Poll on future dx9 card's ( not ati and nvidia ) performance

[Ailuros]

Who tells you XGI hasn't changed the specs ?
After all, they didn't announce anything, no specs given yet, which gives them the freedom to change anything they want.
And there may be more to it than a simple renaming.

Because even the fastest and biggest IHVs take a minimum of 18 months cycles from chalkboard design to mass production under IDEAL conditions. And mind you those IHVs throw in each generation multiple hundreds of millions of dollars for R&D.

 

[Joe DeFuria]

All the indications so far from their publicly announced AGM meetings feedback dictate that their forecasts and targets are to increase the volume of developed chips at the same time from about 10 today to 30-40 (under ideal conditions) in the foreseeable future.

I'm not sure what you mean. PowerVR doesn't develop chips. Are we talking about 30-40 distinct license designs simultaneously available for licensing...or 30-40 productized chip designs in development with 3rd parties?

 

[Captain Chickenpants]

And what percentage of these are actually productized?

I don't really know and couldn't say percentages but just to list info on our corporate web-site we are doing well with the DAB radio sales, not just our own branded ones but other products with our chip in.


Home Cinema products. Quite a few Licensees for MBX, Meta etc. Still sales of arcade related products.

It may also be why other companies that are working on fewer projects, actually seem to ship a higher percentage of those products being worked on....

Possibly. But as we shifted to the more IP Licensing model only a few years ago, and are starting to see the benefits coming through.

I just think that if you want to compete in the high-end PC space, (honest question: do you?), the IP model is a hindrance.

Speaking for myself I definatly want to compete in the high-end, as a company it is not my decision. (They have cleverer and better paid people for that ).


Your points are all valid , but the pros are more compelling to us than the cons. We have investors to satisfy as well as gamers.


(ALL MY OWN VIEWS not IMG's)

CC

 

[Ailuros]

30-40 license designs deriving from X number of produced chips. MBX was a developed chip by ImgTec (at least the prototype) AFAIK. In reality 2 or 3 different models for licensing evolved from that.

I wouldn't guestimate that it's much different with Series5 either; a high end graphics core gets released by whoever and on it's back they'll gain X amount of additional lisences/spin offs for Y different markets, like console, arcade or whatever markets or potential licenscees might demand.

 

[Joe DeFuria]

Speaking for myself I definatly want to compete in the high-end, as a company it is not my decision. (They have cleverer and better paid people for that ).

That's basically my position as well. My feeling is that those "clever people" are deciding that the desktop PC business is not one they are interested in going after with any type of urgency. This may indeed be the "right" decision for PowerVR as a company.

But at the same time it means less and less liklihood of seeing a killer PowerVR product in the high-end PC space, which is what I'd personally like to see.

Your points are all valid , but the pros are more compelling to us than the cons. We have investors to satisfy as well as gamers.

Only, my gut feeling is, that you (meaning PowerVR management) has decided that "satisfying gamers" is not really the business you want to be in. (Again, perfectly valid from a business perspective....just disappointing.)

 

[Ailuros]

But at the same time it means less and less liklihood of seeing a killer PowerVR product in the high-end PC space, which is what I'd personally like to see.

If Series5 gets released on time you'll have a high end PC part from them.

 

[Joe DeFuria]

If Series5 gets released on time you'll have a high end PC part from them.

I thought "on time" has 2H '03....What might be the cause for any delays? Their IP licensing model, perhaps?

 

[Ailuros]

No one EVER said anything about a 03´ release, just an announcement. An on time release should be translated at about the same timeframe as all other PS/VS3.0 hardware.

NV released the NV30 with how much of a delay? Must have been their fabless semiconductor business scheme there at fault. I´d say that chips with even higher complexity might have chances to run into uncontrollable trouble, yet that goes for any IHV out there.

 

[Joe DeFuria]

No one EVER said anything about a 03´ release, just an announcement.

Yeah, I know....the PowerVR faithful always pull the "they're likely to release it in '03!" back in late '02....and then change the story to "announcement" as '03 ends.

Always the optimist, until one is forced to change the story...that's how people end up eating their own body parts.

http://www.powervr.org.uk/

Well it seems that IMGTEC have now confirmed what I've known for a little while now in this EE Times article.

PowerVR Series 4 will not be coming to the PC space and PowerVR Series 5 will be making an apperance in 2003. I can actually be a bit more specific and say that this will be in the first half of 2003 and as the statements from IMGTEC suggests the chip will be extremely feature rich.

So, it was going to make an "appearance" in 1H 2003, but not ship until 2004?

To Teasy's credit, he did say " this isn't all 100% deffinate yet and the proof will be in the pudding."

An on time release should be translated at about the same timeframe as all other PS/VS3.0 hardware.

Depends on what market it's targeted at. And it may be "on time"...the question is, at which market will it be targeted at?

NV released the NV30 with how much of a delay?

Months.

Must have been their fabless semiconductor business scheme there at fault.

No, there are other circumstances behind their own problems. NV30 did ship though, didn't it? nVidia does have parts that compete in the top of the line segment, do they not? The point is, NVIDIA had complete control over the situation to ship or not, and was not at the mercy of STM, NEC, or who knows who else to make that decision.

I´d say that chips with even higher complexity might have chances to run into uncontrollable trouble, yet that goes for any IHV out there.

Agreed...but the fabless semi-cons are in the best position to mitigate that risk.

 

[Ailuros]

Yeah, I know....the PowerVR faithful always pull the "they're likely to release it in '03!" back in late '02....and then change the story to "announcement" as '03 ends.

Always the optimist, until one is forced to change the story...that's how people end up eating their own body parts.

