Kyro 2 - Part 2

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Hey, this is the 'Ghostbusters' of threads... something you can look at over and over again and never get tired of 😉

Cow with legs spread wide either dead or playing 'cello.
 
I want to start giving more freedom to my imagination

I know that they (another videologic division "metagence") are going to introduce DSP (processors) very soon, these DSP can be used in a vast of applications (allmoust everything, from digital radios to graphics). DSP probably means digital signal processors... Not sure 😉 )
<A HREF="http://www.metagence.com/" target="_new">http://www.metagence.com/</A>

These DSP can be put inside a graphic chip(we will have one chip in a board, no multiple chips solutions... ), because 3D is massively parallel it will help to have lots of dsp working in parallel,

these "Imagination" DSP will be programmable they could modify a feature set of a given board via drivers...

I think that kyro 3 will be a kyro 2 with more pipelines and with many DSP cores ...
So kyro 3 will have all the directx8 feature...
maybe even directx9 , directx 10 etc...
:)
but this is only my imagination working...

Damn I make to many error in english... Portuguese language is sometimes in reverse, yap gramatically it's different from others latin languages, very different from germanic languages.

<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 04/26/01 10:49 PM.</EM></FONT></P>
 
what ?

go here:
<A HREF="http://www.xbitlabs.com/news/" target="_new">http://www.xbitlabs.com/news/</A>


<font color=red> <i>"Although the new STM’s chip, KYRO II, can boast a really good price-to-performance ratio, the manufacturers don’t hurry to announce anything built on this chip. Alongside with Hercules, STM’s partner #1, there is currently only one more company, which is about to manufacture graphics cards based on KYRO II - InnoVision. The behavior of other companies is just puzzling. For example, PowerColor refused to launch KYRO II based products shortly after it had actually announced them. Creative was talked to support the new solution of STM, but now it denies intentions like that.
What’s the matter? As we have learned from a number of Taiwanese graphics cards manufacturers, NVIDIA still produces an indirect pressure to make the companies use only its chips. And the manufacturers have nothing to say against. NVIDIA offers a full line of graphics chips ranging from the Low-End to the High-End solutions. If the manufacturers risk to seek for another shipper, they are likely to find no alternative. GeForce2 MX series has some equivalents, while the elder GeForce2 and GeForce3 are unique. Hereat most graphics cards manufacturers have to indulge all NVIDIA’s whims. Hercules doesn’t depend much on NVIDIA. Firstly, this company doesn’t concentrate on graphics cards only, producing a wide range of computer hardware. Secondly, it occupies a big share of the European graphics card market and NVIDIA can’t quarrel with Hercules that easily. So, Hercules has a good chance to get good profits from KYRO II cards, since their manufacturing costs are lower than those of GeForce2 MX cards. Yet it doesn’t work for other manufacturers, that are humble to NVIDIA’s will. But the situation can be twisted by the brand new STM chip of the next generation, KYRO III. Presumably, compared with KYRO II, it is rumored to give a 150-200% performance gain. Should the new chip arrive in time, within the third quarter, STM will be armed with a complete line of products to tilt someone else in its favor. For the time being, NVIDIA enjoys its supremacy over the graphics chipsets sector."
</i> </font color=red>

Nvidia is indeed giving a big push on evolution!!
Preventing others to compete thus enabling the possiblity to sell crap for lots of $$$.
 
I thought of something, a lonely pixel is going to be shaded 6 times (6 texel layers) in a GF2 card. The pixel churns(?) along the GF2 pipeline begging to be colorized but is let down due to the limit of two textures per pass. The pixel says, "oh well I will wait in line again but first I make a pit stop in the frame buffer". While in the frame buffer the pixel is escorted again into the inner chambers and is dressed (setup) again to make the long path. The path of color, the pixel making line. Still only two more textures are added, "Oh I am right here, finish me now, I don't want to wait any longer", says the angry little pixel. Then a deep voice responds "Send him back into the frame buffer". Finally the last escort occurs, a pleased and mighty shinny pixel to be maybe the best ever to dazzle the old master, but wait???? Zoro the Z king comes along and slashes him to pieces. Poor little pixel, poor little pixel, (sob, sob, cry, cry). Feel sorry for all those mighty pixels that are never seen or heard from again.

