Kyro 2 the killer of nvidia ???

Page 16 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
G

Guest

Guest
Since my last post I have been thinking of some of the NON-technical challenges that the Kyro II will face:

1. Since the technology between the major cards so far has been fairly similar, we have certain performance expectations from each card. There have been some differences that took us a little by surprise, like HyperZ and T-buffer. But still most new card reviews go something like this: Well it has a higher clock speed, fewer pipelines but does more textures per pass... At this point most of us already have a pretty good idea of its performance. We then just adjust that expectation slightly for whatever minor new technology was implemented.

This mind set is not only in consumers but reviewers and developers too. Nearly every benchmark is written to test "standard" hardware, and is optimized for hardware that does things the "normal" way. This is not without reason. If new cards come out that offer much better T&L, it makes sense that we find benchmarks with millions of triangles cropping up. What we <i> won't </i> see in this situation are a lot of benchmarks emphasizing overdraw. What would be the point? If the cards do poorly with overdraw, then the applications will avoid overdraw whenever possible, so an overdraw benchmark would not prove anything meaningful.

Now along comes the Kyro line and does things COMPLETELY differently. No Z-buffer? That's impossible! Tiles? ...It is uncomfortable.

It also gives performances that are unpredictable. We shoot through one of our usual benchmarks and the Kyro comes up just short of the GF2 MX. Ah-ha! we say... Now we know where this card fits in. But on the next benchmark, one which we see as very similar to the first one, it beats the GF2 Ultra. Huh? We start to scratch our heads. More benchmarks lead to more unexpected results, and this leads to debates (like this one :cool: ) and fuzzy recommendations from reviewers and retailers. Some people decide to risk going with the new technology, but many people are looking for a clear-cut winner, and this usually means they want all the usual benchmarks dominated in all the usual ways. (Or they want the CompUSA guy to say, "Yeah man, this one kills everything else!") The Kyro II does not provide this–-it provides confusion. Buyers hate confusion. It's almost always easier to buy a card like the GF2 that, though it sometimes loses to the Kyro II, performs consistently well. Again note that the "inconsistency" of the Kyro II has to do with today's software and performance expectations, not with bad technology. But the average user doesn't see this.

So here is one uphill battle for the Kyro II: getting people, from consumers to developers to reviewers to retailers, familiar with and catering to tile based renderers, just like they already are with the technology of the last five years.

2. This is related to number one, but while that was more a mind set issue, this deals with the actual performance decreases that come from that mind set. Developers are going to tailor their applications to the hardware that is most popular. Sure, developers are in a position to know the true benefits of the Kyro II better than most consumers, and they will probably exploit those benefits whenever possible. But when it comes down to a tradeoff between performance on a Kyro II and performance on a GeForce 2, they will go with the GF every time to maximize their market. This could lead to REAL (vs. misunderstood or misplaced) performance issues with the new hardware, which would fuel the doubts of the already confused consumers.

Uphill battle number two for the Kyro II: Proving and <i> maintaining </i> its worth without the developer support that the competition gets.

Needless to say this is greatly compounded by the fact that Imagination Technologies is not yet a major PC player. Had ATI or NVIDIA released a tile-based card, I think the transition would be <i> much </i> quicker.

3. Uphill battle number three: dirty tricks from the competition. I won't spend much time on this one as it is a standard underdog challenge in the business world. NVIDIA is scrutinized, and consequently, criticized more because it is on top for the moment, but companies all over the world pull this kind of stuff every day. If you really don't like these practices then you should go gripe to your congressman or somebody else that could eventually, perhaps, make a difference against them. Currently, "moral corporation" is an oxymoron; singling out and flaming NVIDIA won't do anything except make them and everybody else work all the harder at hiding their dirty side.


In conclusion, I want to point out that over and over in history technology has won acceptance and market share due to politics, not performance. I have made my views clear already in that I don't think the Kyro II is the card to buy right now. But the card does have nice strengths, and the tile-based rendering it showcases has great potential. How are politics, advertising and the other various forms of BS going to affect this technology long term?

Comments? Flames? :wink:

Regards,
Warden

====================
A computer lets you make more mistakes faster than any invention in human history - with the possible exceptions of handguns and tequila. -Mitch Ratcliffe
 
G

Guest

Guest
well Tribes 2 was tailored for nvidia cards and...
kyro is better (more fps, average and minimum)
seriously Sam is another...
kyro is better...

there are too much development time lost because they (tim sweeney,etc... ) have to optimize for traditional cards (they want low overdraw).
and this is also responsible for many cpu time devoted to that.
I remember the issues of glide versus direct3d (tim sweeney was one promoter of glide and 3dfx if I remember) and look now... glide is history...
for most consumers out there the important thing is performance and quality and price, I think that kyro have all of this in a good mix, all the reviews are positive (I mean all). I remember the time when tnt-1 arrived there were lots of negative reviews towards tnt-1 (lack of glide was that feature that harmed tnt-1 the most) yet it's the best seller of nvidia...

you are making nvidia politics giving us confusion !
Nvidia is also unpredictable you never know when kyro will win to Nvidia cards or otherwise.

if kyro is unpredictable then also nvidia get's that unpredictablility. Don't you think ?
;)

and don't forget that unpredictability is with games that are tailored for nvidia cards !!!
if kyro get's that attention we will see even better games for the future !!!
(more development time for more important things)
<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 04/22/01 09:52 AM.</EM></FONT></P>
 

HolyGrenade

Distinguished
Feb 8, 2001
3,359
0
20,780
I tried to draw myself out of this argument, but I had to jump in to make this comment.

No other card was ever gonna have Glide. 3DFX were making sure of that. The only reason developers preferred Glide because Direct X 3 was plain crap!

Once Direct X got fixed up, it started getting more popular.

As for your comments about confusion and stuff, they didn't make any sense.


