S3 & PowerVR

I don't fully agree. The MBX is targeted at the PDA-market, but the Speed is so high that it would give most laptop-chips a run for the money, except maybe for the 2D (cause of the integrated 2D-3D thing ).

Remember, the MBX can run at up to 200MHz; so the fillrate would be up to 400MPixel.

Manfred

 

Seems a bit extravagant for the MBX to have two pixel pipes, afterall mobile phones aren't going to have a screen resolution of much more than 320x240. Dreamcast coped well with 4x that at 60Hz and with a single pipeline. Though, I suppose with 2 pipes the core can run at half the frequency to achieve the same output - and, lower clock frequency + lower voltage = lower power consumption.

 

ronaldo, maybe it's because they are doing some form of FSAA on 320*240? Also it would allow for more ops per pixel => better quality... just guesses on my part though.

Serge

 

I'm not actually sure how many pipes MBX has myself, but another reason why it may have two is to make sure it still does fast 2D - seeing as the 2D is being drawn through the 3D pipes.

 

I'm not sure myself, but DaveB said so :

Quote :

 

Concerning why IMG hasn't had aggressive partners, and Teasy's eloquent explanation of why IMG may need partners: if licensee X storms the market with PVR based products, not only X but also IMG gets rich. IMG may then consider itself wealthy enough not to need licensees, and starts producing its own chips (via a foundry, as do nVidia and ATi), leaving X in the lurch. Alternatively, if PVR products aren't capable of storming the market, there'll be limited interest anyway.

Perhaps, for this reason, potential licensees do not consider PVR worthwhile as a long term investment? Just speculating.

 

Yeah, I guess fast FSAA could be the reason, or as Videologic call it "FSAA for FREE(TM)". It's just that I assumed their 'free' FSAA meant some new tech that would realy give no/low performance drop FSAA, not just the inherent bandwidth saving that comes with tilers.

Oh well, Cebit is just around the corner so all will be revealed... ...or maybe not

 

There is now an additional interesting development. Creative Labs is obviously looking to get into the graphics market in an "aggressive way", based on their announced acquisition of 3D Labs.

This begs, the question....why didn't they buy STM Graphics instead?

 

No significant IP involved, it would still be owned by IMG.

 

Quite possibly, yes.

In other words, IMG's business model itself may be a singificant reason why we don't see PowerVR products in the "high end."

Companies may be "afraid" to persue such markets if they don't have more ownership of the tech than IMG allows.

So, we have 3rd parties with the cash unwilling to take the risk of a "joint ownership" of the technology venture, and we have IMG / VideoLogic without the cash unwilling to make the investments needed.

End result: as I've been saying, lower probability of ever seeing PowerVR products in the high-end space.

 

Your evidence as I've said is overly simplistic in the extreme. It assumes that there is such a thing as a sure thing in the current graphics card market. There isn't.. the highend graphics market is an incredibly aggressive risky business (especially with Nvidia's product cycles) no matter what tech you have and some just don't wish to take that risk. 3DFX were a massive company and had great tech, were are they now?.. there a cusualty of an incredibly competitive and expensive market that will bankrupt even the biggest company if consistent success is not achieved.

In the end you have no technical data to support your point that PowerVR cannot develop a high end card.. have you?, your simply guessing that they can't.

It doesn't amount to not feeling like it.. I've never said that so don't put words into my mouth. You seem to think that business is black and white and you either have a guarenteed success or a failure, its pretty obvious this isn't the case.

Would you like to explain exactly how I have an extreme isolationist attitude toward the graphics card market? Maybe you should get your facts straight before assuming that. I've made it farely clear to everyone that for the last couple of months I've been using a Radeon 8500. I've previously owned a Geforce DDR, Radeon DDR, Voodoo5, TNT2, Voodoo2, Voodoo1, Rage128, S3 Virge ect ect.. what an extreme isolationist attitude I have to the graphics card market eh?

Besides my actual point is you might be interested but your not the sort of person that keeps right upto date with the latest PowerVR info are you?