A story by a saved lost pixel. :lol:
 
Allot of people think this thread (including part 1) is one of the best threads in a while here at Toms and there's allot of interesting info in this thread. The technology being discussed here is something that quite a few people didn't even know existed so its quite an interesting subject compared to most.

OzzieBloke, thanks thats very kind of you I'm glade my explanation was clear. I didn't know a thing about graphics cards a year ago (and I really mean not a thing, I was the sort of person that bought a card if it had more ram and thought that was all that mattered :smile: ) so I think I'm doing ok just from studying the Beyond3d forums and a few articles and reviews.

Powervr2:

<<<<These DSP can be put inside a graphic chip(we will have one chip in a board, no multiple chips solutions... ), because 3D is massively parallel it will help to have lots of dsp working in parallel>>>>

For a nice little article on this subject go here: http://pvr.gamestats.com/Dynamic/Standard.shtml?/articles/metagence/metaspec

Its interesting stuff. Apparently metagence (the name for the DSP tech) would effectively allow IMGTEC to design say two (or more) chips like Kyro II and say one (or more) HW T&L unit all into 1 chip which would be controlled by the DSP. The DSP would be completely programable allowing it to intelegently control all three (or more) cores at the same time varying the resources each core gets depending on the task at hand allowing for incredible efficiency (which I suppose is what PowerVR chips are all about really).

I'd also just like to add that Hercules have just put all there Prophet 4000XT(Kyro 1)/Prophet 4500(Kyro II) drivers on there public FTP. Including W98/ME, W2k, WNT4 so there obviously ready to release the cards very soon. Toms can also now change his review so it no longer says there's no Win2k or NT drivers for the prophet 4500 too.
 
Yep all that hard work that little pixel went through to look its best only to be discarded, so sad. It must feel so used :frown:

Those are the horrors of post rendering HSR and clumbsy texturing caused by traditional immediate mode rendering that so many poor little pixels have to suffer every day. Please....give £1 or whatever you can, to The Home For Discarded Pixels, PO BOX 84959. Just £1 or $1.4 can feed a family of 786432 32bit (a average 1024x768x32 pixel family) pixels for an entire week. Just £85 can buy a Kyro II utilising intelligent deferred tile based rendering which would saves so many future pixels from this terrible indignified existence. Please don't let these tiny lost souls be on your conscience......bring some happiness into a life today.
 
thanks for the <A HREF="http://pvr.gamestats.com/Dynamic/Standard.shtml?/articles/metagence/metaspec" target="_new"> Link </A>teasy...
:)

<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 04/27/01 02:29 AM.</EM></FONT></P>
 
<i><<go here:
<A HREF="http://www.xbitlabs.com/news/" target="_new">http://www.xbitlabs.com/news/</A>>></i>

I couldn't resist... <font color=red>"Hercules doesn’t depend much on NVIDIA. Firstly, this company doesn’t concentrate on graphics cards only, producing a wide range of computer hardware."</font color=red>
LMAO! Creative is a Giant compared to Hercules! They also don't heavily depend on NVIDIA to exist. I can build half a PC with Creative products alone.

<font color=red><<...For example, PowerColor refused to launch KYRO II based products shortly after it had actually announced them.>></font color=red>
I too think Nvidia influenced PowerColor to not produce video cards based on the Kyro II. It's all about money in PowerColor's eye right now. I guess Nvidia made some kind of offer that they couldn't refuse.

=
<font color=green>Are you just lazy or incredibly stupid? -<i>King of the Earth</i></font color=green>
 
On the subject of how long before games start to depend on T&L. Developers are now working on X-Box games and to get the most out of this console they will make the most of T&L. No doubt, many of the same games will be ported to PC and will only look as good with GeForce 3 equivalent graphics cards. The problem to date is that not enough people have had T&L hardware to drive the developers. X-box will change this, and probably quite rapidly.

