Slashdot Mirror

As both a FreeBSD user and Gentoo user, I think the best description would be that Gentoo is BSD for Linux users. As a humourous aside, some friends have also started describing Gentoo as "ricenix: 2Fast2Optimized". ;-)

Gentoo is laid out fairly logically (no idea if it follows the Linux Standards Base though). The main benefit is the total control you gain over your installation - much like you gain with BSD (hence, BSD for Linux users). Though it is achieved through the remarkable Portage package management system, vs FreeBSD which is a wholly maintained o/s, with a very large "ports" system.

The only thing that keeps me from using FreeBSD on my workstation is that I do play some games on Linux, and write software to support game playing on a local Australian gaming network. For those that don't need the fluff that's supported on Linux (games being a primary example), almost everything else is available under FreeBSD. But to save you extra work, Gentoo is probably the way to go (easy to manage once installed through portage).

NVidia already has.

Mmmm, pixies...

(Incidentally, can anyone explain to me how "naked ..." and "... in gossamer lingerie" aren't a contradiction in terms?)

Not Extended... That was EISA, Extended- ISA, wrong century. PCI-X=PCI=Express

after posting some hardware sugs, i forgot about the driver issue:most of the newest 64bit driver work from the
major manufacturers appear to be for Windows(ech). ( Here's a review that came out today.)

The latest Linux drivers from nvidia aren't too old; their last nForce3 update was in Dec 2003 and the gpu drivers in Jan 2004

Tyan have a page of drivers, as does Highpoint, and Adaptec

Look into the suse amd64 message boards - they seem to be having some success...

after posting some hardware sugs, i forgot about the driver issue:most of the newest 64bit driver work from the
major manufacturers appear to be for Windows(ech). ( Here's a review that came out today.)

The latest Linux drivers from nvidia aren't too old; their last nForce3 update was in Dec 2003 and the gpu drivers in Jan 2004

Tyan have a page of drivers, as does Highpoint, and Adaptec

Look into the suse amd64 message boards - they seem to be having some success...

nVidia support the iAMD64 instruction set for Linux in their klunky closed source way. I do not know if YAST will grab iAMD64 nVidia drivers from their web site with a click of a button or if more geeking is needed to get it to work. nVidia

64 bit linux drivers have been out for the nforce 3 since last year. You can grab them here. Opteron 250's official AMD pricing is $851. Street pricing is near $1,000 only because availability is still low. As more vendors pick them up they will drop considerably.

Because there are two Opterons in it. I was talking to Iwill about it today. Yes, single chip solution as in the nForce 3 chipset, as is common knowledge, integrates northbridge and southbridge functions into one chip which reduces latency, and improves performance. Single chipset, two CPUs, less valuable PCB space since no separate southbridge is required, less traces required, easier to make a smaller design. nforce 3 info.

Robert
For real 64bit performance visit VooDoo software tuning and download the 64bit 2004 Longsword Gamez Demo. The Download of UT2004 64-bit English Linux Demo is around 200Mb.

As for organizations beating slashdot to the punch on this one, that's true... but it's good to see this getting even more exposure. :)

GPGPU (General-Purpose computation on GPUs) was a hot topic at various conferences in 2003; a number of papers were published on the subject. At SIGGRAPH 2004 there will be a full-day course on GPGPU given by eight of the experts in the field (including myself).

Mark Harris of NVIDIA maintains a website dedicated to GPGPU topics, including discussion forums and news postings. Well worth a browse if you're interested in GPGPU topics.

I look forward to seeing some of you at SIGGRAPH! :)

--Cliff

Nvidia has already announced Gelato which uses the GPU to render regularly CPU intensive frames for video production. the link is here and the film industry apparently already uses this. I think that it requires the increase 2-way bus bandwidth that PCI-Express offers to be of any use but it's interesting nonetheless. I suspect that with PCI-Express MoBo's becoming more prevalent there will be a new market for arming PCs with non-function-specific (eg. not dedicated to graphic) co-processors that can assist with processing intensive tasks.

Remember NVIDIA's Gelato, which was released a few weeks ago?

Gelato uses the GPU as a floating point processor, in addition to the CPU. I would still love to see a movie rendered in realtime with OpenGL, though.

A few seconds on Google and I found nVidia's mobile offering. A few more seconds and I found this. Undoubtedly ATI has something similar.

I looked for the best link I could find on it...

http://developer.nvidia.com/object/Particle_System .html

Some of the demonstrations there include particle systems whose trajectories are entirely done within the shader. I'm not exactly sure why this would be 2001... I suppose it shows that even back then, some physics functions were possible in the GPU.

