When Martok and I sent in the registration for our Voodoo5's we were amazed to see that they were actually accepted and sent to someone, meaning nVidia is still collecting 3dfx product registration. When the questionnaire asked what games we played, we replied "Anything with Glide". It looks as if nVidia took us seriously! Well it may not be just us who have spammed that to them however (hehe :D ). The NV30 which is rumoured to be called "Omen" and not "GeForce" has a feature called "NVBlur" that is backwards compatible with Glide and possibly a new version of it as well. Many people claim that Glide in 16-bit is better looking than OpenGL in 32-bit (not to mention DX :? ) in many games and I agree with them. If you have superior image quality with less performance hit, isn't that a good thing? Looks as if nVidia is doing something worthwhile considering Ati are now serious competition. Down below is a list of the NV30 specs, remember that 750MHz DDR2 isn't that high compared to what the industry has to offer, Samsung will be making 1GHz DDR2 very soon and the Omen isn't expected until either before or after Christmas so 750MHz will be considered typical for new cards. Also remember that the R300 (Radeon 9700) is compatible with DDR2 and is expected to use higher clocked DDR2 ram as well as a 0.13 core around or soon after the NV30 release. Will it be called Radeon 9900? It seems likely as they are also releasing a lower Radeon 9500 soon. The external TT&L (formally T&L) chip is interesting. 3dfx did that with the Spectre 2000 and 3000. It simply means that you have the TT&L chip separate to the graphics processor(s). What's even more interesting is that there are rumours around saying that the some (or all) NV30 models will have odd amount of chips on their cards. This means that nVidia will have a tri-chip card with TT&L and two GPU/VPU/TMU's like the Spectre 3000. nVidia wasn't wrong when they said that the NV30 would incorporate 3dfx technology. Maybe this will force Ati to make the 9700 MAXX that they said they wouldn't. 8)
AGP 4x and 8x Copper Core GPU Processor 0.13 m default clock 450 MHz 73 million transistors 8 pipelines 128 MB DDR-SDRAM @ 375 MHz x2 = 750Mhz and 256 bit 370Mhz RAMDAC Direct X 9.X compatible Quad Vertex Shaders Dual Pixel Shaders 85 bone capable Lightspeed Memory Architecture III Quad Cache memory caching subsystem Dual Z-buffers for greater compression without losses Visibility sub-system NvAuto advanced pre-charge Advanced Hardware Anisotropic Filtering - A new 12nvx mode that delivers improved subpixel coverage, texture placement and quality. 12nvx mode should deliver 80% better subpixel coverage and placement than previous modes Hardware bumpmapping 4 Textures/Pixel in a single clock cycle "loop back" can also apply an additional 4 textures per rendering pass A separate second copper processor running at 450Mhz an 0.13m as well - handle all T&L now called TT&L (True Time and Lighting) NVAutoShaper a new prechached and precacheable system Programmers can also add shapes to be temporarily stored in precache to speed things up Latest Open GL Backward Glide game capable thanks to NvBlur NView Multi-display Technology Built in re-programmable DVD and TV decoder chip New focus on computational efficiency rather than memory efficiency
Well i always thought glide was better. Back in the day.. *cough* when quake 2 first come out i saw it using glide...and it looked beautiful (rain guns trail was smooth and not just dots), the riva cards weren't bad performance wise...but the quality just wasnt there. Finally they might be doing something about it??..
OMG OMG OMG OMG OMG OMG OMG OMG OMG OMG OMG OMG OMG OMG OMG OMG Nvidia doing something right? And yes Glide is much better than OpenGL :P
Quote: This means that nVidia will have a tri-chip card with TT&L and two GPU/VPU/TMU's like the Spectre 3000. nVidia wasn't wrong when they said that the NV30 would incorporate 3dfx technology. Maybe this will force Ati to make the 9700 MAXX that they said they wouldn't.
Quote from Flammable: I was under the impression that Opengl was based on Glide.
OpenGlide ?:P A More generic version of Glide It's more like a rip off of it, with lower performance and worse image quality. Glide is different to OpenGL, but in many respects its very similar I have been told. For 3dfx cards the graphics settings come under either DirectX or Glide/OpenGL, nVidia will probably do the same. nVidia may also use a name I have heard only once "Demian", it may be a code or something that will turn out to be nothing, but it also may be the new MX cheapie cards. Who knows we'll have to wait and see.
