Slashdot Mirror
A History of 3D Cards From Voodoo To GeForce
Ant sends us to Maximum PC for an account of the history and current state of 3D video cards (single print page). "Try to imagine where 3D gaming would be today if not for the graphics processing unit, or GPU. Without it, you wouldn't be [trudging] through the jungles of Crysis in all its visual splendor, nor would you be fending off endless hordes of fast-moving zombies at high resolutions. For that to happen, it takes a highly specialized chip designed for parallel processing to pull off the kinds of games you see today... Going forward, GPU makers will try to extend the reliance on videocards to also include physics processing, video encoding/decoding, and other tasks that [were] once handled by the CPU. It's pretty amazing when you think about how far graphics technology has come. To help you do that, we're going to take a look back at every major GPU release since the infancy of 3D graphics. Join us as we travel back in time and relive releases like 3dfx's Voodoo3 and S3's ViRGE lineup. This is one nostalgic ride you don't want to miss!"
Personally, I found the article quite nice - it was a nice trip.
Beer is proof that God loves us and wants us to be happy.
at best buy a couple weeks ago... too bad the box was supposed to contain a Nvidia 260... s3, 3dfx, all kinds of old ass graphics boards in the box.. but no 260...
What could possibly hurt the security of the American people more than giving our own government the ability to hide its
Had beautiful graphics and ran on a 386sx with a 128 MB VGA card and a 2D GPU.
So I call Bullshit- the only reason a high powered GPU is necessary is because game programmers have become LAZY.
SJW: a person who perceives an injustice, and while correcting it, commits a greater injustice.
Without it, you wouldn't be [trudging] through the jungles of Crysis in all its visual splendor
Hmmm...is anybody able to play Crysis in all its visual splendor?
The 'MeTaL' acceleration was bullshit. On UT99 I think the software renderer looked about the exact same as the MeTaL. When I popped in a 12 meg Voodoo2, I promptly tore the S3 board apart and threw it away, and popped in a Matrox Millenium MGA card for 2D stuff.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
But the market never accepted them because no matter how thin they made the peripheral slots, the damn things would just fall through the case.
In fact, I've got 3 old 3dfx Voodoo cards I'm willing to part with... cheap! 2 of them complete with TV tuners. Good luck finding Vista drivers for them!
I've abandoned my search for truth; now I'm just looking for some useful delusions.
I used to have an old IBM CGA color monitor that I used on an IBM PC-XT. 8 colors IIRC, including lovely shades of magenta and "brown". I was the envy of everyone on the block. When the first addon graphics cards came out, I got a Hercules card and that absolutely ruled for running Flight Simulator.
These days I'm content to run a year or two behind state of the art. The cards and games (the few I play) are cheaper that way. I think it's pretty much a no-brainer to say that gamers have driven this industry since very early on.
I think I have four Voodoo2 cards. One is my original, and the other three I bought for about $30 when I was in high school. I installed them on the yearbook computers I managed so that the staff could play Half-Life when we had nothing else to do.
"Anyone who [rips a CD] is probably engaging in copyright infringement." - David O. Carson
I remember that I actually had *one* game that supported my Virge/DX - Descent 2.
It did look a better, but was slow enough to make you want to switch to software rendering immediately.
The name "Diamond Multimedia Stealth 3D 2000 PRO" did sound rather impressive on paper.
The article would have been more impressive with screenshots of the games, though.
Save your wrists today - switch to Dvorak
So, don't get me wrong. I love beautiful graphics. I love the immersive environments that they create. The atmosphere of games like Bioshock are great. Even WoW, which arguably has very scaled down graphics, is extremely involved and really pulls you into the game.
HOWEVER...
For as much as I like these graphics, games just do not hold my attention like they used to. I know I'm going to sound like "The Old Guy" with his nostalgic memories, but I spent hours and hours on games where graphics wasn't the primary draw (even for that time period). Heck, I didn't get Legend of Zelda (the original) until well after SuperNES has been out for quite some time. But, I spent so much time on that game, my original Nintendo practically burned itself up.
Basically, the point I'm trying to make is that, while graphics are important to the gaming experience, if a company really spends time on the storyline (Fallout 3, or Bioshock for example), or focuses on the fun factor (Smash Brothers!) games can be just as awesome and fun. It's not just about (or at least should not be just about) the "visual splendor."
Remember when Intel was claiming that you couldn't run an i740 on anything other than an Intel chipset, due to "incompatibilities"? Didn't stop me from using it with a K6, and since Intel did provide documents for that chip, it ran in Linux, too.
I remember when we were big time because we had color. No movement... just more colors than amber or green. Whooo Hoooo! We were the shizzle!!! Especially with our 9600 baud modems!
Then we got movement too, not just pretty words on the BBS, because, yeah, that was before the "real" internet happened.