Warden:

I agree with you points and would like to add that all the big computer manufacturers seem to want "Graphics to the n'th degree" (corny, I know). They would make quite a bit of profit for nVidia. Imagination Tech will find difficulties getting in that area.




<font color=red>"My name is Ozymandias, King of Kings:
Look on my works, ye Mighty, and dispair!"</font color=red>
 
G

Guest

Guest
my comments don't make any sense ?
;)
maybe not for you !!

you said:
"Warden I agree with you points and would like to add that all the big computer manufacturers seem to want "Graphics to the n'th degree" (corny, I know). "

like Dell selling pentium 4 1,5 ghz with a tnt-2 m64 ?
lol<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 04/22/01 02:13 PM.</EM></FONT></P>
 
G

Guest

Guest
<<<<<1. Since the technology between the major cards so far has been fairly similar, we have certain performance expectations from each card. There have been some differences that took us a little by surprise, like HyperZ and T-buffer. But still most new card reviews go something like this: Well it has a higher clock speed, fewer pipelines but does more textures per pass... At this point most of us already have a pretty good idea of its performance. We then just adjust that expectation slightly for whatever minor new technology was implemented.>>>>>>

We only know what to expect from each card when we've seen enough reviews or used the cards ourselves. Before I bought the Voodoo5 I thought it would be allot faster, same with the Geforce DDR. You never know how any card will handle any game until you see a benchmark or play it youself, the Kyro II is not different there.

<<<<<<This mind set is not only in consumers but reviewers and developers too. Nearly every benchmark is written to test "standard" hardware, and is optimized for hardware that does things the "normal" way. This is not without reason. If new cards come out that offer much better T&L, it makes sense that we find benchmarks with millions of triangles cropping up. What we won't see in this situation are a lot of benchmarks emphasizing overdraw. What would be the point? If the cards do poorly with overdraw, then the applications will avoid overdraw whenever possible, so an overdraw benchmark would not prove anything meaningful.>>>>>>

What would be the point of emphasizing overdraw in a benchmark? To test how cards fair in more complex more real environments. To show developers they can program with more freedom, without having to design there games to lower overdraw. Croteam have done this with Serious Sam, have you played that game. Its full of overdraw all over the place. At one point your in a massive room (and I mean massive), the room has loads of supports (which obviously is needed in such a massive room), and loads of monsters running at you, and at any one time you can look and have 30-40 supports infront of you with about 80% of them being covered by the rest. The level of overdraw there is massive and thats not even taking into account the overdraw each monster causes. This sort of scene is very commen in Serious Sam, its a very complex game.

<<<<<<Now along comes the Kyro line and does things COMPLETELY differently. No Z-buffer? That's impossible! Tiles? ...It is uncomfortable.>>>>>>

It does have a Z-buffer, it has an on-chip Z-buffer that can perform HSR 32 times faster then a standard traditional Z-buffer. The tile based rendering is completely invisible to the app or game, so as far as the app is concerned the Kyro or Kyro II is a traditional. I know your trying to say that people will be wary of the tech but how many people buying graphics cards know much about Z-buffers?, or how a scene is rendered?, not that many. They look at the benches and features and thats enough for them. The ones that do understand about z-buffers and such will understand the explanation of the Kyro II tech. There's nothing uncomfortable about it.

Also remember that the method the Kyro II uses has been in several graphics cards over the years all from PowerVR. The tech is not new by any means (4-5 years old), the only difference with Kyro and Kyro II is IMGTEC finally got it totally right. So there's no more HW incompatabilities like there was with early tile based rendering cards.

<<<<<It also gives performances that are unpredictable. We shoot through one of our usual benchmarks and the Kyro comes up just short of the GF2 MX. Ah-ha! we say... Now we know where this card fits in. But on the next benchmark, one which we see as very similar to the first one, it beats the GF2 Ultra. Huh? We start to scratch our heads. More benchmarks lead to more unexpected results, and this leads to debates (like this one ) and fuzzy recommendations from reviewers and retailers. Some people decide to risk going with the new technology, but many people are looking for a clear-cut winner, and this usually means they want all the usual benchmarks dominated in all the usual ways. (Or they want the CompUSA guy to say, "Yeah man, this one kills everything else!") The Kyro II does not provide this–-it provides confusion. Buyers hate confusion. It's almost always easier to buy a card like the GF2 that, though it sometimes loses to the Kyro II, performs consistently well. Again note that the "inconsistency" of the Kyro II has to do with today's software and performance expectations, not with bad technology. But the average user doesn't see this.>>>>>

No it doesn't give unpredictable performance. The Kyro II cards being reviewed are sample cards and the reviews have all shown the same thing (appart from 1 or 2, but then they are sample cards and the same happens in Geforce 3 reviews or any other new card). It consitently beats the MX by a long way. In HW T&L games at low res the MX can win but that has nothing to do with TBR, its down to the fact that the Kyro II doesn't have HW T&L. In most games it gives performance around about the same as a GTS especially in higher res and 32bit. With FSAA it consitently beats the MX and GTS. Yes in some games it even beats the Pro or Ultra but how is this a bad thing? The important thing is its benches in different games don't very as to be slow in certain games. Its GTS, Radeon DDR performance levels in most games and Pro, Ultra levels in other games. The games that hit Pro and Ultra levels are just a bonus for anyone buying the Kyro II.

<<<<<So here is one uphill battle for the Kyro II: getting people, from consumers to developers to reviewers to retailers, familiar with and catering to tile based renderers, just like they already are with the technology of the last five years.>>>>>

All the reviews I've seen have been won over by Kyro II. Croteam have even supported some Kyro/Kyro II only features in there game Serious Sam. Its a big game and I'm sure lots of games will use its engine.