There are many reasons, and you are only picking one presumably because its what you want to believe. Now I realise you will throw that point right back at me but the difference is I hear all the time from extremely reliable sources that will remain nameless that STM are not willing to risk money on going into the high end market and that is what holds back the current tech. For instance STM have still not used HW T&L tech in their PowerVR based graphics cards. Why is that?., is it because PowerVR can't design HW T&L units?.. nope, because they designed the very powerful Elan HW T&L unit used in Naomi 2 4 years ago. So if IMGTEC can do a hardware T&L engine and yet STM have never used a HW T&L engine in any PowerVR graphics cards then obviously that is not down to a lack of skill on IMGTEC's part but to a lack of ambition on STM's part. After all, if Kyro II had been released with HW T&L it would have been quite a highend card in performance at that time.

If neither of us had any other info or preferences toward this subject then the fact that IMGTEC and STM have never entered the high end market despite having potentially excellent tech could mean many things. It could mean that IMGTEC can't come up with high end designs, as you choose to believe. It could mean that STM don't believe there engineers are talented enough to produce chips on those designs. Or it could mean that STM either don't think that themselves or IMGTEC can produce a chip that's guarenteed to beat Nvidia (no chip would be guarenteed that) and so they will not risk the massive amount of cash neccesary to enter the highend market. Why do you think that one of those possibilities is true rather then one of the other possibilities?

I know why NEC never did, because they went for the easy guarenteed money, they had a graphics card that could hang in the high end in the Neon 250 and instead they decided that a deal to put the chip in a Sega console along with Sega arcade machines was the more lucrative deal, and they were correct. They would have never sold around 15 million chips in the desktop market. STM.. who knows for sure?.. I don't.. what I hear from reliable sources is simply that STM are not ambitious in the graphics card market, its not there speciality and there not willing to risk big money in focusing on that market in the long term, exactly what is behind that attitude is something I don't know for sure.

No no'one can assume that a company going for the highend with PowerVR tech will mean a guarenteed success, and that has been my point, problems can always arise. However there is no doubt that given some talented engineers, allot of cash and a little bit of luck a high end tiler would beat the crap out of anything on the market right now (Radeon 8500 and Geforce 4). This is born out simply by looking at the current PowerVR tech, Kyro II has the memory bandwidth and fillrate of a TNT2, yet it beats the crap out of a Geforce 2 MX and is right up there with a GTS.

Are you suggesting that I'm lying when I say that info is publically available? As for the rest well ok then, it has less transistors then Radeon mobility 7500 so it WILL be cooler and use less power, does that make you happy?

Lets wait until STM announce the sale of their graphics division, if the company is an ambitious company that is focused on the desktop PC market, like say VIA or Intel and wants to take on the PC desktop graphics market aggresively and we still see no higher end PowerVR tech then you'll have a good case.

Because someone else got there first.

 

Obviously. That's the point.

The point is, we have yet to see anone actually take that RISK with powerVR technology with respect to entering the high-end market. Again...WHY.

I want to hear some of your personal speculation as to WHY no one is taking that risk. Instead of me offering my opinion as to why no one is taking the risk (tech may not be as good as some make it out to be, their business model is not suited to the PC market), I want to hear YOUR OWN speculation for it.

In other words, instead of you trying to refute MY opinions as to why we don't see such ambition and high-level products...offer YOUR OWN. You have not said anything yet with this respect. And that's what I want to know.

In the end, there is no high-end powerVR PC product. I am simply stating this FACT, and providing speculation as to why we have the reality of no high-end PowerVR chips. In other words, I am not GUESSING that there are no high-end PowerVR PC products. That is a fact. And based on that fact, I am indeed speculating that it MAY not simply be a matter of someone "wanting to build one" for such a product to exist.

Understand the difference?

I won't bother going back to the old B3D forums (yet), but in this very thread, you said: "STM have said their not interesting in the highend market, IMGTEC have also said it is STM's decision not to go for the highend (I've also heard from reliable sources that this is the case many times). "

Then PLEASE, use your own words. Offer your own speculation as to WHY no one has produced a high-end PowerVR PC part. WHY is STM "not interested" in the highend market.