HDGary secures my bank
Kind of ironic title when you take into consideration that on black friday, October 15, 2002, 3dfx's assets were purchased by nvidia. The geforceFX was built using a lot of ex-3dfx engineers, so there was a very literal translation from voodoo to 3dfx. PS, I used to LOVE 3dfx cards, still would, but I've been running radeons since the 9700pro beat the living snot out of the entire geforcefx line.
Is it sad that I am more likely to recognize you and your posts by your sig than your name or UID?
I've yet to see so-called 3D card under $500 that can produce a true holographic display.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
The intro says to include ... other tasks that [were] once handled by the CPU.
In fact, there is a regular cycle of inventing video add-on processors, seeing them spread, then seeing the CPUs catch up and make the older video processor technology obsolete, moving the work back to the CPU. Then, of course, someone invents a new video co-processor (;-))
Foley and Van Dam, in Fundamental of Interactive Computer Graphics called this "the wheel of karma" or the "wheel of reincarnation", and described three generations before 1984.
I suspect the current effort is more directed toward building fast vector processors, rather than short-lived video-only devices. Certainly that's the direction one of the Intel researchers suggested she was headed.
--dave
davecb@spamcop.net
I think I have one of each of these in a desk drawer in my house. Everytime I stick my hand in there I get cut.
This is a sig. It is like every other sig in the world, except that it is mine, and it is different.
lol wut
Without it, you wouldn't be [trudging] through the jungles of Crysis in all its visual splendor
I can barely do that even with a GPU. :)
*runs away*
/* No Comment */
My first computer was a 233mhz pentium with a crazy case layout that wouldn't allow a graphics card to be installed. I remember playing Half Life and Quake 1&2 using software rendering at something like 400x320 resolution, and thinking it looked amazing. How times have changed since then, and my first computer is downright modern compared to many others' here.
I still remember my first Diamond Monster 3dfx video card. I bought it moments after seeing a demo, because it was just that awesome.
I then remember downloading the 3dfx patches for games like Tomb Raider and Interstate '76 (what a great game that was)...
Good times. We take so much for granted these days.
Am I missing a joke or is it an error that the description of the GeForce FX 5800 features the image of a vacuum cleaner? I mean... not that a vacuum cleaner with 15 million transistors is not impressive...
I remember the days of my trusty Matrox card playing Descent and Duke Nukem. Anything that ran on DOS seemed fast.
For shear enjoyment, Rise of the Triad and all of its 2D-ness still gets my vote for all-time game. Who can forget such classic weapons like the Drunk Missile and the Fire Wall? Just pray you don't cross into a hallway that someone had targeted with the firewall at the wrong time.
Good times.
Bearded Dragon
... There was not a single VLB (VESA Local Bus) accelerator on that list. As I recall, the VLB slot was made for video acceleration, so they rather missed the boat by omitting those cards. Starting at Voodoo (except they started with ViRGE) is not a very comprehensive history of 3d acceleration.
Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
You need the Bitchin'fast3D2000
this time the image quality was improved, particularly at higher resolutions (1024x768) where the Voodoo1 struggled.
Interesting, considering the Voodoo2 had a 800x600 resolution limitation
GOD was that thing a damn oven. But damn if the games didn't look (comparatively) sweet on it!
Chas - The one, the only.
THANK GOD!!!
This article is interesting but is heavily weighted towards the consumer gamer, but interesting developments are happening elsewhere as well.
Perhaps the author will do another article with a different focus? Here's a place to start. It's valuable to remember that approximately ninty percent (90%) of all PCs sold are sold with Intel onboard graphics chips.
http://www.google.com/search?q=intel+3D+graphics+history&hl=en&client=firefox-a&rls=org.mozilla:en-US:official&hs=o0&sa=G&tbs=tl:1&tbo=1&ei=fw0TSru0OY7msgODneX4CQ&oi=timeline_result&ct=title&resnum=11
Everything in the Universe sucks: It's the law!
Soon we will have multi-core GPUs
The V2 could only hit 1024x768 in SLI configuration, otherwise, you're right.
Fifty watts per channel, baby cakes.
http://www.cracked.com/article_15732_life-after-video-game-crash.html
I can live with bad grammar in the submissions, and of course in the comments, but can Technical Journalists PLEASE take a few goddamned English courses?
My Babylon
I owned the Voodoo1 piggyback and it was good, but didn't satisfy me. First chance I had, I got a hold of the 5500 beast and had to use the Dremel on my case to squeeze that mother in.
But it was a killer card, giving killer frame rates at high quality.
First rule of holes; When in one, stop digging.
My first 3D video card was a Rendition Vérité 1000; IIRC it was the first card that could do transparent water effects in Quake. Truly a defining influence on my college career.
"Ask not what your country can do for you." --John F. Kennedy
I remember my old Acorn Electron. Elite in black and white wireframe graphics - it was fantastic. The first graphics card I actually bought on its own was a 3D Prophet 4000XT 64MB, which I still have in my parts box. I think I still have a Rage from my old G3 Mac lying around too, and the G3 and G4 I still have still contain their Rage and GeForce 2MX respectively, and still work perfectly.
An intriguing solution to a problem that should never have existed in the first place...