<<<<<This is related to number one, but while that was more a mind set issue, this deals with the actual performance decreases that come from that mind set. Developers are going to tailor their applications to the hardware that is most popular. Sure, developers are in a position to know the true benefits of the Kyro II better than most consumers, and they will probably exploit those benefits whenever possible. But when it comes down to a tradeoff between performance on a Kyro II and performance on a GeForce 2, they will go with the GF every time to maximize their market. This could lead to REAL (vs. misunderstood or misplaced) performance issues with the new hardware, which would fuel the doubts of the already confused consumers.>>>>>

Supporting one card does not neccesarily mean not supporting another. Developers would be supporting the Kyro II and increasing its speed when they add the Geforce 3's 4 texture layers in a single pass into a game.

This sort of problem (when its a problem) will happen for any new card thats not from Nvidia, and even card from Nvidia, look how long HW T&L took to take hold, why did it take so long?, because not all cards supported it. How many games are made to help Radeons HyperZ do its job effectively? Specifically the Hierarchical Z part which needs front to back sorting, not many if any.

<<<<<Uphill battle number three: dirty tricks from the competition. I won't spend much time on this one as it is a standard underdog challenge in the business world. NVIDIA is scrutinized, and consequently, criticized more because it is on top for the moment, but companies all over the world pull this kind of stuff every day. If you really don't like these practices then you should go gripe to your congressman or somebody else that could eventually, perhaps, make a difference against them. Currently, "moral corporation" is an oxymoron; singling out and flaming NVIDIA won't do anything except make them and everybody else work all the harder at hiding their dirty side.>>>>>

If I see Nvidia or anyone else do anything counter competative I'll say what I think of it. If they didn't want people to be angry about the [-peep-] there pulling then they shouldn't have done it or shouldn't have gotten caught. If you want a Kyro II but you would like a choice of boards makers to choose from with different ram configs and different prices and you can only find a Herc board because Nvidia's PDF or strong arm tactics stopped lots of cheaper board makers making a Kyro II, then you'd be pissed off and it'd be your right to show that anger. Why say well all companies do that, its perfectly fair?. What if your living in a country in Asia and there are no Asian board makers making Kyro II because of Nvidia's dirty tricks. You'd end up having to buy something else instead of Kyro II, not because Nvidia gave you a better option but because Nvidia forced you to buy a MX or something by scaring companies into not releasing Kyro II were you live. You'd be very pissed then, what Nvidia are doing will effect peoples lives and those people have the right to be as pissed off as they want if they are restricted in there buying choice because of Nvidia's attitude.

<<<<<<In conclusion, I want to point out that over and over in history technology has won acceptance and market share due to politics, not performance. I have made my views clear already in that I don't think the Kyro II is the card to buy right now. But the card does have nice strengths, and the tile-based rendering it showcases has great potential. How are politics, advertising and the other various forms of BS going to affect this technology long term?>>>>>>

Its the performance of a technology that will gain it acceptance. I have seen allot of BS about Kyro II but also every review I've seen is positive. The forums are buzzing at the moment, allot of people have had there eyes opened to tile based rendering and they like what they see.
 

HolyGrenade

Distinguished
Feb 8, 2001
3,359
0
20,780
That m64 is still making nvidia money. do you really think they give a [-peep-] as long as their customers are using their products?


<font color=red>"My name is Ozymandias, King of Kings:
Look on my works, ye Mighty, and dispair!"</font color=red>
 
G

Guest

Guest
that makes sense Now ! (Irony)

you said:
"Warden I agree with you points and would like to add that all the big computer manufacturers seem to want "Graphics to the n'th degree" (corny, I know). "

they want graphics to n'th degree or they want to make money?
 
G

Guest

Guest
more than 2700 views on this topic ?
:)
more than most reviews of tomshardware !!
lol
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Who could have predicted that one here. Especially when Tom hasn't even reviewed it here except in Germany. KyroII will make some people very happy and others maybe not so happy depending on the game they will be playing. Still appears that most games will fly on the Kyro2 especially the older ones which are plenty. DX8 needs to be fixed by Microsoft fast otherwise Kyro2 will start feeling the dissappointment in some games that has to write to the texture using the cpu vice the Kyro2 chip. I would buy one except for the lack of Linux support now which means a nVidia board is on my near future shopping list. I really don't need another Win9x/W2k game card but really a good linux card which nVidia cards shine at. Looking at a GF2pro but then again a mx may do the trick just fine in Linux for now since I am totally ignorant about Linux presently.
 
G

Guest

Guest
another kyro review this was not so favourable for kyro 2 ...
they used directx8 (we all know about that directx8 bug)
and a low end cpu (duron 700) and pc100

but even with these issues I think that this review proved the superiority against mx with these kind of system...

<A HREF="http://www.hardavenue.com/reviews/3d4500.shtml" target="_new">http://www.hardavenue.com/reviews/3d4500.shtml</A>
 
G

Guest

Guest
Teasy,
Man you totally misunderstood my post and made tons of arguments that weren't even necessary. A lot of what you responded with was agreeing with me and most of the rest had nothing to do with what I was saying anyway.
<font color=blue> What would be the point of emphasizing overdraw in a benchmark? To test how cards fair in more complex more real environments. </font color=blue>
Uhhh, that was a rhetorical question. I was using it to illustrate how there weren't many benchmarks that emphasize things not well supported by current hardware, because nobody sees the point in making them. I didn't say that there wasn't a need for this type of benchmark. I think you just assumed I was attacking the Kyro II and saying that overdraw wasn't important, when that wasn't my point at all. Remember, I was saying how it is going to be hard to get the masses to realize all these points in a timely manner, not that the points (like overdraw) weren't true or important.