Off topic I know (but related), I wonder how X-Box will affect AMD, if developers start spending more time optimising for SSE to get best X-Box performance?
 
I'm pretty sure (not certain but this is what I hear from allot of people) X-Box doesn't have a DX7 HW T&L unit like the Geforce 3 but instead it has 2 DX8 vertex shaders (DX8 vertex shaders are more flexible then DX7 HW T&L but slower, but having 2 DX8 vertex shaders allows for programability at the same speed as hardwired DX7 HW T&L). So yes X-Box games will be written with HW T&L in mind and HW T&L will catch on faster because of the X-Box. But it'll be DX8 programable HW T&L that catches on because of the X-Box not DX7 HW T&L like the Geforce 2 and Radeon have.

<P ID="edit"><FONT SIZE=-1><EM>Edited by Teasy on 04/27/01 06:48 AM.</EM></FONT></P>
 
PowerColor will in fact be releasing a KYRO II board. Well the company will be releasing the board, but NVIDIA has FORCED THEM to release the board using a differnt trade name other than PowerColor. What that name will be is anybody's guess at the moment. It is really sad though to see NVIDIA bully OEM's about releasing products from other companies that compete with their products. It is funny how this compnay has changed in the last couple of years. I think they have been visiting Bill Gates too much lately.
 
That was good, I almost died laughing :lol: . I wished other people had more humor around here. It appears that nVidia and probably ATI has more to fear about the Kyro2 or more exactly the refined technology of TBR then being let onto. First TBR allows for much cheaper cards with great performance. I can see the KyroII becoming very attractive for card builders due to better profit margins not to mention of OEMs. Second for nVidia and ATI to ditch their hardware designs and starting from scratch voiding all their driver work would set them back years. Yet those inefficient hardware designs are driving the video card prices sky high to begin with.

<b>These are my best guesses or predictions:</b>

ATI, introduction of the Radeon SE which will show faster benchmarks in most benchmarks except for FSAA over the Kyro2. Nvidia already has two cards that shows better performance on most benchmarks. What this shows is that the Kyro2 is not the best performing card in the publics eyes which means the companies that cater to the public as in OEMs and card manufacturers will be pressured into using ATI/Nvidia cards due to a general public perception. Even though what the OEMs will be installing are stripped down versions of Radeons or GTS MXs that don't even perform as well as the Kyro2, the general perception of ATI or nVidia being better does have influence. I think what was mentioned earlier is politics does play a part. Just think, how many GF2 Ultras where sold? Very little compared to the rest of nVidia lines yet the perception that nVidia is better helped to sell many MX's and TNT2's (to OEMs).

Second, rapid price reductions, mostly to get rid of merchandise in stock but also to blur any new designs out and to slow down the introduction rate meaning less of a fan base and voice in the market. Just think of 10 million Kyro2 owners praising and recommending a Kyro2 not to mentionthe KyroHolics that will spring from this. That in itself establishes a brand name and promotes many other people in buying a Kyro2. Resulting in loss sells to the other folks making video chips(a.k.a. ATI, nVidia). Look at the so called ATIist and nVidiots that are on most computer hardware boards, regardless of facts they just promotes one company worshipping whatever the company produces.

Lastly a earlier introduction of the RadeonII and nVidia upgrade to the GF3 I think will result.

The kyro was contained pretty much by this method and origianlly the Kyro had driver problems which looks solved now. So Imagination does have some work cut out for them but if they perservere and introduce a killer Kyro the technology I think will catch on big. ATI may have some stuff up their sleaves, for one thing the Radeon drivers can do tiles with textures, buffers, AGP textures with the onboard ram. Now moving some of that on the chip and you will see something similar to the TBR of the Kyro.

<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 04/27/01 09:58 AM.</EM></FONT></P>
 
Here is a review of a RadeonSE:
<A HREF="http://www.pcpop.com/info/20010426/" target="_new">http://www.pcpop.com/info/20010426/</A>
It is in Chinese so unless you read Chinese looking at the test graphs and PowerStrip data gives you an idea of whats coming up. Note: it can overclock to 250mhz from its normal 230mhz speed.