I had an AGP Permedia 2 with 16MB when everyone else was dicking around with PCI Voodoo 2s with 12. The V2 was slightly faster, but the Permedia would not only let you do larger-window OpenGL (XGA, for example) but it also had superior lighting effects. And, you know, actual OpenGL, not just MiniGL. I kept that until the GEforce 2 came out and have only deviated once... to an ATI Radeon 9600 XT that was a total lemon.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
True indeed. Ahhh, the memories - getting a new screen & popping a hernia to lift it onto the desk, then having to buy a new card to be able to experience the highest resolution.
Then - discovering - after getting updated drivers via the (snail) mail - that your card could not support a decent refresh rate at the highest resolution...
Then...finding out that your PC could not actually keep up with the data that certain apps wanted to write.
Then, oh the joy of AMD 486 overclocked Intel clones that drove the VGA straight of the CPU pins - what was that called again? - just a few excotic video cards, but that worked really well - combined performance better than Intel's 486DX2/66....oh dear, time for me to lie down...
Vista isn't even as mature as Voodoo2. That's like comparing the proven worthiness of Jack Lalanne to a loaded trailer of dirty stinky logs that some fool hacked from a mountain of shit. Of'course Lalanne can pull them, but remember this: you need the right double-ball hitch. If this were Chuck Norris, we'ld be using a Pintle hitch for goodness' sake, and you know how those don't slack well in compared to the firmness of Jack Lalanne's double-ball hitch. A Pintle hitche has too much play that would allow sagging, especially when the trailer doesn't have brakes itself (or they were secretly cut) where a foreign foe on that mountain of shit near Redmond would cause the load to jacknife...
Try to imagine where 3D gaming would be today if not for the graphics processing unit, or GPU.
The first thing that came to mind when I saw this sentence was a world where ray tracing was a better developed technology and 3D objects had spent years being much less blocky than they are in computer games.
Blazing Spiders
"obsolete, moving the work back to the CPU."
Except memory and memory bandwidth is not keeping up to move every function to the cpu, we'll see "co processors" like GPU's, etc, for a LONG time to come.
Stereoscope cards can be had for well under $500.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
There's a couple of blatant omissions fron this article:
1. Matrox Millenium I/II - Matrox's best card until the G200 came along. The Millenium II was, at the time, one of very few cards that could be bought with up to 32MB of RAM. Many entry-level 3D workstations running NT4 shipped with such a configuration.
2. Matrox Mystique/Mystique220 - I STILL have one of these AND it's in service. Matrox believed that speed was king, so they designed the MGA1064SG chip for just that, but failed to add features like bilinear filtering, transparency, and mip mapping. As a result, games flew on these cards, but tended to look like utter crap. Both versions had the ability to be upgraded to 8MB or RAM or to full-on video capture and compression using the Rainbow Runner capture daughtercard (which is why I still have/use mine).
3. PowerVR PCX2 - Superceded the PCX1, faster than the original and also an add-in accellerator like the Voodoo1/2 with one two major differences: 1. It didn't require a pass-through cable for operation. 2. It could render 3D in a window as well as full screen. It was also one of two 3D chipsets with native API support by Unreal at it's launch (the other being The Voodoo chipsets). It had, in my eyes, only one major problem - no alpha blend transparency. It could do transparencies, just not alpha-blended. It did have it's own API, PowerSGL, and games coded in it (like Unreal and a Japanese game called "Pure Vex") could look quite good and were pretty fast as well. A few games had after-the-fact patches that added PowerVR support (Mechwarrior 2). Interestingly, the PCX2 could scale much better than the faster cards of the day. I'm not sure of what it's upper limit was, since most reviewers stopped testing it after a while.
4. Savage4 - The Savage series of chips from S3 had their own API called MeTaL. Unknown by many, Unreal (in later patches) and Unreal Tournament both supported MeTaL and through it S3TC. Unreal Tournament 99 looked it's absolute best when run with a Savage4 and the extra textures installed from the second CD. The S4 also had full scene AA, though I doubt anyone ever bothered using it.
5. S3 Virge - The 3D image quality of the S3 Virge was rivaled only by the Voodoo (this was repeated several times in magazine reviews). No other card delivered 3D that looked as good at the time... It was still unbearbly slow.
6. i740 - The Intel chip was one of VERY few that could run Quake III Test when it first appeared thanks to its complete OpenGL ICD.
7. 3DLabs Permedia 2 - Known, but not known... The Permedia 2 was everywhere for a minute. Most card companies were pushing this entry level 3D workstation chip as a 3D gaming platform. Performance wise... well... it kinda sucked. It was missing some features, but thanks to 3DLabs' bulletproof OpenGL ICD, it was one of few cards on the market that could properly render the particle effects on Quake II AND could run Q3T on arrival. Superceded by the Permedia 3, which WAS a better chipset in every way, but still not competitive against the likes of Nvidia and 3Dfx.