Hmmm... well I was going to answer several other things point by point but I'm going to skip most of them as I don't know where to start (my post will be long enough anyway. :wink: ) Almost everything I said you took in a completely different context. For example, many of my statements you countered with technical arguments. Technology was not the point! The point was that there are many influences which are out of the control of the technology. Again, I WAS NOT ATTACKING THE KYRO II.
<font color=blue> It does have a Z-buffer . . . There's nothing uncomfortable about it. </font color=blue>
It does Z-buffer calculations on the chip, which is about as far from a traditional Z-buffer as you can get without getting rid of it all together. But again, this was a <A HREF="http://www.m-w.com/cgi-bin/dictionary?va=rhetorical" target="_new">retorical question</A>.

As for you thinking it is "comfortable," that's because you have studied the technology until you have become comfortable with it. That is great for you, but most people aren't there yet. My point was that it will be a while before very many people are there with you.
<font color=blue> This sort of problem (when it's a problem) will happen for any new card thats not from Nvidia, and even card from Nvidia, look how long HW T&L took to take hold, why did it take so long? </font color=blue>
That is exactly my point: new technology takes a while to catch on. When it is radically different and coming from an outside company (as in outside the mainstream PC graphics arena) then the problem is only worse.

Why did you even make this statement anyway? You seem to be coming to the rescue of the Kyro II by putting NVIDIA and ATI in the same boat with them, when I was clearly talking about battles that face ALL new/different technologies. I was applying this to the Kyro II because it <i> is </i> new and different. I never once said this was from a <i> fault </i> in the Kyro II, so there was no need to rescue it.
<font color=blue> If I see Nvidia or anyone else do anything counter competative I'll say what I think of it. </font color=blue>
Hey speak on man. Say all you want. I wasn't trying to interfere with your first amendment rights or anything. But I will speak my mind too, and there are two things that really bother me about most of the NVIDIA "strong-arm" flames:

[rant]
1. Whining without acting. If something is really that important to somebody then they need to put their money where their mouth is and try to change the situation. Whining on a forum isn't going to help <i> stop </i> the behavior. It will just encourage better <i> hiding </i> of it. Really I don't think anything will stop it, but I have respect for people that try to change things they see as bad. I don't have much respect for people who love to point the finger but won't try to solve the problem.
2. Don't assume that ATI or anybody else is a saint just because NVIDIA got caught. This one really makes me sick. Wake up to the real world boys and girls! You can talk all you want about the general lack of ethics in the business world, and give plenty of examples. But <i> very </i> few people are in a position to truthfully pick, say, ATI over NVIDIA as being more ethical. Most all of us have only the press and rumors to give us our information on this. We might as well pick a company out of a hat and start flaming it.
[/rant]

(By the way Teasy, these two remarks aren't aimed at you as I don't really know where you stand on this. They are aimed at anybody in general who wants to whine up a storm about NVIDIA being "bad" while other companies are "good," based solely on what they read in the press and hear in forums. Readers will have to decide for themselves if it applies to them.)

Last, and probably most important, as this may be core to your misunderstanding of my whole post:
<font color=blue> It's the performance of a technology that will gain it acceptance. </font color=blue>
Have you had your head in the sand the last 20 years or more? I completely agree that this is the way it <i> should </i> be. But it is simply not this way at all. Look back: VHS vs. <A HREF="http://toyvax.tucson.az.us/~vance/betaphile.html" target="_new">Beta</A>; Windows vs. OS/2; DirectX vs. Glide; digital audio tapes; DOS vs. all of it's imitators and practically any other OS; K6-2 vs. K6-3; the <A HREF="http://www.tuckerclub.org/" target="_new">Tucker '48</A>... (OK so the Tucker was impractical to build at an affordable price, but its technology was still outstanding. :cool: )

I don't very often just come out and say somebody is wrong, but in this case I'll make an exception. Your last statement is in complete denial of historical fact. Politics ALWAYS affect technology, and not often for the better. Sorry man... it sucks, but it's true. :mad:

Regards,
Warden
 
G

Guest

Guest
I just read the new 3d cards article over at anandtech. It says the Kyro 2 is actually competition to the GeForce 2 Ultra and the GeForce 3. What kind of bull-crap is that???!!!

Also, it says if you don't need a card for a long time i.e 9-12 months!

9-12 months is supposed to be a long time??? does that mean 2-6 months is a short time for owning a 3d card. How many people here upgrade every 3 to 6 months?

Me thinks, some one leans a bit towards the Kyro!!!


<font color=blue><b>Who is General Failure and why is he reading my hard disk?</b></font color=blue>
 
G

Guest

Guest
yes it's a competition for the high end geforce 2 !!!

if you go and find some reviews you will find that when in lower depths geforce 2 is better (but I think that even a tnt-1 can achieve playable fps on those depths )
if you start doing FSAA or higher depths kyro could mantain more consistent performance...
so kyro 2 will be playable in depths going from 640x480x16 to 1600x1200x32, or being playable at none FSAA to 4x FSAA .

Yes it's a budget price card with a high end performance...
and it does fine with todays T&L (like a high end geforce 2) if and only if coupled with good cpu !!!

<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 04/23/01 02:07 PM.</EM></FONT></P>
 
G

Guest

Guest
<<<<<<<<<Man you totally misunderstood my post and made tons of arguments that weren't even necessary. A lot of what you responded with was agreeing with me and most of the rest had nothing to do with what I was saying anyway.>>>>>>>

I think most of them were neccesarry. You said allot of things that I beleived to be wrong so I put my opinion in there. The post was directed at me after all.

<<<<<<Uhhh, that was a rhetorical question. I was using it to illustrate how there weren't many benchmarks that emphasize things not well supported by current hardware, because nobody sees the point in making them. I didn't say that there wasn't a need for this type of benchmark. I think you just assumed I was attacking the Kyro II and saying that overdraw wasn't important, when that wasn't my point at all. Remember, I was saying how it is going to be hard to get the masses to realize all these points in a timely manner, not that the points (like overdraw) weren't true or important.>>>>>>>

Really, lets go back to your original comment "What would be the point? If the cards do poorly with overdraw, then the applications will avoid overdraw whenever possible, so an overdraw benchmark would not prove anything meaningful"

If it was a rhetorical question then why actually put a question mark there?. Also you said a benchmark which is made to test how cards run with a certain level of overdraw would not prove anything meaningful, because any app would try to keep overdraw to a minimum. This is untrue. How do you keep overdraw to a minimum in outdoor scenes, short answer....you don't. Look at serious Sam, that game doesn't even try to reduce overdraw. The game was written to forget about the overdraw problem and make a good game, so a benchmark like Villagemark or Temple demo is meaningful.