Aravi I get my sources from many different sites and piece them all together to form my discussion here. As nVidia haven't officially released any information yet all of the above information has the possibility to be incorrect, however I doubt that very much. Some things like clock speeds and fill rates ect will most likely change by the release date. ATi are still deciding which feature to through into the Radeon 9500, any graphics card company can change their minds along the way so even if the above specs were real they may change anyway.
Quote from Seqo_ae: still can't beat that Anisitropic Filtering no performance loss...;) Hmm i think i read something about this, they used special drivers to achive this by forcing the user to use lower settings without telling them, ill have to go back and find the artical
Both Nvidia and ATI make their drivers to run at lower detail in 3dmark to acheive a higher score and Nvidia also include this for Q3 (not sure about ATI for Q3)
BTY the Radeon 9700 and Omen both have no performance hit with 4xAA (and probably some settings of AF). ATi are good at that sort of thing, but I have to say that the Spectre had no performance hit about two years ago. Here is the piccy of it, maybe if a tri-chip Omen comes out it may look something like it. [Image: http://members.optushome.com.au/bobsmurf/pics/rampagecard.jpg]
or if you can get UT running in OpenGL with DXTC/S3TC running then...wow.
and...huh? lower settings...for what? the card? clock speeds etc?
don't understand, but i am sure that this cannot be the case for the original radeons. i am also pretty sure the 8500 don't have this either...
can't imagine any game without AF....it would just be...bleh~
(this was my first reaction after going from a 233 with 4mb S3 non-3D accelerator. i was shocked when i ran half life with these weird lines appearing...then Q3...then UT. i thought there was something wrong with my video card. rang up the retailer and asked them what the problem was. etc etc got AF switched on and haven't gone back since :D)
By lower details i mean just that. The drivers for the Nvidia cards are programmed so that if you are running 3dmark or Q3 it will lower the detail level so that you get a higher score in the case of 3dmark and higher frames in th case of Q3.
ATI put them in for the 8500 i beleive. Im not sure about the earlier Radeons because it was only discovered a few months ago
[quote=Seqo_ae]or if you can get UT running in OpenGL with DXTC/S3TC running then...wow.[/quote]
hehe.. perhaps I should clarify..
UT in Metal w/ the high res S3TC textures ;) Going on many peoples opinion that Glide is better than OpenGL/DX... Metal was ub3r sexy.. better than Glide IMO. I think it was only ever used in the UT engine though.. UT and WoT supported it, not sure about anythnig else.
Martok: You're right there.. they all do it one way or another.. Think of what was it SiS and their graphics chip rendering only every other frame or whatever it was.. something on Anandtech about it. Also things like optimising their drivers for 3DMark performance... lame. But it's the same with motherboard manufacturers as well. For example default voltage for 3.3v rail is quite often running at 3.5 volts which of course feeds your ram, AGP, etc. So they can.. optimise their BIOS/wring every last bit of performance out of their board and '[i]cheat/i]' to get higher scores in benchmarks. Whether or not you believe that's cheating or just required to maintain system stability these days it's upto the individual.. I couldn't really give a stuff :)
Asus don't use 133.3MHz FSB. Thy use 134MHz so when comparing two mobos with the same specs they run a fraction of a second faster, that's lame. nVidia use drivers that are optimised for 3Dmark (EvilMark or nVidiaMark you decide), as Martok has said. They load certain settings for each individual section of the test to squeeze every little bit of performance out, that's also lame. ATi discovered nVidia were doing it so they did sort of the same thing. Can you blame them? If the competition is getting much higher benchmark scores than they should wouldn't you? It's sad when the industry does dodgy shit like this as it&'ll only make people think a card is better than it really is. That means gamers like us. I don't really care because I never compare or even look at any EvilMark results of any card. You want to benchmark a card then benchmark it in the games you are playing or want play and have some performance reserve for future games.
Quote from Seqo_ae: if you can get UT running in OpenGL with DXTC/S3TC running I wouldn't know, I don't use either of them.
Quote from Seqo_ae: can't imagine any game without AF....it would just be...bleh~ Mind reader.
Quote from Seqo_ae: i was shocked when i ran half life with these weird lines appearing...then Q3...then UT It's called "aliasing/jittering/jaggies/stairs".
