony.
That's fine with me. I'm not particularly happy with 3DFx. The Voodoo3 series was decidedly low quality, they really don't give a hoot about OpenGL beyond the minimum required to run Quake III, and their windows drivers aren't always polished. Matrox is a good company, and they care about OpenGL, but their drivers are less than polished. Neither company offers the power and polish of NVIDIA. Their drivers are stable, fast, and pro-quality. I'm pissed that they don't support BeOS (although if you use BeOS, check out BeNews today. Apparently BeBits is involved in a top-secret project to bring GeForce drivers to BeOS.)
However, in the end, it's what's more imporant to you. I prefer getting the ultimate in performance, and don't mind rebooting to do it. You do mind rebooting and I can understand that.
Please tell me you're joking.
SDL is pretty limited though. For some games, it's enough, but for a real commercial game it's not that great. (Yes I know real commercial games have used SDL.) For most platforms, SDL is just a wrapper for the native services with less features. That leads to lower performance, and a "least common denomenator" API. What is really needed is for some smart cookie to get a bunch of people together and work on a portable API like DirectX.
Of course, you have the problem that no code is truely portable if it takes full advantage of all the feature of the OS. For example, write anything to take advantage of ALSA, and you're limited to Linux. And it's not just people writing non-protable code. If you code just for POSIX, then you end up with a game that really doesn't take full advantage of the user's machine.
Look, I care more for quality than a peculiar attachment to a particular OS. Right now, OpenGL on all the other consumer cards sucks compared to NVIDIA's. They're the only ones I can use for 3D Studio. They also have the most stable drivers, and I'm benifeting from all their driver work even though I've got an older card. That's real quality.
Put blame where blame is due. Lambaste Linus and his desicion to not put in a stable driver API.
Actuall according to another guy, DirectX8 is completely designed around/for NVIDIA card. NVIDIA is aweful big in pushing for the feautres they want during conferences, (DirectX is planned by hardware and software vendors talking about what features should be put in.) Given the fact that NVIDIA's cards totally dominate in terms of features, I wouldn't be surprised. I do remember, though, that DirectX6 looked awefully like a TNT.
Of course Matrox has been working on OpenGL since the G200 and they are STILL half-baked. (Quake works, 3D Studio doesn't)
Yes I'm pissed at NVIDIA for Be thing. But given that the company does so much for it's users in a day where videocard companies are generally down the tube, I forgive them for it.
NVIDIA is still toe-ing the XFree86 market. But they're doing it whole-heartedly, and the drivers have improved quite a bit.
As for the time it took to release, remember this. The specs for XFree86 weren't really set that early in the game. The driver ABI changed up until right before it's release. Also, an OpenGL driver is a complete implementation of OpenGL. Not only does it bang interrupts, but totally implements the entire pipeline. Functionally, it does just as much as Mesa, but is slightly easier to code because most functions don't need emultors. It's a wonder they got it out in the time they did. As a preempt to any OSS comments, remember, this isn't just a driver, it's an OpenGL implementation. It's also the highest quality GL implemenation on a consumer level card. It is simply too much to ask them to OSS it. They'd be giving away all the tweeks to the GL pipeline, not just register level info. Matrox is really hurting for a good OpenGL ICD. Is it really fair to ask NVIDIA to give them one?
Weird, I had no problem back in the Riva128 days. Of course, take a look back then. The main reason there were no incompatibility problems with Voodoo back then was because all games were designed with 3DFx in mind. Of course 3DFx had no problems because everyone tested on their card. Of course, take a look at what NVIDIA did back then. The brought DirectX6 support and OpenGL ICD's in, putting a big hurt on 3DFx's Glide monoply. Of course take a look at the now. NVIDIA drivers have been more or less rock solid since 3.x, they have great OpenGL support, and they release high quality products. Compare this to 3DFx that released the totally flaky Voodoo3 3500, that still doesn't have good OpenGL support, and still is in the back of the pack in terms of features. I'm suprised you didn't go all the way back to the NV1 and the load of crap that was. Of course, NVIDIA has made a huge turnaround. The Riva128 was their first real competitive card (and only their second commercial design.) With the TNT they experianced some growing pains early on, but ever since the driver issues cleared up (I got a TNT several months after they came out, and I've had no problems) and so far, the only problems they've had is the fact that their boards are so demanding of current that lower quality motherboards have problems.
About the Microsoft thing. What is wrong with writing code, protecting it, and making a profit? It is those kinds of feelings that get people to think of OSS people as communist. Not that YOU release your software for free, but don't want to give other people THEIR freedom to release THEIR software how THEY choose.