<<<<<<<Hmmm... well I was going to answer several other things point by point but I'm going to skip most of them as I don't know where to start (my post will be long enough anyway.) Almost everything I said you took in a completely different context. For example, many of my statements you countered with technical arguments. Technology was not the point! The point was that there are many influences which are out of the control of the technology. Again, I WAS NOT ATTACKING THE KYRO II.>>>>>>

I answered allot of your posts with technology because in certain cases I wasn't sure if you knew certain stuff about the PowerVR technology. I also wanted to make sure others on the board took these technical points into consideration when reading your post. I don't think I took to much of it out of context even when I read it again now.

<<<<<It does Z-buffer calculations on the chip, which is about as far from a traditional Z-buffer as you can get without getting rid of it all together. But again, this was a retorical question.

As for you thinking it is "comfortable," that's because you have studied the technology until you have become comfortable with it. That is great for you, but most people aren't there yet. My point was that it will be a while before very many people are there with you.>>>>>>

Please do NOT point me to a dictionary, I know what rhetorical means. But how am I supposed to know if something is a rhetorical question if you put a question mark there. I just thought I'd mention the on-chip Z-buffer for people that weren't sure what you meant in your post. Not all my comments were necessarily pointed at you even though I was replying to your comments.

I also feel I explained myself well in this point. I pointed out that as long as the card emulates a traditional perfectly then tile based rendering in invisible to both the person running the card and the app (this is how the Kyro II works). People aren't going to be scared just because a site talks of tile based rendering. For the ones that don't understand what it means they simply think its a new feature just like HyperZ. All they look for is the performance figures and the general impresion of the reviewer on the quality of the card.

<<<<<<That is exactly my point: new technology takes a while to catch on. When it is radically different and coming from an outside company (as in outside the mainstream PC graphics arena) then the problem is only worse.>>>>>>

I'm sorry but I don't see your point here in relation to the Kyro II. Your talking about tile based rendering as if its like HW T&L or something saying it will take time to cacth on. Tile based rendering doesn't need to catch on. It doesn't need developer support to do what it does. There is nothing you can do to optimise for a traditional card that will hurt a tile based renderer like Kyro II.

<<<<<Why did you even make this statement anyway? You seem to be coming to the rescue of the Kyro II by putting NVIDIA and ATI in the same boat with them, when I was clearly talking about battles that face ALL new/different technologies. I was applying this to the Kyro II because it is new and different. I never once said this was from a fault in the Kyro II, so there was no need to rescue it.>>>>

I posted that because I felt your comments made it sound like the Kyro II would be slower in the future because its somehow incompatible in some way with games written for traditional renderers. This is wrong, anything you can do on a traditional is fine on the Kyro II and developers can optimise for a traditional design without hurting performance on the Kyro II. For instance there's nothing developers can do to stop games having a minimum of 2 overdraw in todays games. So it doesn't matter how much a game is optimised for a traditional to remove overdraw. They can't get rid of much of it in software and so the Kyro II will always have the advantage in that sense.

Overdraw is only going to increase even if developers keep making wastful algorithms to try to do in sofware what Kyro II does 100% perfectly in hardware.

<<<<<<Hey speak on man. Say all you want. I wasn't trying to interfere with your first amendment rights or anything. But I will speak my mind too, and there are two things that really bother me about most of the NVIDIA "strong-arm" flames:

Yeah and I'll speak my mind back. You were trying to say that people shouldn't flame Nvidia for what they've done. Why shouldn't I or anyone else flame Nvidia if we don't like something they've done, as I explained if something effects me I'll speak my mind about it.

<<<<<Whining without acting. If something is really that important to somebody then they need to put their money where their mouth is and try to change the situation. Whining on a forum isn't going to help stop the behavior. It will just encourage better hiding of it. Really I don't think anything will stop it, but I have respect for people that try to change things they see as bad. I don't have much respect for people who love to point the finger but won't try to solve the problem.>>>>>>

What you call whining I call speaking out and this does help. Speaking out and making sure this info is available on forums actually hurts Nvidia and gives them a very bad rep. They will try this [-peep-] less if they get such a bad rep that there sales are hurt. What could I do if I wanted to stop Nvidia doing what there doing, I don't even live in the country which they base themselves in. Also how did you know I hadn't done anything about it or that anyone else talking on a discusion board hasn't already done something about it, you don't know that so you shouldn't make that assumption. The reason I assumed you were directing the original comment to me was because I had mentioned something about what Nvidia had done and voiced my disaproval to it. And since the post was pointed at me I assumed it was said to me, is that a stupid assumption to make.....I don't think so.

<<<<<<(By the way Teasy, these two remarks aren't aimed at you as I don't really know where you stand on this. They are aimed at anybody in general who wants to whine up a storm about NVIDIA being "bad" while other companies are "good," based solely on what they read in the press and hear in forums. Readers will have to decide for themselves if it applies to them.)>>>>>>

They may not be aimed at me directly but I am someone who dislikes what Nvidia do in this area very much and I have voiced my disaproval quite a few times on message boards and my own webesite so I will reply to it. I certainly won't go as far as saying Nvidia are bad and ATI or IMGTEC are good. But I will say Nvidia are bad, until I see direct evidence of IMGTEC or ATI doing similar things then I won't tar them with the same brush. Its innocent until proven guilty and Nvidia have been proven guilty by there PDF which IMO was a discrace.