It looks like the third chip on the Omen may be a memory controller chip to provide a huge amount of memory bandwidth. That would concur with the rumour about nVidia working on computational efficiency rather than memory efficiency, as they will not need it because of the large raw memory bandwidth. If that's true then that means there will be; 1 TT&L chip 1 Graphics core 1 Memory controller This would also give then better cooling 8) as the chips are spread apart, not a bad idea although we all know where they got that from. :wink: ATi will be making MAXX cards for workstations ect, but it's also likely we (gamers) will see a 9700 MAXX if the Omen is fierce competition for the Radeon. *rubs nipples*
Mr Kwok Yuen Ho, the CEO and chairman of ATi Technologies has given a speech during the press announcement and presentation of the Radeon 9000 and 9700 at China... "ATi does not have plan to make Radeon 9700 to be based on 0.13 micron process but the next generation R350 will and it is scheduled to be launch by first half of next year and supports DirectX 9.1" "Based on the Radeon 9700 technology with parallel connection of up to 256 (R300 VPUs), ATi is currently examining the possibility of having Dual Radeon 9700 VPUs on a card. Also Quad display outputs is possible with such card. However, there is no official schedule on when such card will be launched. Most likely the target segment for Dual VPU card is for simulation market."
Um..."Most likely...simulation market", but what if the NV30 canes the 9700? As I have said before, competition is always good. :D
Yeah there is no reason for them to need to release a dual chip card. Basically ATi is releasing now but NV30 won't be out for 3 or 4 months. That gives them a huge head start and it's quite likely that Nvidia will lose a lot of ground here. They're probably not all that likely to make it back too quickly either.
The original maxx cards were expensive enough back then... add to that the massive power requirements of the two chips and then designing a board to stick everything on, etc, etc... too much money for too little gain. It would be a high end card and high end means low demand... maybe they will but I highly doubt it.
Quote from Imitation: Yeah there is no reason for them to need to release a dual chip card. Basically ATi is releasing now but NV30 won't be out for 3 or 4 months. That gives them a huge head start and it's quite likely that Nvidia will lose a lot of ground here. They're probably not all that likely to make it back too quickly either.
The original maxx cards were expensive enough back then... add to that the massive power requirements of the two chips and then designing a board to stick everything on, etc, etc... too much money for too little gain. It would be a high end card and high end means low demand... maybe they will but I highly doubt it.
Here's another point. Doesn't the Radeon 9700 require an extra power connector already (Molex??)
Martok, i was referring to Vinc's post..sorry for the misinterpretation..
lower setting is a little different to lower details.
from what i know of the ATi catalyst drivers you force the higher detail level. you can set the card to 'performance' or 'quality' detail, where the card auto detects what would be better performance or better quality.
i use custom settings... forcing the higher quality...so i lose frames in just about everything.
i tested this with 3Dmark 2k1SE and what settings you choose makes a difference...(in the range of 750 points...on an orginal radeon that is huge...) and you can see some of the detail drop...not huge amounts that i could notice...but it was there and became annoying.. :)
Quote from Imitation: too much money for too little gain. Unlike the Rage Fury MAXX or Voodoo2, 3 or VSA100 SLi products that would deliver around 1.5-1.8 times the performance of a single chip when running with two chips, ATi have developed technology to make multi-chip configurations work more effectively such as around 1.9-2 times the performance of a single chip. If people really want such a card then they will have to pay for it. They are making MAXX cards for visual systems regardless of weather they make a gamers card or not so we will see them. I'm just saying it would be nice to see a gamer level MAXX card again regardless of the cost.
Quote from Jay: Anything that guzzles a lot of current way above standard just about goes into my bad books automatically. So your saying you don't like most hardcore visual simulation cards then? Sometimes companies put an AGP Pro connector on their cards for more stable current, but not every motherboard has one of these. It was reported that the GF256 was a "guzzler" and was unstable without the Pro connector. nVidia have had some problems with some motherboards not supplying enough current reliably through the AGP slot for the GF4 and that is a bad thing. Some people solved this as some motherboards have options to increase voltage ect, but you shouldn't have to do that and not all motherboards have those sorts of options. So ATi put a floppy connector on the 9700, I think this is a good idea and one of the many reasons I like the 9700. Vinix have put a molex connector on an SiS Xabre 400, not because it needs it, but because they believe its more stable and therefore more reliable.