There's also the Matrox G400/450, which I still have 4 of in service at home (DH for the wife and 450's for three of my kids).
Fifty watts per channel, baby cakes.
I remember my first AGP card, a 128bit Number 9.
They were certainly short lived.
Haha. Took me a few seconds, I must be slow today! (Most likely the reason being I have to work on a Mac).
They forgot to mention that, jumped straight to PowerVR series 2.
The PCX2 based cards (from Matrox and Videologic) where quirky little cards that deserved a mention...
As other's have pointed out, MeTaL was the Savage 3D/4/2000's native API.
What no one has brought up is that S3 DID have a native API for the Virge series - S3D. To their credit, S3 DID try to get devs to write for S3D, and a few did, but you try writing a game with a software renderer and having to add support for Glide (Voodoo), S3D (Virge), RRedline (Verite), MSI(Mystique), ATI 3D CIF (Rage), AND Direct X... Which at the time would have been the dreadful DX3 and you see why no one really had much interest in it.
Fifty watts per channel, baby cakes.
I saw GLQuake. It was awesome.
First post! (just in case I am...)
I LOVE me some vintage video cards. It's like a mini history of gaming and PC enthusiasm. I used to have a pretty good collection of Voodoos and geforces, and even some very old Vesa Local Bus Hercules cards along with other ancient devices and they were all functional, but it just got to be too much of a hassle when I moved into my own apartment. I wish I had some of them old cards. Others I gave away to those needy for parts (I.E. kid gamers that couldn't afford current gaming setups), and while I wish I still had them, I'm glad I was able to benefit someone else with them instead of simply getting back a mere fraction of their original value from the ol' electronics store.
The eternal struggle of good vs. evil begins within one's self.
I think it's OK that they started with the Voodoo era. That was the first major leap in no-nonsense 3D gaming accessible to anyone who could afford the card. I've been computing since the early 80s so I have a pretty good perspective on it.
I remember the first thing that blew my mind was Unreal, the single player game, then UT. The way the level designers used the card's features were great, along with a deep, rich color palette.
They omitted the PERMEDIA 2 from 3DLabs.
If my recollection is correct it was the first windows compatible accelerator with full OpenGL installable client driver (ICD) as opposed to OpenGL Mini- Client Driver (MCD).
At the time it was the only option for GLQuake. Not strictly speaking a consumer level card, most users had one for using Autocad/3DMax/Maya or similar 3d based applications, but was cheap enough for gamers and hobby 3d developers.
By the time Quake2 came out the Riva 128 or Voodoo had overtaken the Permedia 2 for performance, though their OpenGL drivers were targetted specifically for Quake2 and lacked stability for general OpenGL development.
It seems to be an article biased towards nVidia, which makes for some pretty entertaining reading of the notes about each card generation.
Also, I'm wondering if the person/people who wrote that article were even around when some of the earlier cards were coming out. Talking about the Voodoo 2, but only mentioning SLI in a footnote? If they had actually been there when that card was released, they would've known about how amazing SLI was, and why nVidia used the same acronym for their own SLI (which along with the ATI SLI isn't anywhere near as good as the Voodoo SLI).
The article was EVERYWHERE!!! Someone actually got an S3 Virge to run GLQuake at 60fps using a Virge. I can't remember if the guys that did it used the OpenGL ICD that was included with the game Crime Cities by Techland (the game included an ICD for the Virge and the Matrox Mystique because neither had OpenGL drivers) or another ICD, but I remember this being the talk for a hot second.
Fifty watts per channel, baby cakes.
Poorly designed for 3d applications, borderline for the rest, and leeching off of the system memory to boot!
Banging the gong on this one, for those onboard GPUs are the main reason why the 3d club are buying upgrades.
First rule of holes; When in one, stop digging.
The first 3d accelerated game I played was Quake 2 on a Matrox M3D PowerVR card.
http://commons.wikimedia.org/wiki/File:Matrox_m3d_powercr_card.jpg
What is the first 3D accelerated game you remember playing? Mine was Grand Theft Auto on my 3dfx Voodoo card. I had a Trident as my primary card. My first impression was that everything didn't seem so pixelated. I'm sure that if I were to see the same setup today, each of the pixels would seem to be as big as my pinky, but at the time, the textures just seemed so smooth and well blended.
Several voodoo variations are missing, all by Quantum3D.
Obsidian Pro 100DB-4440: I remember reading a magazine review of this thing, noting how the reviewer said something like "GLQuake at 120FPS is unplayable". This was a single card SLI **Voodoo 1** card.
Obsidian2 X-24: Single card Voodoo 2 SLI. Two Voodoo 2's on one board in SLI configuration.
Obsidian2 S-12: The only Voodoo2 made on an AGP card. The only reason I remember this thing is because I actually own one. It's in storage right now, but a pic of it can be seen here: http://www.geocities.com/subject28/mystery2.txt
I bought it to compliment my Matrox Mystique 220 w/Rainbow Runner which I was completely unwilling to give up.
Fifty watts per channel, baby cakes.