<<<<<<Have you had your head in the sand the last 20 years or more? I completely agree that this is the way it should be. But it is simply not this way at all. Look back: VHS vs. Beta; Windows vs. OS/2; DirectX vs. Glide; digital audio tapes; DOS vs. all of it's imitators and practically any other OS; K6-2 vs. K6-3; the Tucker '48... (OK so the Tucker was impractical to build at an affordable price, but its technology was still outstanding.)>>>>>

Just about every product mentioned there were two or more different formats trying to do the same thing and all totally incompatible with each other. VHS and Betamax were two different formats of tape recorders. Windows v OS/2 were to different operating systems. DirectX and Glide were 2 different API's. Tile based rendering is invisible to the developer with the Kyro II so optimising for a traditional doesn't hurt the Kyro II. So every example you quoted is nothing like tile based rendering against immediate mode rendering because you don't have to support one or the other.

<<<<<<I don't very often just come out and say somebody is wrong, but in this case I'll make an exception. Your last statement is in complete denial of historical fact. Politics ALWAYS affect technology, and not often for the better. Sorry man... it sucks, but it's true.>>>>>>>

Yes it does but the problem is you seem to have a misconception of tile based rendering as if it needs special support or something like VHS or Betamax did. As if optimising a game on a Geforce will hurt a Kyro II. So no I'm not wrong and its arrogant of you IMO to say that. Neither of us are totally wrong but you seem to simply have a misconception that the Kyro II's TBR needs support which is doesn't. In cases were two products are so different as to be incompatible one will inevitably lose badly and often it loses because of polotics, like VHS and Betamax. But the Kyro II is not different in that way. The architecture is different but the conpatability is the same and anything that will work well on a traditional will not work badely on a tile based renderer.
 
G

Guest

Guest
here is a swedish review...
kyro 2 against geforce 3 and geforce 2 pro
;)
<A HREF="http://www.sweclockers.com/recensioner/hercules_k2-3.shtml" target="_new">http://www.sweclockers.com/recensioner/hercules_k2-3.shtml</A>
you must take note that these are average benchmarking...
in mdk 2 WITH T&L enabled 32 bits 1600x1200 kyro 2 gets 39 fps against 45 on geforce pro. but the problem for geforce pro is that kyro 2 achieves always better minimum fps rate than traditional cards (ex geforce 2)
so it will be more playable with a kyro 2 in 1600x1200x32 with mdk2 with T&L than with a geforce 2 pro card !

geforce 3 is better of coursse ;)
but 500 us$ ???

this review was on a duron overclocked to 950 mhz if it were on a pentim 4 1,5 then we would see even better results for kyro 2 ...
kyro 2 scales better than any other card !!
<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 04/23/01 02:09 PM.</EM></FONT></P>
 
G

Guest

Guest
Quake3
in 16 bit the Kyro 2 only beat the GeForce MX. I thought you said Kyro 2 does FSAA without a Performance hit. It even failed fsaa 1280*1024.

In 32bit it only beat the pro by to frames on 1024*768. Again failed the FSAA in 1280*1024.


MDK2
16bit: Kyro2 loses against all but the GeForce mx
32bit: Same again.


Alice
16bit: same again!
32bit: in 1024*768 slower than all but mx. in the others its the same as the pro.


3DMark 2001
Even slower than the mx. in 1280x1024 in Game 1 the mx gets 4fps compared to the 14fps mx. In Game 2 and 3, it comes on par with the mx.


Also it cannot be overclocked much as it was first supposed to run at 166Mhz. but stm decided to sell it at the overclocked speed of 175Mhz.

And you think the Kyro 2 is competition for the High end GeForce 2 Cards? LOL!!! Its barely competition for the low end.



<font color=blue><b>Who is General Failure and why is he reading my hard disk?</b></font color=blue>
 
G

Guest

Guest
the only cards that they put in there to compare was:
geforce 2 pro
geforce 3
kyro 2
geforce 2 mx

only the geforce 2 mx is in the low budget !!
and we already know about that bug on directc8 (try to find here all the posts about directx8)


You should start using your head a little !!!

when we are on 16 bits and in lower depths EVERY CARD (even a tnt-1) can Achieve playable frames !!! my geforce 2 pro gets 300 fps ... duh !!! so what...
when it matters those geforces start to suffer really bad (higher depths 32 bits color, FSAA) and this LOW BUDGET CARD START GETTING GEFORCE 2 PRO SCORES !!!!
HOW ABOUT THAT ?

and that is a issue !
because when we start using future games (lots of overdraw)
these geforce 2 pro will start to suffer even more !!! UNLIKE KYRO 2
serious sam is an example and of course tribes 2 !!!

on the planettribes:

Being the resident tech guy here @ PlanetTribes, I pride myself with having the latest hardware in my hands. I recently received a review sample of the new Kyro II video card. What was the first thing I did? I popped it in the Win98SE testbed (1.1ghz Athlon TBird) and fired up Tribes 2. The results are phenominal! Out outperforms my GeForce 2 Pro by far. I managed to crank up T2 to 1280x1024x32 with EVERYTHING at full and did not see a dip below 45fps - in a nice 50-60 player game. My GF2 Pro can't hardly do that at 1024x768. I plan to write up something more detailed on it in a short while, I just wanted to inform you of what preliminary results I've pushed out of it. So far, it looks like a GREAT card.
"
then he put directx8 , we already know that there is a bug that prevent kyro to use all the power !!