<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 04/27/01 10:56 AM.</EM></FONT></P>
 
I see a whole lot of posts on various message boards with people saying they love the idea of the Kyro II technology and a small lesser known company coming out with such a nice card, and stuff like "did anyone see this card coming?, its awesome I'll be getting one as soon as their released". IMO there will be quite high sales for Kyro II in the retail market and it will get a good unser base which will help the next PowerVR card. I don't really think that the reason the Kyro 1 didn't do that well was because of politics from Nvidia or ATI. Nvidia AFAIK really didn't bother with the Kyro 1 and didn't see it as a threat. For proof of this look at the Powercolor Evil Kyro. If Nvidia had seen the Kyro 1 as a threat they'd have done the same thing as there doing now with the Kyro II by persuading Powercolor not to use there name for the card and producing scanderlous PDF files and bullying board makers like INNO3d not to make the card, which AFAIK they didn't do with the Kyro 1. The reason the Kyro 1 didn't do that well was it didn't have performance that would reach out from a review and smack you in the face like the Kyro II has. It was a card that started off slower then an MX in low res and slightly faster at high res and 32bit because of unoptimised drivers. Then after most of the reviews had already been done loads of driver improvments were made and now its quite a bit faster then the MX but not many people saw this. Also it cost the same to preduce as the Kyro II so it wasn't cheaper then the MX when first released (£115 here in the U.K at release for a plain 32mb Kyro 1 compared to £85 at release for the Kyro II 32mb TV-Out). Now the Kyro II comes out for a budget price and convincingly sweeps the MX aside by as much as 70%+ at high res and challanges the GTS in 16bit and beats it in 32bit and beats the Radeon DDR too and also has great FSAA. Its a very impressive card and really leaps out at you from a benchmark. So the Kyro II will be allot more succesful then the Kyro 1. All its publicity about its great performance from sites like Anand has already seen to that. You should have seen the Anand forum after that first Kyro II preview. It was in a state of shock and people couldn't beleive what this card, that most of them had never heard of until that preview, could do. Allot of those people will be buying a Kyro II. I agree though the Kyro II won't be some massive hit but it will build on the minor success (compared to the Neon250) of the Kyro 1 card and make a user base and build awarness ready for NP2 and NP3 (Kyro 1 and 2 were NP1, that was there code name).

Also the Radeon SE probably won't be much under $300 ($250 online) at release if its released at all (at the moment its still up in the air whether ATI will release it or not), it certainy won't be in the Kyro II's price range.
 
The RadeonSE as far as I see it is a image card. Meaning something that performs well but probably won't sell that many. Kinda like the GF2 Ultra as a status symbol for a company. Anyways we will soon see how happy people are or not with the Kyro2. When it will be placed in 10s of thousands of machines and conflicts start to arise. Hopefully not. The RadeonSE may still be canned and the Radeon2 will shortly come out instead. Still the price is the driving factor in this and the Kyro2 has that potential advantage. The Kyro2 is only a chip for a video card which does have limitations like all the other video chips produced. Plus it is competing with the GF2s, RadeonDDRs to start off with. ATI does claim games around christmas time to take advantage of the 3rd texture unit of the Radeon. We will see by then how all of this stacks up.

Plus do you really believe the DX8 bug on making the cpu write to textures on the Kyro2 will dramatically increase the Kyro2 benchmarks when fixed? Greater than 30%? Or more likely a 10% increase. Which would mean that it would continue to score the lowest in 3dMark2001 benchmark a benchmark possibly indicating future game potential. A hardware T&L optimize benchmark.

Plus the thousands and thousands of hardware configurations and different degrees of Operating system clutter and updates all that could cause significant KyroII problems. This could get pretty ugly. So we will all be eager to see how effective Imagination updates the drivers.

Allot of games havn't been tested on the Kyro2 but shortly they will be. What problems are lurking around the corner for the thousands and thousands of games available today?