And simply telling me that there are "high risks" involved is not good enough. Because you (and I) don't know if they are lack interest in the high-end market, full stop, or if they're not interested in persuing the high-end market specifically with PowerVR technology because they don't think they'd be successful with it.

By virture of your arguments in this and other threads. I have yet to see you cast blame on IMG for, well, anything. Every time someone points a finger at IMG, you shift the "blame" away from them, and onto their partners, etc. Also, your willingness to bet parts of your personal anatomny on the execution of IMG and it's partners, doesn't help that image.

And what difference does that make? You would think that someone with the "latest PowerVR info" such as yourself wouldn't be in danger of eating his private parts in a lost bet. How could someone with such "inside info" as yourself possible lose a bet that concerns PowerVR, against someone like me that doesn't "keep up to date" as you say?

Maybe it's because I apply a bit more common sense, or maybe it's because I don't take every rumor I hear PowerVR, and spin it in the most positive and favorable light.

No, it's because I have not been presented with any other viable alternatives.

Great...you've just possibly outlined a weakness in PowerVR tech. It's designs may be "complex" for the typical engineer. But doesn't this go against the "fewer transistors are needed" theory? Shouldn't PowerVR's designs be EASIER to manufacture compared to comptitive designs, becuase they use fewer transitors?

WHY? You keep offering this "reason", but it has no substance. In other words, this reason can very well be used to back MY OPINION, becuase you have not told us WHY you think that reason could be valid. In other words, it could mean that the PowerVR tech does not offer ENOUGH advantage over competitor's tech, to give any reasonable assurance of success in the market.

Again, the point is, BOTH of those possibilities can be used to back MY OWN opinion. Again, you are not asking yourself the underlying reason of WHY STM or any other partner doesn't think they could COMPETE (or beat?) nVidia with a PowerVR design. All you keep saying is that these partners may think that the competition is too tough, so they "choose" not to compete. What you haven't asked yourself, is WHY don't they think they could compete with a design that suppossedly offers significant price / performance advantages?

Guranteed money? No such thing. You mean the shrinking volume of new Dreamcast hardware and now NEC has no high-end PC products? Some guarantee. You mean guaranteed short-term money. So the question is, why didn't they still keep a license for making PC products beyond Neon / Kyro?

And you wonder where I get the "they simply don't feel like it" thing from, when you make statements like that?

Then why don't you SPECULATE. That's what I'm after. There's only one reason why STM would not be "ambitious" about the high-end market...because they don't feel like they can make a lot of money in it. And PART of the equation, must be a consideration of the technology they would use to go after that market. Yes, there are other considerations too.

Uh, no. I'm suggesting exactly what I said: "A lot of "shoulds", "coulds", "not exacts", "someone elses" in there. Not exactly convincing. " Translation: I'm not conviced by your statements with no quoted references or evidence.

No, becuase just because something has fewer tansistors, doesn't mean that it will be cooler and require less power. Does the Kyro core have any "power saving" tech to reduce the power used when certain parts aren't active, for example?

I doubt it you see it that way. You've already given yourself an out. You said "and wants to take on the PC desktop graphics market aggressively."

So all you'll be saying in the future is that "well, they bought STM but "simply decided not to be aggressive." And we'll be having the same argument. I'll keep on asking you WHY they "simply decided" not to be aggressive, and you'll keep on saying it has nothing to do with IMG or the tech.

Huh? Where did we read about someone else buying STM "first?" There have been no announcmenets of an STM buy. Seems obvious to me that Creative Labs had their CHOICE of who to "buy" to enter the high-end mainstream market, and they chose 3D Labs.

So again, why not STM?

 

STM are not a business focused on the PC graphics market.. Their graphics division is not much of a priority in their business (which is obvious considering recent events) and they probably feel that considering the extremely competetive nature of the highend part of the desktop graphics market they don't want to shift priority away from other sectors of there business into their graphics division. In the end whatever is behind STM's decision not to enter the high end graphics market doesn't change the fact that they have shown a lack of ambition in the graphics card market and that is the comment you seem to be arguing with is it not?