Ultimately, GPU's are going to be able to function as highly parallel, highly powered vector processors. Many of the cards are going to be able to handle physics as well as, or better than CPUs, at the expense of possibly graphical power.

What I'm ultimately more interested in is some kind of API for physics, if it's even possible. In that case, graphics companies would have to write drivers that properly load balance physics tasks between CPU and GPU.

That would indeed be interesting.

0.nvidia drivers
1.amsn
2.firefox
3.mplayer+codecs
4.ogle
5.btdownloadcurses.py


After that I restore my porn/mp3 collection. I can usually go for several weeks before having to install anything else. As I need them, I install..

6.RealOne player (to listen to BBC 2)
7.WineX
8.$Nes_Emulator
9.mpgtx (for joining mpeg files)

Look at screens. Graphics cards have improved massively (electronics), screens (optics) used to be 1024x768 quite a while back, and typically aren't more than 1600x1200 now. The LCDs will hopefully change that though, since they're much more scalable (make more pixels) than a CRT (move beam faster).

See, this is why I want one of these. Unfortunately, there are no Mac-compatible video cards (yet) that can drive this puppy, and the PC cards that can still cost thousands of dollars. Still I must say, 3840x2400 at 22" (204dpi) is nothing to sneeze at. (or on.)

Then again, my Dell laptop packs 1920x1200 resolution into a 15" display, which is razor-sharp in its own right (150dpi). Jarring to realize that's barely over 2 megapixels.

Of course, as soon as they have a WHUXGA screen (7680x4800) on a 12" laptop, I'm there.

On another note, I haven't been keeping up with my 3D like I used to, but some software, such as Renderman, can do distributed rendering on a single frame, and then automagically merge the results. I don't think Brazil offers this yet (could be wrong?), but they're working on it (under the name of Banshee, bottom of page. If your renderer of choice offers such a feature, you could build some serious distributed rendering for $12k.

Indeed. NVidia's FAQ for this group says " It is the evolution of NVIDIA's acquisition of Exluna in 2002"

The good thing is that it is Nvidia, who do the best OpenGL implementations around.

2004 will see the first handheld devices using the same 3D technology that powered the Dreamcast gaming console.

Of course nVidia and ATI and others are also going to release 3d for mobile phones.

In the last video game generation people were shocked at the unbeliveable power the consoles had. The n64 featured an advanced 64bit 100MHz MIPS RS4?00 chip with SGI level 3d graphics designed by SGI for $200. Only a few years before that a slower 32bit 33Mhz MIPS 3000 chip with worse graphics would've cost many thousands of dollars. Just wait a couple years and we'll have $20 watches with gigs of memory to replace our iPods and more power than the xbox ;)

I thought most renderman stuff was rendered, not raytraced?

I'm sure you've seen the raytracing OpenGL examples such as nVidia's

The OpenGL 1.0 pipeline is great for games, CAD, etc. The future of OpenGL is basically "here is a really powerful parallel processor (AKA pixel shader) and some memory (AKA textures), go use/abuse this in anyway you like."

There are a lot of people working on General Purpose ways to program the GPU/VPU such as BrookGPU. Moving forward graphics chips look less like old style OpenGL where the chip is hardwired to support up to 8 lights, gouraud shading, and a texture, and more like a giant processing farm that will be good at certain tasks (render farm) and worse at others (compile farm). I belive raytracing will be one of the tasks future GPU/VPUs are good at.

Being integrated with the motherboard doesn't make a performance difference on any board I've ever seen. It still goes over the PCI bus, it's just not using a slot. Creating a separate bus just for the ethernet port would be too expensive.

Not really if the motherboard chipset was designed from the get go to support somthing like that, it was mentioned earlier in the posts about the Intel i875 as well as the nForce3 based systems.

yeah, cuz they're reall hard to find on the nvidia site.

http://www.nvidia.com/object/winxp-2k_archive.html

The first time I set up NWN under Linux and ran it under TWM I almost came all over myself! It gave me better graphics, better performance, and a better gaming experience. Thank you Bioware, Nvidia, and Suse.

One has a Tux sitting next to it and one has a blue screen on it. And the caption is "And I could have spent HOW MUCH less?")