I think, companies striving to be "the" speed king is a good thing. As as soon as they have the fastest card on the market, or one of the fastest, they now have to technology and the options to create lower market cards easily. Having the fastest card on the market, for any amount of time also gives a name for your company that people will respect instead of think of you as one of those "Hey, I heard of those guys, are they some kind of OEM company?" companies :P
Quote from Odyssey: nVidia have had some problems with some motherboards not supplying enough current reliably through the AGP slot for the GF4 and that is a bad thing.
Would this be the insufficient AGP signal strength on Via based motherboards problem?
No AGP 4x unless you explicitly enable it in the driver.
Quote from Odyssey: It was reported that the GF256 was a "guzzler" and was unstable without the Pro connector. nVidia have had some problems with some motherboards not supplying enough current reliably through the AGP slot for the GF4 and that is a bad thing. Some people solved this as some motherboards have options to increase voltage ect, but you shouldn't have to do that and not all motherboards have those sorts of options. So ATi put a floppy connector on the 9700, I think this is a good idea and one of the many reasons I like the 9700. Vinix have put a molex connector on an SiS Xabre 400, not because it needs it, but because they believe its more stable and therefore more reliable.
Quote from Jay: Look Odyssey, don't start this video card bullshit again. I think everyone's had enough of it. Settle down and shut up.
Jay he is saying that just because its power source is a plug direct from the power supply doesnt mean it guzzles power. He is also stating a fact that many people considered some Nvidia cards as power guzzlers, just as you labeled cards that have a molex connector as power guzzlers.
You started it, he finished it by stating facts that some cards guzzle power without the use of an external connector :P
Quote from Imitation: Odyssey: Limited market. Yes that's correct, however how can I explain this to you in a simpler form.....hmmm, how about this.... The Car industry is just like the computer industry, they make their top models the best of the best regardless of price. How many people like Ferrari for their F50 and other top models? How many did they sell and for what price? It doesn't matter because having the best car out will increase their lower speed and priced cars because that company has the crow of "the fastest car". How many people heard that the GF2 was the fastest card out at the time and went straight out and bought a GF2MX card? A shit load. Now the 9700 will be out shortly, how many people are looking at 8500 and 9000? A shit load. The point of the story is the companies fastest card will help sell their lower cards regardless of the price of the highest model, that's why they should make a 9700MAXX.
Yes I know that.. but this isn't about why they should.. it's about if they will and I will agree that it's quite possible that they might do this. However until we see what happens with NV30 there doesn't look to be much reason for a 9700MAXX. 3 months after NV30 comes out ATi would be scheduled to release their next product. Because nVidia has pretty much fucked up and missed a product cylce it upsets things a bit... now it's time for nVidia to play catch up - at least for the not so distant future.
Like you know.. NV30 is .13 micron so that should give nVidia a chance to crank up the speeds to regain some lost ground to ATi. But in the time that it takes NV30 to come out ATi may have perfected their process enough to run .13, crank up their speeds and negate any performance gains NV30 ultra/pro/etc.. will provide. Basicaly nVidia have fucked up and it's going to be something that will take a fair while to recover from. Anyway there are so many things that could happen and so much that we don't know about there isn't much point in speculating. I'll shutup now.
Ati have stated that they will be making MAXX cards, we're just not sure if they will be coming to the gamers market. The 0.13 micron Radeon (R350) is supposed to be out about Q2 or Q3 2003 (probably midyear). I don't think nVidia have "stuffed up" as you have stated. They are bringing a card out that will be worth the wait. For the time being their current cards are selling rather well so they have nothing to be worried about....except for one thing. They are intrusting the 0.13 micron process of the NV30 to a company that has never produced 0.13 micron chips before, a gamble in my mind. Ati have done the right thing to stay away from it for now until 0.13 has had all its nails ironed out.
My position is this. SiS will be making a Xabre II with DX9 support (Q4 2002), but that will most likely be aimed at the budget market. Matrox have only released one of the four Parhelia products they promised (where are the others?). 3Dlabs P10 is interesting, but will be aimed at the professional market. Ati have a good product (R300) that currently could only be bettered by being multi-chip. Nvidia's NV30 looks good on paper and I would be very surprised if they screw it up.
I'm currently sitting on the fence waiting for both the R300 and NV30 to come out before I would decide on which one to buy. I will watch them both very closely because it's almost guaranteed that one of them will soon find a home in my computer.
Everybody knows my opinion on past nVidia products, but if they make the NV30 pass the "Odyssey Test" then good on them, I'm hard to please. (pls note that was not a comment referring to anything sexual)