How are they screwing over the OSS community? The used some code. I'm sure a lot of people have used some code. They keep their drivers closed. I've actuall gained new insight into that one. An OpenGL driver is a full implementation of OpenGL. Not just a driver that bangs interrupts, but something that handles everything from glVertex on down. NVIDIA has THE highest quality OpenGL ICD in consumer space. In Windows-land Matrox and ATI are both struggling to get high quality OpenGL drivers. (Read the interview with Matrox's OpenGL guy in this month's MaximumPC.) Asking someone to give that away is simply too much.
Significantly, I mean you can go out and buy 128MB of RAM for that. Also, V5600 won't be out for awhile, and the price should go down as DDR RAM prices go down. (A big chunk of the increased price is the 230MHz DDR RAM)
I'm saying that quality and price ARE the only IMPROTANT factors. So is service, and support. NVIDIA aces all of these. I'm just saying that it is not in one's best interest to go support companies who really aren't as high quality over something like Open Source DRIVERS. You're free to do it, but just realize that by doing so you have no right to complain that computer companies these days don't care about product quality, and care even less for their users.
As for alternatives, you could buy ATI and live with crappy drivers, or buy Matrox and live with (comparativly) crappy performance and bad OpenGL support. Desicions, desicions.
Maybe I'm wrong, but...
If quality is what counts (you said they make quality products) then why is there a new card every 6
months or less? Also, why new drivers nearly every week?
There are two ways this could be viewed:
>>>>>>>>>>>>>>
Quality counts, but so does speed. Unlike most companies, NVIDIA manages to keep one from being exclusive of the other. NVIDIA cards are all high quality, but they are released often so that NVIDIA stays on top of the industry. Unlike 3DFx they are not resting on their asses waiting for everyone to catch up. As for new drivers, why is a new Linux kernel released every few weeks? Does that mean that the kernel isn't high quality? Of course not. NVIDIA continually improves the quality of their drivers. There are always tweeks you can implement, and NVIDIA is taking already high quality drivers and making them better. You're complaining about that?
It would seem to me that the wisest choice would be to release a solid card with solid drivers, rather
than a card plagued with incompatibility issues and drivers that constantly improve performance on
each release.
>>>>>>>>
Where is there any indication that NVIDIA's cards aren't solid? The NVIDIA drivers are rock solid. (Have you actually used the Windows ones?) They're all fast, stable, and very compatible. NVIDIA is one of the only cards a workstation 3D user could seriously consider using with apps like 3D Studio. As for compatiblity issues, they're in your head. There have always been problems with crappy hardware and NVIDIA's boards. There were problems with NVIDIA's boards and the early Super 7 ones. There were problems using them in lower-quality boards because of the large amount of current they use. The Via KX133 chip had problems with them. Of course, they were also incompatible with the Thunderbird Athlon's, so it seems that the fault lies with VIA. NVIDIA cards are big, and take a lot of current. If your motherboard manufactuer isn't building stuff to spec, then how can you possibly blame NVIDIA for compatiblity issues?
3DFx doesn't use AGP direct texture execution. This means that textures must be copied into texture RAM first before they can be used. In cards that support "real" AGP, textures can be used directly from system memory. It also lacks fast-writes and sidebanding, and essentially uses AGP as a 66MHz PCI slot.
AGP has no problem with multiple processors, it just can't handle multple devices. That's why Voodoo5 is on one card. However, AGP has no problem with multiple processors, as evidenced by the ATI Rage Fury MAXX, which uses multple Rage 128 Pro chips. It seems that 3DFx didn't engineer support for AGP texture execution into the VSA-100 chips. To 3DFx's credit, it really doesn't matter because in current games AGP makes less than a 1% differences. However, future games are going to be limited both by the lack of AGP texturing, and by the fact that even the 128MB Voodoo5 6000 only has 32MB of RAM (because each proc needs it's own copy of the textures.)
Actually, it won't be 10x more powerful than the PS2. More like twice as powerful. That's nothing. In the same situation, the original Playstation whooped Nintendo's ass even though the N64 was more than 3X faster, and featured stuff like texture filtering and full screen anti-aliasing (which PSX doesn't support.) Of course, Nintendo actually knew the console industry, and knew how to deliver a simple, stable, easy to use product. Microsoft is bringing too much PC garbage into the XBox to be able to do that. XBox will not be simple, stable, nor easy to use. It will require updates, and upgrades, and patches. Nothing against MS, it's just that it doesn't understand the industry. None of this takes into account the fact that Playstation 2 is coming out more than a year earlier. As history shows us, that's a killer. PSX whopped Nintendo largely due to the fact that N64 came out more than a year later. Lastly, Sony is a lot bigger than Microsoft, and they have more developer support and consumer mindshare. MS is going to be eaten for lunch.