<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 04/27/01 01:13 PM.</EM></FONT></P>
 
Well in Giants, were rendering into textures is used for advanced water effects and shadows (amongs other things), the framerate drops by 10fps when those 2 features are enabled with DX8. I haven't got anyway of knowing exactly what increase the Kyro II will get in all games because different games use different features but what I do know is 3dMark2001 uses rendering into textures in low detail tests and then adds dynamic shadows (again more rendering into textures) for the high detail tests so I can see a high increase in speed when the bugs fixed. Also lets face it the Kyro II is closer to the GTS in benchmarks seen so far then the Radeon DDR. Any speed boost will put it close to the GTS in 16bit (which is were its still a little slower then the GTS, in 32bit its faster).

Serious Sam already uses all 3 TMU's on the Radeon cards. It fully supports the full 3 TMU's per pipe on the Radeon and it still doesn't impress in that game falling behind the GTS and well behind the Kyro II. Don't get me wrong I like the Radeon and I have a Radeon myself (its not in my system at the moment though) but its not quite upto Kyro II speeds in most benches out so far.

On 3dmark2001 ,when any game is anything like a 3dmark series test I'll then beleive that non-hw T&L cards are in trouble. I've said it before, the high detail tests in 3dmark2000 are still much much slower on my PC then any game I've ever played (and I play all the new big games). And why? because 3dmark2000 is supposed to run like crap thats why. Think about it, does 3dmark2000's high detail helicopter test look anywhere near as good as most games out now?.....it looks rubbish compared to Giants so that should show you the difference between a benchmark and a game. A game developer wants to make their game look great and run great to sell copies of the game, and a benchmark developer wants there bench to look great and run like crap on anything but the best hardware. The engine of Max Payne might be in that test but its not Max Payne. Its completely made by madonion and its in no way an indication of the real game performance of Max Payne. For a start it deliberatly uses an incredibly wastful reflective floor that doubles the poly count in the high detail test and also doubles the fillrate and bandwidth hit that the low detail test takes, appart from the reflective floor nothing much is added to the high detail test over the low detail one. Do you think Remedy will make a game that uses reflective floors all over the place which makes the speed of the game slow down by half?, if they do I'll turn the reflective floor off and double the framerate and so will most others because the reflective floor simply isn't worth the trouble.

<<<<<Allot of games havn't been tested on the Kyro2 but shortly they will be. What problems are lucking around the corner for the thousands and thousands of games available today?>>>>>

I think your really underestimating the amount of people that have Kyro cards and report problems to all the PowerVR forums. I myself have tested close to 60 games on this card, with 3 problems and those problems were only when TC was forced (one was the Giants TC problem which is now fixed and I haven't tried Rune or Dues Ex to see if there fixed yet). Over the last months I've had 4 new drivers from IMGTEC. They don't release all of these to the public like Nvidia but they wait until a decent amount of problems are reported as fixed by webmasters like me and PVR-REV and Paraknowya etc and then they release a much improved driver to the public. If I don't find a problem with the drivers then there good drivers because I don't just play standard PC games. I play PSX emulators, Arcade emulators, N64 emulators, and allsorts of stuff and Kyro does fine with all of them, this shows a general maturity in the drivers. When I got my Radeon I installed it (with latest drivers at that time) and I had allot more problems with it then I've had with the Kyro. In conclusion IMO the Kyro II drivers I'm currently using in my Kyro 1 (which are now publicly available) are very stable and much more stable then my Radeon. I don't have a problem with any games I play, I download allot of warez so I play allot of games. The only problem I have is with DX8 thats it. The 2d stability is great too. Anyway as you say we'll all see the quality of the drivers soon when Kyro II is released, but I'll say with absolute certainty that there will be no big problem with these drivers when more people start getting the Kyro II.

Something that lots of those PCI only mobo owners out there might be happy to hear is a Herc 4500 PCI will be released along with the AGP version, so all those people stuck with only the option of a PCI MX now have a better faster option.

Check it out here: http://www.acidhardware.com/reviews/3dprophet4500/index9.shtml
 
<font color=red>"Plus do you really believe the DX8 bug on making the cpu write to textures on the Kyro2 will dramatically increase the Kyro2 benchmarks when fixed? Greater than 30%? Or more likely a 10% increase"</font color=red>

Hum... I think that the increase in performance will be greater on lower end cpus ...
a greater boost on a duron 700 than a athlon 1,2 ghz...
 