You would indeed have to be a soccer mom or busy single parent (meaning: little or no knowledge of computers due to lack of time/interest/intellect) to fall for such a ridicilous ad. Windows no longer bluescreens anywhere near as often or as easily as it used to, as anyone in the know will tell you if you asked. Yes, it's still possible to bring the system down with a dodgy driver. But please name one desktop operating system where this can't happen... At least on Windows, I can use my perfectly common graphics card with full 3D and not have to worry about crashes.

You know, silly stuff like reliable, robust video and sound drivers.

It's funny, but Linux is in much better shape for video drivers than audio ones. Since the game-capable graphics market only includes two companies, Linux is already adequately usable.

But since soundcards are technically easier to make, there's many more brands still in active use. Many gamers who buy the latest NVidias to squeeze a few more FPS or pixels might still be satisfied using motherboard audio output, or a $2.50 PCI soundcard.

Linux audio support is close to adequate... but unfortunately, the Alsa Project's longstanding philosophical refusal to move software mixing into the central driver means you still can't expect Linux to run games on any random piece of desktop PC hardware.

I read Microsoft's requirements page on 1080p HDTV, and that's why I didn't buy the T2 DVD edition which features an HDTV version of the movie: my CPU is an Athlon XP 2500+. I tried playing the 1080p trailer, and indeed it was choppy as hell, although playing the 720p version was fine (barely).
Then I bought myself a GeForce FX5900 (PDF file, 209 KiB), which provides hardware acceleration for HDTV content; I played the 1080p trailer again with MPlayer, the video was smooth, and CPU charge was about 70%.
I don't know how this hadware acceleration works (I would expect it to be bound to a specific codec), but it certainly works damn well.

+4 insightful? More like bold face lies. Try looking toward the bottom of http://www.nvidia.com/object/linux_nforce_1.0-0261 .html. GPL'd drivers for the Nforce2 chipset, including audio and ethernet. And it's not like it was hard to find. Just three clicks off of Nvidia's front page.

And for what it's worth, the Asus A7N8X-Deluxe is the best motherboard I've ever owned. And if you take a look at oneline reviews, you'll see I'm not the only one who thinks so.

Who's trolling now? Nvidia supplies FreeBSD drivers for XFree86.

There are many "window managers" available for MS - if I understand correctly, these just use the Windows API to do their stuff. Snap-to, focus-follows-mouse, automatic-window-previewing during Alt+Tab switching, multiple desktops -- all these are available for free on the MS platform - mostly, IIRC, by installing the 'TweakUI' Windows software that Microsoft makes available on their website as a free download for power users. MS doesn't make these options available for normal users because of the 80-20 principle -- in their "regular" s/w they implement what is required by most of their users.

Two more examples:

1. I recall seeing a basic multiple-desktop manager on either download.com or the Ziff Davis network (don't recall) that implemented multiple desktops on a Windows system. It was open source.
2. Nvidia has an "Nvidia desktop manager", bundled free with it's Detonator drivers. This one is much easier to use than UNIX WMs and offers some cutting edge functionality: extensive hotkey, transparency, and multimonitor support, as well as two features I haven't seen on Unix WM's yet -- window color keying (useless) and partitioning *visible* screens into multiple areas with grid-lines (potentially very very useful). Here are two links to Nvidia's s/w:
Nvidia Link
Another link

"The only reason the public has stuck with Windows as long as it has is simply because they are familiar with it."
Also because it's simpler to use (compared to Unix/Linux), works well (except for security), and has scads of desktop app. support: "developers, developers, developers..."

Well, nVidia has support for their graphics cards on 2.6. As for the other hardware you'll have to google yourself. The nVidia link wasn't paticularily hard to find.

I do know that HP and one or two others are shipping Itanium2 servers full-steam at the moment, despite Intel's recent 64-bit malaise. It's almost a given that all of those will be running 64-bit Windows. I've seen a demo of a HP 64-bit workstation running 64-bit Windows, and it was really nice. It even had accelerated video drivers, but I don't know what video hardware.

If you comment to this guy, make sure that you emphasize the need for a Linux port of ForceWare. This is (IMHO) the closest thing to a TiVo competitor for the PC world. Right now, it is Windows-only.

If it weren't for the GPL and the non-binary rule, NVIDIA could actually release a linux based PVR ISO for their nForce boards. Oh well... maybe they can do something in BSD?

I moved my scanner to my linux server and installed "sane". I installed "sane-twain" (free/OSS software) on my XP box, and it then accessed the scanner on the linux box quite happily. Some of the icons weren't as pretty as the windows driver, but all the same stuff was there.

Later I installed a dual-boot setup on my workstation. I used XP less and less because it was so SLOW and getting slower - I don't install much new software once i get set up either - and yes, I ran AdAware and anti-virus software.