I'm just curious, how low and boring does your life really have to be before you decide that it's a good use of your time to post something like that? I don't even use Linux and I still think you're a waste of time and space. Your parents must be proud.
Intel managed to sell me the "Extreme" Graphics i740 back then and when I look back it just me sad again.
Of course today I would not buy an Intel graphics card even if they named it Larrabee - oh wait they already did?
sure I had one these (or similar) for my Amiga back in the day (yeah, the amiga in its sunset years had some 3d-capable addon gfx cards released for it here in europe. warp3d.library was the de-facto standard opengl subset API used on the amiga).
It was just a PC 3D accelerator chipset with a slightly different interface, I presume most permedia cards were for PCs not Amigas.
It never made it to mass production from what I recall, but they could at least honor the legacy of 3dfx. Maybe there are still some on ebay...
Its nice to see Kyro and KyroII, but they never sold maybe more few 10Ks. **
When exactly did video processors become obsolete? It's obvious that if you have a new CPU and an old GPU that the GPU isn't going to help much, but otherwise it seems that GPUs of the time were always faster than CPUs, starting with Voodoo's.
Member of the 7 Digit UID Club
Just a few things off the top of my head...
-The Rendition Verite (V1000) was the first "true" 3d card that I can remember. It was capable of running accelerated Quake and the included Indy Car Racing II was breath-taking at the time with it's high framerate (30fps), high resolution (640x480), and bilinear filtering. I remember sitting in amazement as I watched the ICR2 demo play when I first bought this card. It was truly the first amazing 3d acceleration I had ever seen in my life. Probably the closest thing to it was the Race Drivin/Hard Drivin arcade game, with all of it's blocky polys and 10-15fps goodness.
-The article gives us a huge list of 3d cards, but little depth into the actual performance or market penetration of any of them. Most of these cards are barely worth mentioning. Even at the time, some of these cards sold in miniscule numbers or performed horribly. That's not to say that they shouldn't have been mentioned, but how about focusing on the major players a little more?
-How about 3dfx's "16-bit color is just fine!" rhetoric when Nvidia upped the ante with 32-bit color? I think the reason that 3dfx neglected to move to 32-bit color at the time was because they had tested it and performance on their cards was horrible. Hey, I bought 3dfx's bullshit at the time, mainly because I wanted to run Quake 2 as fast as possible.
I could go on, but the bottom line is this article isn't much of a "history" at all. It's nothing more than a list of cards with specs and brief summaries. How disappointing.
Looks like that can do SLI too! I want 2 just so I can play Duke Nukem 12: Your mum says Hi
If I was witty I'd put something funny here but, as it stands, I am not and have just wasted seconds of your life
What's the matter, scared?
I've got a 'genuine' Hercules Voodoo rush integrated 3d card (which had a VESA feature connector which I wanted, long story, didn't work.) to test with.
I have to say, this article didn't sufficiently emphasize the importance of the introduction of the GeForce and the GeForce 3. Almost every other graphics card was just "more" and "faster", but not the huge game-changing revolution that these two graphics cards represented.
Before GeForce, everything was all about accelerating rasterization - the act of filling in triangles.
With the first GeForce, lighting and transform was put into silicon. This was *huge* - this means that real math processing units were put into hardware. Scene complexity went up drastically, since we were finally able to push a lot of the more expensive operations into hardware.
With the GeForce 3, we had the introduction of the *programmable* graphics pipeline. This was a huge game changer - for the first time, the developer was limited only by their own intellect and creativity what kinds of things could go into the hardware. This was the beginning of what could be considered the first mass produced commercial stream processing unit. The graphics card has become a general purpose computational unit, a blazingly fast computational unit with applications into fields that have absolutely nothing to do with computer graphics.
I'm not sure what the ultimate evolution of the stream processor will be, but it still has the potential to really change the fundamental architecture of how future computers will be designed. Stream processors might eventually displace CPUs as the main computational workhorse in a computer.
--
#include <malloc.h>
free(your.mind);
There is a much better history of graphics cards at: http://accelenation.com/?ac.id.123.1&CFID=5425096&CFTOKEN=25318798 IMO of course. It has a lot more details and gives you a better sense of the "mood" of the industry
This was actually a fairly important piece of silicon (CL-GD5465). At Meltdown '97, this was the chip that powered Microsoft's first "Memphis" (Win98) demo as well as Intel's prototype Klamath/Pentium II chipset's special feature: AGP. I actually have a machine that has one of these onboard; An IBM Personal Computer 300GL. It's a P2-333 (originally 266) with a whopping 256MB RAM. It's my home domain controller, but originally it was one of my daughter's computer. It DOES support 3D under 98, it's not great, but it's also not a Virge. G-Police played reasonably well on it as did Incoming.
Fifty watts per channel, baby cakes.