I found this interview with a man of nvidia here is the link
<A HREF="http://www.hwzone.it/html/text.php?id=205" target="_new">http://www.hwzone.it/html/text.php?id=205</A>


<font color=red> [Marco] What about ATi and Kyro boards? Do you see them as challengers for the GeForce 3?

[Dan] Not at all! We think that even a GeForce 2 GTS can do better than those product, so GeForce 3 is really way ahead for them </font color=red>

yah right if you play at 16 bits then geforce 2 gts is a little better but not at 32 bits or higher resolutions or even FSAA, if geforce 2 gts will do better than kyro 2 in the next months (in sales) then being evil pays off...


<font color=red>. Though, we see them asOne may view them as competitors for our MX range of boards but we feel that we have the best solutions at all price points. Especially with our new line of faster MX products...</font color=red>

your new lines of faster Mx (faster???? lol)

<font color=red>[Marco] The present prices for a top-level 3d card are much higher when compared to those from some years ago. How will your price politics change, now that there are not many competitors left, at the same level of NVIDIA?

[Dan] The high end has been pretty consistently priced in the $399 to $499 range for the last two years. Price politics won't change much, mainly becouse the biggest part of the price you have to pay for a cutting edge 3D card is due to the fast DDR memory used. Anyway, we'll be able to cover all the price segments, from workstation (top high end) to desktop. Desktop segment is then divided in performance (GeForce 3) and value (GeForce 2 GTS and MX) products.</font color=red>

:) or because you only know the brute force way !!!
nvidia don't (or want) make smarter design cards
they want to take our money from us!
 
hey what a xx ?
comparing a geforce 3 with kyro 2 ???
lol
comparing a card that is 1/4 of the price of geforce 3 ?

It must be better, it should be better than kyro 2 !!!

<A HREF="http://www.aceshardware.com/Spades/read.php?article_id=25000223" target="_new">http://www.aceshardware.com/Spades/read.php?article_id=25000223</A>

<P ID="edit"><FONT SIZE=-1><EM>Edited by powervr2 on 04/27/01 05:07 PM.</EM></FONT></P>
 
Well I hope you are right and the Kyro2 makes a very big splash with great success. That the DX8 issue is resolved by Microsoft promptly. Since it is a DX8 issue then comparing the differrent cards on OpenGL games would probably show the real potential of the Kyro2. In which case Serious Sam (A OpenGL game) also QuakeIII shows outstanding performance of the Kyro2.

Looks like after 9 months after the Radeon was official lauched 11 months from time of annoucement the Kyro2 caught up to it in some benchmarks and games, except for the video stuff. The T&L issue will also be seen in the coming months when new games like Max Payne are officially launced and benchmarks are conducted that is if it is different from the 3dMark2001 benchmark that used the same engine. In the mean time I will be enjoying my Radeon in all my games at max settings.

Ace Hardware is coming up shortly with a Video Card guide including the Kyro2 and a overclocked Radeon 225mhz representing what the Radeon SE can do if it is launced. Frankly I think ATI should just concentrate on the Radeon2 and start making annoucements so as to not loose potential buyers. I am sure ATI will respond to the new competition. Nvidia is not standing still either, look at the dramatic price reductions goin on. Now the question is when will the Radeon2 be delivered and in what variations. Comparing the Radeon2 and GF3 against the Kyro2 will make the Kyro2 look like a budget and a rather low end card. Still all the new cards available now will play the new games just fine for the next 6-12 months.

The current video cards surpass what the developers are delivering. Taking 18 months to 32 months to develope the new high end games while the video cards are being significantly updated at 6 months intervals caused a significant gap in cards ability and usage.


Do you know if a mobile Kyro chip solution is in the makings? Seems like an ideal chip for a mobile unit due to its low overhead.

Are there any official roadmaps for the Kyro line of graphic chips? Not the unofficial versions on the web sites.
 