Eventually, I only ever fired up windows to run Quickbooks. Now that I have Crossover Office installed, I don't even do that (crossover runs the native windows quickbooks just fine).

A few weeks ago I used Partition Magic to downsize my XP partition (which I had done once before) to make more room for linux. My XP partition was 15GB with about 3GB spare, while Linux was 8GB with no spare.

(un)fortunately, Partition magic trashed my XP partition..... so what did I do? stress? no... I just said "well, I don't use it, so why recover/re-install it? Partition Magic then proceeded to do a wonderful job deleting the XP partition and moving/resizing the Linux Ext3 partition. I now have a lovely 23GB linux partition with loads of free space. GNU parted provides similar capabilities on linux, though I have yet to check it out in person.

The best thing, is that I have a WinRadio card. Winradio stopped developing their linux drivers shortly after releasing a working open-source driver a few years back. Someone started a sourceforge page and updated the original driver. They haven't done any work on it for almost a year, but i was still able to download it and with about a day's work yesterday, I have my winradio card working on kernel 2.6. (yes, I have contacted the sourceforge page owner about sending the updates so everyone can use it).

Someone is going to say "but i can't write software so what good does that do me". My answer is that I don't write 99% of the software on my linux box. I just contribute where i can because i want to - it doesn't matter if I draw a few graphics, write code, make a web page, or do nothing at all, I can still use the work of people like myself.

The best part is that I don't have to start from scratch - I don't have to start writing the driver all over again just because Winradio don't want to update the drivers for my old card, and won't give me the source code. (although to their credit winradio do provide a windows driver for XP, even for this, their oldest card) Another example is the NVidia drivers - the official ones don't support Kernel 2.6 yet, but due to the open source component (the core of the driver and GL code is closed source), I can get a 2.6 driver from a third party, who, just like myself, did it for himself and released the result to the public.

Right now I have ALL my hardware working quickly and well, even though some of it is 5 or 6 years old, and ALL of it is 3+ years old, and I'm running the latest version of the OS.

I just can't get that anywhere else.

You're about to say "but I can't get drivers for the latest gadget". Well if the vendors followed the Winradio and NVidia examples, by releasing a linux driver, you wouldn't have that problem.

Their current installer will try to use a pre-compiled binary driver, or of one isn't in the file you download, it will look on their site for one. Failing that (or if you choose to roll your own anyway) you can compile the included source.

Thay do this to protect their code, while making the driver easy to port to new kernels.

I'm running binary portion written by NVidia, and the source portion ported to 2.6 by someone else. NVidia Linux Drivers Ported to 2.6

For example:

• SuSE 9.0 Pro for amd64.
• SuSE Enterprise Server 8 for amd64.
• Mandrake 9.2rc1 for amd64.
• Fedora Core1 test1 for amd64.
• Gentoo's amd64 info
• Red Hat Enterprise Linux AS & WS are available for amd64.

I've been using Gentoo's amd64 stuff for a little while on my new Shuttle Box. Things are generally good although there are still a lot of packages that are masked. KDE is also problematic which may be a turn-off for some people.

A colleague just got a new dual-opteron Workstation from Pogo and is running SuSE 9.0 pro for amd64 and is rather happy -- just about everything plays nicely.

Multimedia has significant problems on both systems. No flash player for 64-bit, mplayer and related multimedia requiring 32-bit codecs. Nvidia amd 64 drivers require some patching if they work at all, at least as of last wednesday.

Otherwise quite happy with all of these. Mandrake claims to have multimedia stuff working properly (see above link for info) but wants to eat my partition table so I haven't checked it out yet.

--
"Now you'll see why they call me the Velour Fog" --Zapp Brannigan, 25-star General & Cpt.

Why doesnt your FreeBSD support your nVidia GeForce card?

Download the official FreeBSD drivers from nVidia.com and it should work fine.

http://www.nvidia.com/object/freebsd_1.0-4365.html

With the state of 3D and video card driver support, openGL, etc... they still don't run as well as on Windows systems

WineX allows you to play some of the newer games. Though there are a bunch out there that will be MS Windows only. If you NEED to play PC games, either dual boot or buy a second PC with a cheap KVM and have a ball. I personally use just Linux and a PS2 to handle any gaming needs.

... and here's the latest 3D stereo drivers.