Yeah I remember Myst. In fact at the moment - with a little bit of struggle - I'm replaying Myst III. The struggle is that this Win98 game is being played on an Athlon64X2 running Windows XP-MCE with a nice $50 ATI HD 4550 card. Made Myst V play great, but even setting Win98 compatibility mode Myst III still presents a couple hiccups.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
I had one in a combination with Matrox G400 dual-head. Ran Quake II in OpenGL like a charm. Actually that was the only thing it did.
Best $30 I ever spent back in 98 or so.
Wearing pants should always be optional.
I'm getting 1000fps on my 4x3 pixel monitor with all the eye-candy turned on.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
I love that on the scale it reaches 420
Hell of a lot cheaper than the 3Dfx cards of the same time, and VQuake looked spectacular.
Unfortunately, they couldn't keep up when Quake2 came out, and the V2100 was a total dud.
I still have every 3d card I've ever bought, from the Voodoo 1 with its clacky relay and passthrough cable through to my current Nvidia 9600gt. Add the cost of them all together and they probably owe me a very nice car.
Hal Spacejock: Science Fiction with Nuts
As the original gamer generation(s) age, they have an ever expanding sense of what a videogame can acceptably look like. While a gamer of a more recent generation may (tho not always, of course, god bless them) find it difficult to make the visual transition to ye olde games, older gamers in particular often find they are as happy with decades-old depictions of a gameplay evironment as some more cutting edge.
With that in mind, may I recommend a brief smattering of games from genres or eras that we often forget:
The good old days of insane fast paced simple, straightforward multiplayer FPS, Sauerbraten: http://www.sauerbraten.org/
Incredibly deep dungeon creation/management simulation, in glorious ANSI, Dwarf Fortress: http://www.bay12games.com/dwarves/
Straight-up honed turn-based hack-n-slash roguelike, Angband: http://rephial.org/
Kooky realtime multiplayer roguelike dungeon crawl (yes, multiplayer and realtime) MAngband: http://www.mangband.org/
After decades of gaming, I think many of us come to realize that its the quality of gameplay that matter far more than fidelity of depiction. ASCII kobolds and twitch rocketlauncher firing FTW!
Was for me too -> http://www.google.com/search?gbv=1&hl=en&q=APK+3dFx+Tuning+Engine&ie=UTF-8&sa=N&tab=iw &/or http://www.google.com/search?gbv=1&hl=en&q=APK+3dFx+Tuning+Engine&ie=UTF-8&sa=N&tab=iw
("Savoir Faire IS everywhere...")
Well, used to be @ least (was fun) - only I learned that hardware KEEPS CHANGING, & so do its programmatic parameters, + more, so, imo @ least? Programming around that is a chore that NEVER ends... but, still fun for me around the 3dFx Voodoo family, which imo? Really started it all, for 3dGaming!
APK
P.S.=> From a Voodoo I by 3dFx in 1999, to a NVidia GeForce 8800GT OC by EVGA - things are a LOT more powerful nowadays, & that is great imo! "Oh, the times, they are a changin'..." (I'll look back on this 5++ yrs. from now laughing @ what I have now most likely in the way of a videocard)... apk
Yeah, Myst. That was a great game, not the least of which was the *suspense*. I.e., the suspense of waiting for the f*&%ing CD-ROM drive. I owned a "blazingly fast" NEC Intersect 2x CD-ROM drive (yes, 2x; yes, external SCSI). I'd click on something, that thing would thrash around for a few minutes, and then I'd get that wonderful HyperCard-dissolve-into-next-eerie-scene. I remember my parents angrily telling me to "go to bed" because the NEC was so damn loud and was keeping them up. It was fun, but, thankfully storage is relatively fast and cheap now.
That joke made my sense of humor murder my sympathy.
"P.S.=> From a Voodoo I by 3dFx in 1999, to a NVidia GeForce 8800GT OC by EVGA - things are a LOT more powerful nowadays, & that is great imo! "Oh, the times, they are a changin'..." (I'll look back on this 5++ yrs. from now laughing @ what I have now most likely in the way of a videocard)... apk" - by Anonymous Coward on Tuesday May 19, @10:43PM (#28021099)
Yup, it's been a LONG weekend, & I am beat from building up a fairly big garden today (laid in with brick border, around 200 of them, & large stones (for keeping plants warmer @ night since they hold heat + they stall weedgrowth around the plants), I also hauled 15 buckets full of those stones today, & then laid them out inside the brick borders in the sun - was tiring, thank goodness it was afternoon sun though (5 pm or so)
Anyhow?
Yes - You'd most likely feel it too & it's getting near 11:30 here now, so I am tired: Thus, this is me, correcting MYSELF as regards my dates above in the quoted section - The Voodoo I? Hey - it was earlier than 1999, it was 1997 iirc for the Voodoo I by 3dFx...
APK
P.S.=> Sorry about that gents - lol! I've got to correct MYSELF, before one of you nitpickers here gets me... apk
"Fuck NVIDIA's proprietary and undocumented hardware!"