Prices are coming down quick so the GF3 is not 4 times the cost of a Kyro2 at least in the States. When the Radeon2 comes out it should compete price wise with the GF3s.
 
Looks like STmicro is getting all they can from the Kyro2, this is fresh from Ace's Hardware which today will be publishing on the web a Video Card Guide
<i>
Until then, a few words about the "overclockability" of the Kyro II. It is true that the review samples come with 5 ns memory (you can see "-50" on the chips) and that retail samples will probably ship with 5.5 ns. <b><font color=purple>However, I can honestly say that the memory chips are not limiting the overclockablity. The core does not get hot, but even with improved cooling (a giantic fan blowing air on both the chip and memory chips) we could not get it past 185 MHz. IMHO, the Kyro II is limited by its architecture and that might also explain why the Kyro I was such a bad overclocker (1-2% overclockable).</font color=purple></b></i>
Looks like what you get will be a slower ram version Kyro2 when it hits the street with no chance of overclockability. My Radeon easiliy goes to 200mhz/200mhz stably which does indeed improve the benchmarks. With some ram heat sinks I am sure I will be able to go to 205-210mhz which would smoke the Kyro2 in most benchmarks except for FSAAs. With the core of the Kyro2 being pushed to its maximum speed I see a real potential of high rate of failure rate of the Kyro2 boards. So now I think the Kyro2s that where given for reviews where hand chosen and may not represent what us buyers would be paying for. That wouldn't be the first time this occurs. Well we will see how a Kyro2 stacks up against a overclocked Radeon at Ace's Hardware a video card that is 9 months old competing against STMicroelectronics best still to be released chip which is overclocked to its max speed.
 
these were sample cards !!!
we must wait for some final board to conclude..
but even with final boards there will be some boards better for overclocking than others...
some would get 190 some don't etc...

the 5.5 ns can easily go to 190 mhz, I think...
Maybe I am wrong enlight me...
 
That was a good review of the GF3 and Kyro2 which included other cards as well. Giants at 1024x768x32 had some interesting results🙁GamePlay)
<A HREF="http://www.aceshardware.com/Spades/read.php?article_id=25000229" target="_new">http://www.aceshardware.com/Spades/read.php?article_id=25000229</A>

. . . GF3 - 38FPS

. . . Radeon - 37FPS just 3% slower then a GF3! :lol:

. . . GTS 2 - 29FPS Radeon 28% faster :smile:

. . . Kyro 2 - 26FPS <b>Radeon 42% faster</b> (oops the Kyro2 couldn't beat this 9 month old card here :frown: )

Even at 16bit mode the Radeon was able to beat the GTS 2 card besides the Kyro2 card. Once again showing everybody the card you buy should depend on what you do and sometimes the games you play.

Now Serious Sam benchmark shows a very awesome result with the Kyro2 and my hat is off to the Hercules 3d Prophet 4500. Only thing faster is a GF3 period. Excellent!!!

Now with QuakeIII 800x600x32 FSAA4x for QuakeIII the GTS 2 beats the Kyro2 barely while the Kyro2 beats the Radeon hands down. In Nascar racing even the Radeon beats the Kyro2 FSAA4x at 800x600x16??

Formula 1 Grand Prix tests using different methods of filtering shows a rather steep degradation from bilinear filtering to Anisotropic from 120FPS to ~21FPS! 😱 While the Radeon maintain a rather decent consistent 70FPS :smile: . I not only use 16x but also 128x Anisotropic filtering on my Radeon which is just plain awesome for improving visual quality. Even a GeForce MX beats the Kyro2 easilily when Anisotropic filtering is used. What gives?
<A HREF="http://www.aceshardware.com/Spades/read.php?article_id=25000232" target="_new">http://www.aceshardware.com/Spades/read.php?article_id=25000232</A>

I really don't see the Kyro2 clearly beating the Radeon or GTS 2. Where advance games are soon to be hitting the shelves the GTS and Radeon will look even better.

<P ID="edit"><FONT SIZE=-1><EM>Edited by noko on 04/27/01 10:24 PM.</EM></FONT></P>