"
Okie I formatted and everything on that machine, I found out why i had a dongle decrease in 32 bit - my own stoopidity... I ran the 32 bit with EAX on :D - and didnt figure it out, heh so I just formatted for nothing. at 1600x1200 the Kyro2 REALLY CRAPS OUT... 10-20fps... unplayable - seen it dip to like 5 too. But since I updated from DX7 to DX8 on that machine (during the format) it dropped from 45 FPS average outside in fights to like ~37-40... ugh...mmm... its about matching my GF2 Pro now... I paid $300 for my pro, I wish this little $150 Kyro 2 was around then... all in all its VERY playable @ 1280x1024x32. I'm looping it in Q3 for an hour or two just to make sure its stable. I'll post more exact numbers (I WISH T2 HAD A TIMEDEMO OPTION) later
"

take note on the drop because of the directx8 bug !!!

;)
you said that kyro doesn't overclocks well !
of course...these are not finish products they are only samples...

<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 04/23/01 05:36 PM.</EM></FONT></P>
 

HolyGrenade

Distinguished
Feb 8, 2001
3,359
0
20,780
Lets see some benchmarks with the kyro showing valid competition against the high end nVidia cards (Gf2 Pro + Ultra, GF3? yeah right!!!) before you start making these claims.


<font color=red>"My name is Ozymandias, King of Kings:
Look on my works, ye Mighty, and dispair!"</font color=red>
 

Ncogneto

Distinguished
Dec 31, 2007
2,355
53
19,870
Why should the Kyro beat those cards it is after all 1/2 to 1/3 the price. That is like expectiing an mx to beat an ultra, come on gimme a break. The fact that it even stays competitive in some test( and in some cases actualy winning) is a testament to the new technology ( or refinement of an old one). This card will be priced in the neighborhood of the radeon Le and gforcemx, and when put up against those cards it looks like a winner. When kyro-2's big brother comes out ( soon I hope) We may actually have a third contender in the video card market. Be you an Nvidia, Ati, or Kyro fan this is good news for all of us. Maybe it will cause nvidia to slash prices like Intel just did. Even the biggest of Intel fans should be patting AMD on the back for that one.

A little bit of knowledge is a dangerous thing!
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
Thanks for the link for the review which I found interesting. This pitted a Herc MX against the Herc Kyro2. On a more typical system which a person would own, a Duron 700mhz. Which is faster than a 800mhz celeron and faster to the many non coppermine pentium3 systems out there. The Kyro2 may scale well but the test shows on this review that it scales down pretty quick also. The MX kepted up and surpassed it in a number of tests except for FSAA where the superior FSAA of the Kyro2 showed. Sold as a budget card in a budget system it competes with the MX. With a higher power system it competes with a GF2, Radeon and sometimes the GF2 pro. The FSAA competes with Voodoo5 and GF3 in effectiveness. A very interesting card and I sure wish the Kyro2 was tested in a 1.33mhz T-bird system overclocked to the gills to see what this card can really do. Seems like the card is weak with weak cpu's and very strong with strong cpu's. Talking about confusing but also a blessing for those who will be updating there system in the near future knowing that their video card will act like a brand new performance video card. Still for a budget system or a relatively old system the Kyro2 doesn't seem to be the best choice when a MX which is going for half the price performs well against it and beating it in a number of areas. Also now in the U.S., a GF2 delivered to your door cost $131, GF2 Pro 32meg to your door $154 and then a GF2 Pro 64meg w/ 5ns ram $189 on your doorstep. Plus the reported driver problems as reported by AnAndTech and the DX8 problem could be a trying experience for someone until the corrections are made. Still the Kyro2 is not a new chip where IMGTEC had to program new drivers, it has been out for awhile as a slower Kyro. So by now I would expect the drivers to be relatively mature in which the reviews indicate that they are not. So some of my questions would be if ever a TBR video card will have good drivers with few problems. So far the record indicates otherwise. I am almost concluding unless you have a 1ghz system or greater then another video card especially with T&L would probably be better.

<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 04/23/01 11:18 PM.</EM></FONT></P>
 

noko

Distinguished
Jan 22, 2001
2,414
1
19,785
I am sure ATI is rethinking their strategy. Kyro2 looks like an outstanding OEM buy especially with the many Pentium4s with TNT2s and believe it or not RagePros!! That is if the price is low enough. Plus the Kyro2 is priced right in there with the Radeon lines, direct competition now but also Nvidia GF2 reg/pro cards have entered that price range as well. So I expect ATI to do something since the Radeon hasn't really change in the last 8 months. So maybe a Radeon SE will be introduced while the Radeon2 comes out early. Still the sluggish sells of computer goods I think also plays a part, look at what happen to the GF3 price lowering to less than $400, GF2 Ultra less than $300, GF2 Pros less than $200 and MXs less than $100. Hopefully a strong third contender will be present in the market while ATI responds promptly otherwise I think they will be hit hard.

<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 04/24/01 04:00 PM.</EM></FONT></P>
 
G

Guest

Guest
I have a friend that choose to buy a tnt-2 m64 and a athlon 1,2 with 256 mb pc133 because he wanted to save some money!
(he said that in the future he could buy a good video card if and only if he find that he really needs it, he don't play much... )
please 700 mhz will not be the typical cpu with kyro 2, I can buy an athlon thunderbird 1 ghz for 200 us$ here in portugal
;)
a tnt-2 m64 costs (here in portugal) roughly 60 us $

noko you are right!
Geforce's prices are getting very low so low that manufactures are losing money (creative wants to drop graphic business because of that)
kyro 2 may give them some freedom to lower the prices even more without losing money...
;)
drivers problem?
the only problem I can see is from directx8 (microsoft responsability)<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 04/24/01 01:18 AM.</EM></FONT></P>
 
G

Guest

Guest
anandtech have a very interesting article !!!