I didn't say you were guessing that there was no highend PowerVR PC products I said you were guessing that PowerVR could not make a highend PowerVR design.

How should I know exactly why there not interesting in going into the highend PC market? I don't work for STM, the most I ever get as an answer to why there are no high end PowerVR cards is that STM have not choosing to go into that market and they have not choosen not to do high end cards through a lack of high end PowerVR tech.

err if that were true (which its not) then it wouldn't be isolationism.. more like zealotry. By isolationism I assumed you meant that I never assosiate with any other graphics card.

I don't shift blame from IMGTEC. When someone says something like "IMGTEC need to speed up the release schedules for their next few products, they also need to make a high end product and" for instance, I merely tell that persons what I have been told many times. That STM set the release dates for thier PowerVR based graphics chips and STM decide which market to go into not IMGTEC. STM decide which design they will base there product off and then licence that design from IMGTEC. IMGTEC cannot force STM to licence what would be considered a higher end design nor can they go over to STM and beat them up if they release late. If a PowerVR product comes out late then that is in the end STM's responsibility as its there product. As for the part about the bet.. well you don't think I might have been joking when I suggested I would actually do that if I lost do you? Maybe a sense of humour transplant is needed here.. STAT.

I wasn't aware it was already after September?.. wow how time has flown since earlier today when I could have sworn it was still early March. In other words do not assume I have lost that bet yet.

I don't base my bets on rumour, as for common sense it was anything but common sense to suggest that it was going to be over 8 month until a new generation of PowerVR products (Kyro III). Tell me exactly what were you basing that on?.. it seemed totally unlikely to me then and it seems totally unlikely to me now. I admit though I have been known to be very optimistic about some rumours in the past, but thats just my nature, I'm an optimistic person most of the time.

And you can't think for yourself then?

Talking of spinning things to put them in a possitive light for IMGTEC you consistently do the opposite. I suggest that maybe STM's graphics engineers are not talented enough to make a good chip on a highend design and you assume that means the design is overly complex, that is one possibility, the other is that STM's graphics engineers are not as talented as those at ATI or Nvidia for instance (not that I'm saying that is true, its just a possibility).

I didn't say this specific point was IMO the reason behind not having a high end PowerVR card, I said it was one of the possibilities.

Clearly there is a large technical advantage though.

I have given my opinion on that many times already, thinking you might be able to compete or beat Nvidia in the high end doesn't mean any company will drop what there doing and move a massive amount of resources into the highend market, STM are not primarily a graphics chip company.

When compared to the desktop graphics market Sega arcade and console machines is guarenteed money.

Hardly short term money, they made the deal with Sega in 98 and 4 years later their still selling chips in Naomi 2 machines.

I don't know, do you?

Yes I do wonder where you get it from and I continue to wonder where you get it from as I haven't said it. Statements like that??.. I'm simply telling you what others have said to me, STM don't want to go into the highend graphics chip market regardless of what tech is available. The desktop graphics market is not their companies focus, they don't put much money into their graphics division and the highend market is a risky business, riskier then most.

I have already done so, I have suggested many possibilities that someone could take from the fact that there is no high end PowerVR card. Like maybe IMGTEC can't do a highend design because there no talented enough, or STM can't do a highend design because there not talented enough ect. Why exactly do you wish me to speculate?.

No, but that could be added, as for heat, my Kyro II has a small heatsink and no fan and isn't very hot at all.

Oh yeah man you caught me.. well done, I simply said that so I would "have an out"?? If the company is not willing to go into the graphics market aggressively then there not likely to try to take Nvidia and ATI on in the high end are they?

Would you kindly not tell me what I will be saying in the future, you have no idea what I will be saying in the future.

If thats the truth as I know it then yes I will say that, if I find otherwise then I won't, if you don't believe that then that's your problem, I suppose we'll see won't we.

I don't believe I said that you read that did I?