Ah, Tom's Hardware. Not trying to be negative, but IMHO, they are a terrible source for tech information, and the bulk of their reviews contain startling errors, conclusions that defy reason, glaring omissions, and sensationalized reporting. The majority of those writing the reviews clearly have no idea what they are talking about, at least regarding the subject they are reporting on. Overall, I would rate them slightly above HotHardware.com.

Tom himself, as far as I can tell, is on the ball and knows his stuff VERY well, but he doesn't write articles much anymore, and obviously doesn't read them either. It is a common practice among hardware enthusiasts to quote Tom's for the humor value, trying to see if the author of the latest article is even more clueless than he was in his (or her) last article.

To be fair, they do have some excellent articles occasionally, and were the first ones to dare publish information on Intel's unstable Pentium III 1.13GHz processor, but unfortunately these seem to be the exception rather than the rule.

Also, as has already been stated, XGI is hardly a new company. Of course, these bits of SiS and Trident are in completely new territory if they are trying to compete in the high-end gamer's market. Considering that this is their first real foray into that market, I think they have done an amazing job. I'd say give them the benefit of the doubt until they prove otherwise. Remember, even the (once) most respected companies in the field can faulter, and that XGI has something that is even in the same ballpark as the most seasoned of players is an impressive feat.

Ah, Tom's Hardware. Not trying to be negative, but IMHO, they are a terrible source for tech information, and the bulk of their reviews contain startling errors, conclusions that defy reason, glaring omissions, and sensationalized reporting.
The majority of those writing the reviews clearly have no idea what they are talking about, at least regarding the subject they are reporting on. Overall, I would rate them slightly above HotHardware.com.

Tom himself, as far as I can tell, is on the ball and knows his stuff VERY well, but he doesn't write articles much anymore, and obviously doesn't read them either.
It is a common practice among hardware enthusiasts to quote Tom's for the humor value, trying to see if the author of the latest article is even more clueless than he was in his (or her) last article.

To be fair, they do have some excellent articles occasionally, and were the first ones to dare publish information on Intel's unstable Pentium III 1.13GHz processor, but unfortunately these seem to be the exception rather than the rule.

Also, as has already been stated, XGI is hardly a new company. Of course, these bits of SiS and Trident are in completely new territory if they are trying to compete in the high-end gamer's market. Considering that this is their first real foray into that market, I think they have done an amazing job. I'd say give them the benefit of the doubt until they prove otherwise. Remember, even the (once) most respected companies in the field can faulter, and that XGI has something that is even in the same ballpark as the most seasoned of players is an impressive feat.

Stalker seems to be overlooked quite a lot. Still it has Nvidia's attention. Tom Ohle from Bioware is looking forward to Stalker as well as Andrew Edelsten from Auran. And I know I am keeping both my eyes on this gem :)

this seems to be pretty common. newest drivers, like any kind of software, may take a little while to become fully stable.

tell us about it

FreeBSD drivers dont get updated for about 6 months
latest (2nd since 11nov 2002) was released back in 1 July of 2003

I wonder if Nvidia is droping Freebsd support? but thats a nonsense just to think about it. Since we can run manly as much games as Linux community does, and the drivers prove to be stable.
Well for some, after some tweaks around :\

Or they are just waiting for the 5-Stable to came out? But that will only happen in march/april 2004 atmost when 5.3R gets out and that is a bit scary to think about it :(

The prob isnt keeping on the payroll 1 or 2 bsd developers. I manly use Freebsd(oki some loonix too :P ) and what i had paid for my 5 cards plus judging by the every day contact i find in irc support chans, forums/mls or in gameservers all around the world, more and more ppl are using freebsd with nvidia for multimedia/desktop purposoes.

The current drivers, miss some linux supported features , better opengl cinefx, glx1.3, dual monitor, agp kern module and nvidia's ago gart surport, nforce2 integration and device listing, 3d glitches and all sorta of bugs during gameplay, even if isnt a big release, just making small improvements at a steady pace like every 3 or so months will make us smilley and felt that nvidia _does_ care about FreeBSD!!

xpto

• I went here and ordered.
• I followed the instructions provided here.
• Installed a bunch of stuff from Windows Update. You know, stuff like this, but for Windows.
• Rebooted
• I downloaded this.
• Rebooted one more time.

Some things in life are needlessly fucking complicated and obscure. For everything else (fun, productivity, selection, support, choice) there's Windows.

This message is brought to you by the Common Sense League Of Wisconsin. "Vacating Parent's Basements And Saving Minds One Copy Of Windows At A Time (TM)".

My GOD you people are insane.

Mod away, like I give a flying fuck.