<A HREF="http://www.anandtech.com/showdoc.html?i=1461" target="_new">http://www.anandtech.com/showdoc.html?i=1461</A>
 
G

Guest

Guest
Teasy,

Here we go again: :eek:
<font color=blue> The post was directed at me after all. </font color=blue>
I looked back at my post and realized it WAS a reply to you. Actually this was not intended, which is why I didn't use your name at the top. The post was just tossed out to everybody to generate some discussion about political issues surrounding graphics technology. Sorry... I can see now how that was not clear. :redface:
<font color=blue> Really, lets go back to your original comment "What would be the point? If the cards do poorly with overdraw, then the applications will avoid overdraw whenever possible, so an overdraw benchmark would not prove anything meaningful"

If it was a rhetorical question then why actually put a question mark there?. Also you said a benchmark which is made to test how cards run with a certain level of overdraw would not prove anything meaningful, because any app would try to keep overdraw to a minimum. This is untrue...

...Please do NOT point me to a dictionary, I know what rhetorical means. But how am I supposed to know if something is a rhetorical question if you put a question mark there.
</font color=blue>

1. A rhetorical question is said for effect, but is still followed by a question mark, as it, like any other question, would be grammatically incorrect without one. (Look <A HREF="http://www.clearcf.uvic.ca/writersguide/Pages/RhetRhetQuestion.html" target="_new"> here</A> and notice the example questions.) You determine if a question is rhetorical by the tone of the statement, and it's the tone of my post that I think you got all mixed up. This is a fair enough mix up on your part, as it is hard to infer the intended tone from text alone (especially when I screw up and direct it at you like I was continuing our last argument. :wink: ) It was supposed to be a <i> friendly </i> discussion of how politics <i> might </i> affect the Kyro II, for good or for bad.
2. My comment, "What would be the point," etc., was not a statement of my personal beliefs. It was supposed to capture the general "political attitude" (sorry that was the best phrasing I could come up with) of the last few years, in an attempt to explain why there are currently few benchmarks that really stress HSR. The Radeon is out now and has HSR, and now Kyro II and GF3 are out with even better HSR, and even a game is out that has a lot of overdraw. The time is ripe for overdraw-stressing benchmarks and they are certainly meaningful <i> now </i>. But I don't think very many people saw this 6 months ago, and benchmarks (and even more so, games that we use as benchmarks) take a while to develop.
<font color=blue> ...you seem to simply have a misconception that the Kyro II's TBR needs support which is doesn't. </font color=blue>
This statement seems to mostly sum up the rest of your complaints. You seem to think that I don't properly understand TBR, that I think it will take a lot of support from developers and that it's almost incompatible with current hardware. Well this is not really my view. It is true that I didn't know a lot about TBR when this post started (397 messages ago!) but I have since learned a lot. I understand that it is transparent to the game (which I knew all along.) I also do not think that developer support will be a <i> huge </i> issue, but I do think it will effect things some.

Let me give an extreme example: <A HREF="http://www.fastgraphics.com/images/reviews/000037-06.gif" target="_new"> Treemark</A> and <A HREF="http://www.fastgraphics.com/images/reviews/000037-04.gif" target="_new"> Villagemark</A> on a Kyro II vs. a GeForce. (I know you already know what these look like but I linked them for the convenience of those who don't.) Click <A HREF="http://www.fastgraphics.com/reviews.php?id=37&page=2" target="_new"> here</A> to see the source of these benchmarks. The TBR of the Kyro is transparent to both benchmarks, but yet the scores vary like night and day on both cards. This is because each engine tries to draw things that the other card is weaker at. Again, this is an extreme example and no game engine is going to be so totally unusable on the opposite card. But let's say that a hugely popular future game, like Unreal 2, comes out tomorrow and its engine tends to draw stuff more like Villagemark does. (Not to the same level, but just more so than it draws like Treemark.) This would make the Kyro card a very popular card indeed, while limiting the future usefulness of the GeForce series when you consider how many games are likely to use the Unreal 2 engine. Change that around and say Unreal 2 renders a landscape more like Treemark. Then the reverse could happen with card sales.

I say all this only to illustrate how developers--even though they don't have to write in specific TBR support-–can affect the future usefulness of <i> any </i> card. How much is up for debate. I do think that how developers write their games is influenced by politics, some of which is started purposefully by the companies involved (ref. your favorite NVIDIA gripe. :wink: ) I provided this just as food for discussion-–not to say that the Kyro is going to be left out.
<font color=blue> Just about every product mentioned there were two or more different formats trying to do the same thing and all totally incompatible with each other... So no I'm not wrong and its arrogant of you IMO to say that. </font color=blue>
Let me quote your original statement that I was responding to: <i> "It's the performance of a technology that will gain it acceptance." </i> To me, this reads like you were saying technology is unaffected by politics, and is always judged by performance alone. This is what I meant was "wrong," as politics do affect technology in many ways. My examples, like Betamax and such, were to prove only one point: that politics do affect the acceptance of technology. I was not trying to say that the Kyro II was in the exact same situation.

Now perhaps I should have read your statement to mean something more like this: <i> "The technology of the Kyro II will win over the bad politics because of its superior performance." </i> Though I don't completely agree with this statement, I am certainly not arrogant enough to say you would be wrong in saying it, as it is a valid and defendable position.

If I did indeed misinterpret your statement above, then I apologize. I just couldn't see how you could say that technology always succeeds or fails based on performance alone–-we do not live in utopia! But I see now that you probably weren't saying that.

Last thing and it's about the NVIDIA flames: It's your right to flame them and it's my right to call it whining. :wink: It's nothing personal on my part. I also don't really care if you flame some company. My main gripe is this: when considering the ethics of modern corporations, I think, sadly, that "guilty until proven innocent" fits much better. I get really tired of seeing NVIDIA made out as the bad guy compared to other companies when those companies almost undoubtedly do the same things, or would if they were in the position to. The fact that the press has not caught them yet is <i> not, </i> in my book, proof enough that the other big players aren't just as bad.

By the way, I appreciate the fact that no matter how hard we have disagreed, you have always answered with facts and research, and not just shot off a bunch of baseless flames like many people do. :cool:

Cheers,
Warden