That seems obvious to you why?.. because thats what you want to beleive? STM and IMGTEC have been in negotiations with a certain company ever since they announced that there graphics division was for sale. Just because STM have not announced anything yet it does not mean they have not already completed the deal to sell their graphics division, a few sites have recently claimed that a certain company is waiting till Cebit to make a big announcment about the purchase of STM's graphics division, I suppose you'll see soon enough wether that's true or not considering Cebit starts tomorrow.

 

How many times can you repeat this same argument? So, PC graphics is "enough" of a priority for STM to introduce a chipset in the highly competitive mid-range segement, but "not enough" priority to introduce a chip-set in what could arguably be less competitive high-end maeket?

WHY did STM re-enter the graphics chip business to begin with? I keep telling you...because they felt they had an opportunity to make money.

Yes, but the PART of my argument that you keep ignoring is the possibility that the REASON for their "lack of ambition" in the graphcis market is BECAUSE they don't feel IMG TECH'S TECHNOLOGY (or business model) gives them reason to BE ambitious.

And is that not a valid possibility? Now, let me refine that thought to be absolutely clear. I am guessing that the combination of PowerVR design, in addtion to the PowerVR "business model", does not give any chip-maker the confidence needed that they COULD be successful in the high-end market.

ANYONE can "design" a high-end part. It's the ability to take that design and move it into a working and shippable part that matters.

The point is, you won't even guess. You keep on telling me that my "guesses" have no basis, but don't offer your own guess.

Zealotry it is, then.

And that is an example, IMO, of a zealotrous(?) statement. You are neglecting to consider the fact that part of STM's decision will be based on the technology that's presented to them. You don't place ANY "burden" on IMG for the decisions that IMG makes.

You are ASSUMING, that going into the negotiations, STM already decided to not make any "high end" parts. You fail to appreciate the liklihood that STM CONSIDERED all of IMG's designs, and based on what they saw, THEN decided that "high-end" wasn't a good idea.

Again, you wash the hands of any "blame" on IMG TECH. You are ASSUMING that IMGTEC provided STM with an attractive high-end part, which STM just decided "not to do." You are ASSUMIUNG that any delays are due to STM engineering, and not to a combination of STM engineering, PowerVR design, and IMGTECG business model.

Is it ultimately the responsibility of STM to deliver on the scedule they set forth? Yup. Can there be reasons for late delivery that are beyond the control of STM engineers, and rest on IMG TECH engineers? Yup.

Obviously. The point is, no one that was not 100% sure of himself would make a bet with such outrageous terms. Perhaps some common sense is needed here...

True. But I'll ask you this question...would you make the same bet again given the current circumstances?

I was basing my guess primarily on the problems everyone's been having with advanced fab processes, and on the historical track record of IMGTECH / partner deliveries.

Um, I HAVE thought for myself, and my opinion is the best that I can come up with. Now you're asking me to think for you, to come up with an alternative to argue against myself? My, that's a clever strategy you've got there.

Again, I am asking for YOUR opinion (besides "the partners just weren't interested in the high-end market) for why the partners haven't made a high-end part. Alternatively, provide your opinion/guess as to WHY these partners aren't interested in the high-end market, considering PowerVR tech is "supposed" to have a significant price / performance advantage.

Then why is it that IMGTECH and it's partners never seem to meet the expectations of the PowerVR "camp", while at the same time end up producing when and what I expect them to be? Could it be that what you see as the "opposite of optimism" is actually a realistic point of view, rather than negative?

I didn't assume anything. I flat out said it was a possibility. The whole point of that exercise (apparently wasted) was to get you to realize that the examples you are giving do not refute my opinion. Because in all the examples you give, the "reason" behind those examples can point to problems with IMG tech or its business model.

I've been trying for the past dozen or so posts to get you to go more DEEPLY into your reasons. Don't just give a skin-deep response. You said not talented enough to make a chip based on a "high-end" design. What about talented enough to make a chip based on IMG-TECH'S high-end design?