I do, everytime I try to run my linux drivers on it... hah, oh yes, I went there
(just kiddin, I only ever have problems with ati's drivers (windows and linux) and i've always been lucky enough to never have problems with nvidia drivers, (again, windows and linux))
The revolution will not be televised... but it will have a page on Wikipedia
We're there already. A GeForce GTS 250, for example, has 128 cores.
Ok.
Actually, I have a very different view of where we might be without dedicated graphics hardware.
The introduction of hardware acceleration removed the incentive to develop real-time software rendering techniques.
All 3D hardware has been designed for just one rendering technique: Drawing triangles one at a time with a depth buffer.
Couple this with the problem that for a long time none of the major graphics architectures were open (which prevented any APIs apart from OpenGL and DirectX to be developed), and you're left with a very stagnant set of usable rendering techniques.
There hasn't been enough effort in real-time ray tracing, and very little in combined approaches, (eg using the standard per-triangle method to compute the 1st ray when performing ray tracing). It's very rare to hear news of new rendering algorithms being developed, and yet there's plenty of scope for more.
We've also been limited to not using procedurally generated models, given only a few arbitrary blending modes (thus making transparency far uglier), been forced to use only a handful of supported scalar types (which varies depending on whether you're calculating vectors or pixels), and a million other issues that programmers would've avoided by using a CPU. And your games and applications have surely suffered.
At the bottom I will describe a few techniques that I would like to use, which would be possible on a fully programmable CPU, but are not (AFAIK, at least they weren't a year ago) possible on current hardware.
Now, compare the reality above with what we would have had if we hadn't used GPUs:
- We'd've focused very early on on high degrees of hardware parallelism, meaning overall computing power would be higher and rendering frame rates would probably be comparable to present.
- We'd've developed many new and superior rendering techniques, like those mentioned above.
- We'd've been able to write our own APIs, instead of having to choose between Windows-only (DirectX) or inferior feature set (OpenGL), both of which have less than optimal coding interfaces.
- And lastly, there'd be far more Linux gaming as developers would have no reason to force a proprietary single-OS API on us (DirectX), which is the primary reason for the lack of uptake of Linux amongst gamers.
Now we're heading back in the right direction:
- GPUs are becoming more general purpose chips (CUDA etc).
- We're possibly seeing the first steps in a return to memory sharing between graphics and applications (like with Larrabee). Of course we need to increase memory bandwidth to make this worthwhile.
- Next we'll get access to the raw instruction sets, and the ability compile and run arbitrary code on these chips.
And then we'll be back to square one - (specialist) CPUs doing our graphics.
Yet we could've gotten here far earlier if we didn't take the GPU diversion.
In my opinion, what we really did is we sold out on the long future of rendering and gaming for a few years of good times.
ASIDE: Two (of thousands of) techniques that we could be using now, if it weren't for GPUs.
1: Per-pixel curved surfaces (Eg spheres, cones, spline planes).
For these you just need a programmable depth value and a programmable "edge" shader that allows you to specify the start and end pixels of each line.
Eg:
A sphere 100 pixels round in an orthographic projection can be drawn in 1 pass with no unnecessary pixel calculations by:
Calculating the edge start and end points per line as sqrt((50^2) - (Y - CentreY)^2).
Per pixel, calculate the Z by CentreZ - sqrt((50^2) - ((X - CentreX)^2) - ((Y - CentreY)^2)).
It gives a perfectly smooth sphere and it renders far faster than multiple triangles.
Last time I tried, with OpenGL, about a year ago, the above was impossible, although I did make it work partially by ignoring depth and using a triangle that fully enclosed the sphere.
2: Subtraction from surface. (Not sure if there is an official name).
You agree with me.
Yep, I was right. The Mystique they have listed is the G200 Mystique. The author seems to have the old Mystique/Mystique220 mixed up with the G200 Mystique.
The original Mystique was based on the MGA1064SG chip. This was one of the first chips to have an internal RAMDAC. The internal RAMDAC ran at 170MHz on the original Mystique, and 220MHz on the Mystique220 (which I have). The chip name on the 220 was also changed to MGA1164SG, even though the only change to speak of was the increased RAMDAC speed.
The Mystique G200 listed in the article, however, is based on a completely different generation of chip. The G200 Mystique has a RAMDAC speed of 230MHz versus 250MHz for the Millenium G200 (which I ALSO still own). The chip that the G200's used was simply called MGA-G200 (or, in my case, MGA-G200A-D2... not sure what the difference is).
Fifty watts per channel, baby cakes.
I'm currently running nothing but old cards..
Matrox Mystique 220 w/Rainbow Runner Capture card
Matrox Millenium G200 (8MB) - NLX form factor
Matrox G450 AGP (16MB) x3 - NLX form factor
Matrox G450 Dual Head AGP (32MB)
Matrox G450 Dual Head PCI (16MB)
ATi Rage Pro Turbo 2 AGP (8MB) NLX form factor
Cirrus Logic Laguna 3D (4MB on motherboard)
Nvidia Quadro AGP (64MB) NLX form factor
Nvidia Geforce 2MX AGP (32MB)
Nvidia Geforce 6200 AGP (256MB)
I keep the NLX form boards around 'cause all of my kids' PC's require it for AGP use (IBM 300PL's, 2 towers and i desktop). One of them has the Quadro, but only for a game that the two of us play together online. The other three rarely every use 3D so the G450's work fine for them. I have two more machines that also use the NLX form factor. One has a Rage in it, the other the G200. My wife uses the G450 DH, though she's only using one display. Eventually, I'll upgrade them all to 64MB Quadro cards (Elsa Gloria II's) since they're so cheap these days.