NVidia has now for a LONG time supported the AMD64/Opteron platform (as well as IA-64, but who cares - UT2004 sure isn't going to appear on IA-64)

Here is a link if you're too lazy to check it for yourself -

The only thing you could POSSIBLY accuse Nvidia of not having done is providing 64-bit Drivers for FreeBSD (or drivers for any of the other BSDs). But even then

My current set up is an Athlon XP 2400+ pared with a KT266A-chipset motherboard. I had NO problems with the chipset under Linux or FreeBSD - ever, ever, ever. Sure - VIA has issues with new chipsets, but hey, they're human too - who doesn't. If you stick with the A-releases you should be more than fine.

I have, however, encountered it impossible to run Windblows XP without SP1 - without the harddrive being unusably corrupted within the first 30 minutes of uptime. But thankfully - that horror is over. No more Mr. Balmer and Microsoft on MY computer - or any computer within my proximity, ever.

Anyways - I can't wait to start porting my home-brew kernel onto x86-64 whenever I come across a cheap Athlon64.

nVidia drivers run *natively* under FreeBSD.

You can also run Linux OpenGL software via FreeBSD's Linux support without a recompile.

nVidia FreeBSD Native Drivers

The newer cards that you mentioned are mainly geared towards 3D performance because their biggest market (the gamers) want the utmost 3D performace they can get. If you want good 2D performance, you'd probably be better off buying the cards that are geared towards the workstation market (like nVidia's Quadro NVS line) instead of the gaming market. Also, as has been mentioned, DVI-capable cards and an LCD will give an improvement.

Before that you used to say "render me with [phong|gouraud|flat] shading" and the whole thing looked uniform. Shaders programs let you do cool things like features [e.g. skin, roughness to things, etc...]

I am a 3D coder guy, and your description is grossly accurate. Allow me to be more accurate. :-)

In 3D graphics, models are composed of vertex sets (all the edges) and fragments (pieces of the model). For performance reasons, you want to keep all the models on the video card, and not re-transmit them through the system bus every frame.

Example: a walking person.

In a "vertex soup" system, you would have to carefully move every single vertex, every frame, then upload that to the video card, and let the video card rasterize the model (=convert it from points into a picture). This is not very practical.

Fragmenting your model into pieces (torso, shoulder, upper-arm, joint, forearm, etc.) lets you keep all the pieces on the video card and move the different pieces around, much like a jointed wooden doll. Keeping it on the video card means you don't have to spend a long time waiting for the data to be transferred to/from the CPU, so you can see a *huge* speed improvement.

Now lets say you want to modify the model, maybe to make the joints look beter.

Without shaders, you would have to modify a bunch of points on the model, then transmit it back up to the video card, then render it, every frame. That is very slow, but easier to transmit a small portion, like a joint, every frame than it is to do the entire model.

Shaders try to help you keep the model on the video card. Rather than modifying and re-transmitting a model fragment, you send a shader program and give it instructions every frame.

What they can do

Vertex shaders let you modify :

• vertex transformations (moving points around),
• normal transformation (changing what direction is 'forward'),
• texture coordinate changes (changing where the artwork on the model sits),
• lighting changes
• coloring changes

Fragment shaders let you modify:

• Stuff that takes place between vertex points, such as surface smoothing,
• Textures (changing the artwork that is wrapped around the points),
• Texture application (changing how the artwork is wrapped),
• depth cueing (fog or vanashing into the distance),

You can do an awful lot with those shaders, much more than the few things you mentioned. Some of the big things right now are hair/fur, grass/sand/dirt on terrain, and perturbed surfaces.

But there are a lot of things shaders cannot do. They only operate in specific places in the graphics pipeline. They must be small and operate within very tiny time and space requirements. They have limits on execution control (no looping). They have very limited input, and output can't effectively be re-used in other places.

What I don't get is why didn't they just make the GPU a generic RISC with say 32/32 registers [ALU/FPU] and a set of instructions that fast graphics would require [say saturated X bpp operations, fast division, etc...] That way you have a processor you can just upload code to.

Shaders need to be very small and run very quickly, since they may be run in the hardware hundreds of millions of times each second. One of the biggest complaints I hear about shader languages is that you can't loop, and you can't do more general programming on the GPU. But shaders have to live within very tight rules. They function as a slow pit-stop or detour within a VERY high-speed rendering pipeline.

But lets just think about it; Let's briefly consider the GeForce4 Ti 4800. It has an advertized performance of 136 million verts/sec and 1.23 Trillion ops/sec. If they had only a single pipeline, it would ne