No, clearly there is not a large "technical edvantage" when you consider the entire product and the target market. Isn't that the point we are arguing about? Clearly, there is a SPECIFIC memory bandwidth advantage when it comes to fill-rate. This does not translate directly to a large technical advantage for the overall product. Isn't there a semi-annual discussion on these boards about large poly counts and relative bandwidth advantage?

Let me put this another way so that maybe you'll understand.

If IMG-TECH came to STM and said "with this design, you can build a high-end part that will sell retail for about $10 US, and will be 1000x faster than ATI's and nVidia's high-end"....do you think STM would say "nah...we're not "interested" in the high-end.

Obviously, there is SOME POINT between the absurd example above, and what IMG-TEC actually offered, that would entice STM (or some other chip-maker) into the "high-end." And obviously, IMG-Tech has not offered any chip-maker such a product. Now, where that "point" is will vary from chip-maker to chip-maker, as every company has different risk/reward points.

But the fact that NO company has apparently been persuaded by IMGTECH's "high-end design", should tell you something.

Yes, I gave you my GUESS. The PowerVR tech and IMG's business model doesn't particularly favor a profitable business in the PC space. Now I'm asking for your guess...again.

Well, that's the first time I've seen you specualte that perhaps IMGTECH isn't "talented" enough to design a profit-making part for the high-end.

But that's not what I'm after. What I am trying to get you to speculate is nothing specific. Just SOMETHING that cannot be explained by my own guess. Because you are continually harping on my opinion as overly negative and pessimistic, yet when you try to refute it, you only provide things that could be actually be explained nicely by my opinion.

Basically, if you find my guess so "pessimistic" or otherwise repulsive, I'm looking for ANY alternative opinion from you that cannot possibly be explained by my gess.

And what would that cost in terms of redesign? How long would it take? ATI's and nVidia's current parts already have the r&D and design cost refelcted in their parts sale price. Again, you have made an overly "optimisitic" presentation of Kyro as a mobile part, because you haven't considered these things.

On the other hand, if a company is presented with tech and a business structure such that they believe they will be successful in the high-end, they will take on the competitors.

You keep on presenting your stance as if these chip companies decide on what markets they will go after, before even looking at the tech they will use to go after it. Doesn't it make more sense for the business to look at the tech available, and THEN decide which market(s) are appropriate? Sure, a company might have a pre-disposition to concentrate on one market, but if the same tech can expand their business, don't you think they'd consider it?

And? This means that negotiations with other companies are off limits? If someone doesn't offer STM a better deal than whatever company you're talking about, STM wouldn't go for it?

Gee, I can play that game too. Just because Creative announed their deal a few days ago, doesn't mean it wasn't "done" earlier than that. I thought you didn't make your decisions and guesses based on rumors? Guess that doesn't apply here...

Look, when I asked you for a reason Creative didn't buy STM, you said it could be because someone else bought STM first. Anything's possible, but if don't base your guesses on "rumors" then how can you reach that conclusion given the fact that Creative already announced their intended purchase, and no one has done the same for STM?

 

Creative didn't start negotiations with 3DLabs yesterday, nor do deals like that happen over night.

Anyone interested with some insider contacts can simply ask how long Creative has been researching to expand into alternative markets.

You may excuse my interruption now. Carry on with the usual cross-quoting pattern *ahem*

 

: Creative didn't start negotiations with 3DLabs yesterday, nor do deals like that happen over night.

Certainly true but that doesn't mean that the deal was sealed without CL considering STM's graphic division.

 

http://www.theinquirer.net/13030205.htm

 

He he ...cool...this gets more and more interesting, if this comes out true, then the possibility to see a "high-end" powerVR product in the "near" future has to be quite big now when considering that VIA is way more into the PC market than ST ever was.....or ??

/ Tompa

 

Nah...I'm sure Via just "isn't interested" in the high-end. They never have been before. It doesn't matter what tech IMG or STM has, they just don't want to be in that market

All sacrasm aside, If VIA is the STM purchaser, I hopw that they DO go after the high-end, and deliver a PowerVR based part designed to make full use of the latest high-speed 128 bit DDR memory available!