Fifty watts per channel, baby cakes.
And interesting blast from the past. But whatever became of Matrox's Headcasting technology and why don't we see more of it?
Shai Schticks:"You don't make peace with friends, you make peace with enemies"
You're still using a Mystique 220. How about the original Millenium? This is from lspci -v:
00:0a.0 VGA compatible controller: Matrox Graphics, Inc. MGA 2064W [Millennium] (rev 01) (prog-if 00 [VGA])
Flags: stepping, medium devsel, IRQ 12
Memory at d2000000 (32-bit, non-prefetchable) [size=16K]
Memory at d3000000 (32-bit, prefetchable) [size=8M]
Expansion ROM at [disabled] [size=64K]
This is in a server though but still, the card is 15 years old now. I got my money's worth on that one, that's for sure.
"Doesn't Support Opengl or DirectX nor fits any case known to man"
A friend of mine had this card back when it was released back in 1996. I think that was the first 3D card that I ever saw in action. Came with a game, I believe it was Virtua Fighter. Only a few Sega games were ported for the card. It cost him around $950AUD.
3DFX had some funky dithering algorithm in their drivers that they claimed allowed their 16-bit framebuffer to output something that approximated 22-bit (yes, 22-bit) colour. This was possible, iirc, because the card rendered internally at 24-bit and then dithered down before passing to the framebuffer.
It did work, kinda. In fact, if you're lucky enough to have a system with a Voodoo2 (or similar) lying around you can see for yourself. Run a game (Thief 2 or Quake 3 or something) in 16-bit on a modern card, and then on a Voodoo2. There is a perceptual increase in quality on the Voodoo2, at least to my eyes. Colour gradients are smoother, there's less banding. A 3DFX card is probably still the best way to play those old games that only had 16-bit colour support.
don't forget their 256x256 texture limit size, i love my voodoo banshee but
it's not all bad business that made them go under
Okay, this is probably hopelessly offtopic, misguided and just plain wrong...
but...
IIRC once upon a time (about 20 yrs ago) there was a specialized hardware card that you could put into a PC that had a chip that ran forth native operations in hardware. Apparently it blew the doors off any other CPU type thing of the day. It was reported that it was difficult to get this special CPU over 5% usage because it spent most of its time waiting for your PC to catch up. However, being quite specialised and expensive it didn't catch on before Moore's law caught up.
But I wonder what would be possible if a GPU manufacturer actually implemented some kind of Forth engine in one of their chips??
At the lowest level the forth language amounts to about 60 very low level, very primitive ops such as "push", "pop", "inc", "dec", "0branch" etc that operate on a simple stack machine.
Out of simple commands, you then created more complicated commands which were just pointers to the simple commands in the library.
Since GPU's are chips that do simple things very, very quickly, imagine what a "forth primitives in GPU hardware" device could achieve!
Dunno, just guessing.
OTOH consider how many things have become mostly software-based after quite some time of needing to be implemented in hardware.
One that hath name thou can not otter
I've got an old 3D labs wildcat which actually looks very similar to that board....
From page 4:
Yep, it's the curse of the F-G siblings...
Each was built with limited programmability and on near-current but cheper technologies, to get a good cost/performance ratio, and then either did not advance, thus becoming shelfware, as described, or instead or gained more programmability sufficient to make the ma CPU in their own right. In the latter case, someone then invented a new DPU to offload the CPU of all that annoying graphics stuff (;-))
--dave
davecb@spamcop.net
Technical journalists would also do well to avoid Nvidia's "Graphics Processing Unit" coinage (with the dividing line they placed between DX6 and DX7 for marketing reasons). For all the impressive ALU power and "GPGPU" frameworks of today's graphics accelerators, they are not autonomous enough to warrant the GPU moniker. (Or else I want my mainboard's NPU, APU, SPU, UPU acknowledged as well...)
You need the Bitchin'fast3D2000
Wow! Imagine a Beowulf cluster of these!
Dude... you're playing word games and can't accept that you're wrong.
At it's heart it may be a crippled Voodoo2, but it's not a proper Voodoo2. There was ONLY one PROPER Voodoo2, with a pass-through like EVERY OTHER Voodoo2 card, ever made and that's the Obsidian S-12. The Banshee is not a V2 proper. You know what I mean, don't play stupid. You were horribly wrong about the S3 Virge and MeTaL and now you wanna try to redeem yourself here???
log off.
Fifty watts per channel, baby cakes.
troll penis.
last post