You wouldn´t be better off discussing that specific matter with the original source then wouldn´t you? I don´t have necessarily to share everyone elses opinion either. Metcalfe stated in October 2002 that Series will debut in 2003.

I don´t care what everyone else says or thinks, but what I personally think or say in public.

Depends on what market it's targeted at. And it may be "on time"...the question is, at which market will it be targeted at?

How many times do they have to state that they´re targeting the high end market until people actually realize it?

No, there are other circumstances behind their own problems. NV30 did ship though, didn't it? nVidia does have parts that compete in the top of the line segment, do they not? The point is, NVIDIA had complete control over the situation to ship or not, and was not at the mercy of STM, NEC, or who knows who else to make that decision.

At least with ST Micro layout and mass production wasn´t overlooked by PowerVR. What if I´d tell that this time they will?

Agreed...but the fabless semi-cons are in the best position to mitigate that risk.

See above. All I was trying to say is that if there should appear any uncontrollable factors that would delay things, is that there chances in all cases with high complexity chips. NV40 was supposed to be released before the end of this year; please tell me that I´m so inherently wrong in rather guestimating a paper launch. Likewise ATI should have normally released it´s next generation part this fall too (one year past R300) etc etc. All I can see at the moment are revamps of revamps. That´s exactly the best position you´re talking about here I figure.

And while we´re at it PS/VS3.0 products all had to experience minor delays and that not alone because of chip complexity. Ask Microsoft or ram manufacturers if they have anything to say about it.

Why can´t we just finally wait until this year runs out, if and what they´ll announce and then we can catch up from there?

 

[Joe DeFuria]

You wouldn´t be better off discussing that specific matter with the original source then wouldn´t you?

Or with you who said "No one EVER said anything about a 03´ release, just an announcement. "

I don´t have necessarily to share everyone elses opinion either.

Of course...then just don't speak for everyone. Because SOMEONE (there are only a handful of devoted PowerVR fans on this board) certainly was implying an '03 release.

Moving on...

Metcalfe stated in October 2002 that Series will debut in 2003.

Yes he did. And specifically, debut on 0.13 micron...which implies, imo, the debut of a part, not just tech.

How many times do they have to state that they´re targeting the high end market until people actually realize it?

How many times do I have to state that I'm less concerned about what they claim to be targeting, than I am with their ability to bring it to market.

Take an EXTREME example: what if BitBoys said they planned to bring a high-end part to market later this year? (No, I'm not putting PowerVR on the same level as bitboys...but I hope you understand the point.)

At least with ST Micro layout and mass production wasn´t overlooked by PowerVR. What if I´d tell that this time they will?

Then I'd say that's a step in the right direction.

See above. All I was trying to say is that if there should appear any uncontrollable factors that would delay things, is that there chances in all cases with high complexity chips.

Agreed, and this was always the case.

This doesn't change the fact that there has NEVER been a competitive high-end PowerVR PC part since the original PCX-PCX-2. In other words, PowerVR isn't any more isolated from the same challenges as everyone else...they have an additional layer of challenges.

NV40 was supposed to be released before the end of this year; please tell me that I´m so inherently wrong in rather guestimating a paper launch.

I'm not even too confident in that. Won't surprise me to hear nothing of NV40 or R420 'till next year.

Likewise ATI should have normally released it´s next generation part this fall too (one year past R300) etc etc. All I can see at the moment are revamps of revamps. That´s exactly the best position you´re talking about here I figure.

Yup...just like always. And yet, despite there being revamps of revamps in the past...PowerVR was not found to compete.

Why can´t we just finally wait until this year runs out, if and what they´ll announce and then we can catch up from there?

I'm more than happy to wait. You just don't seem to be happy that I'm not holding my breath at the same time.

 

[Ailuros]

Or with you who said "No one EVER said anything about a 03´ release, just an announcement. "

I had public company announcements in mind.

Of course...then just don't speak for everyone. Because SOMEONE (there are only a handful of devoted PowerVR fans on this board) certainly was implying an '03 release.

That´s a pathetic reasoning to handle an argument. I speak only for myself and I don´t have to add to every other post that it´s all IMHO. Again address them about that and stick to what I am saying when you´re talking to me.

Yes he did. And specifically, debut on 0.13 micron...which implies, imo, the debut of a part, not just tech.

And who guarantees as of yet that it won´t? Who says that an announcement can´t be done with a full technology presentation? No one but PowerVR/ImgTec employees can possibly know such details at this stage anyway.

Take an EXTREME example: what if BitBoys said they planned to bring a high-end part to market later this year? (No, I'm not putting PowerVR on the same level as bitboys...but I hope you understand the point.)

Officially we´ve been made aware that they won´t develop for the time being another graphics part, after Axe was dropped. On the other hand we´ve been also repeatedly made aware that PowerVR is in fact developing a new part. What am I missing here? BB would have developed a high end part in a couple of months or what in such a case scenario. John said on these boards himself that they´ve been working on S5 a bit less than 2 years.

PowerVR isn't any more isolated from the same challenges as everyone else...they have an additional layer of challenges.

Advantages as disadvantages like with about everything. The advantage in that case would be what John mentioned a page or so ago; they haven´t had neither the necessary "compatibility" transitions from former parts, nor anything else to distress them from but just concentrate on that project. Plus you seem to forget that a rather crucial manufacturing process has been ironed out by a 3rd party and that others can always learn from their design mistakes too. You don´t believe that IHVs do not in fact put competing products under the "microscope" do you?

I'm not even too confident in that. Won't surprise me to hear nothing of NV40 or R420 'till next year.

I wouldn´t be surprised to see three high end PS/VS3.0 parts at about the same timeframe next year, yet if you´d ask for guarantess or bets I´m telling you out front that I´ll pass. I can´t be sure about anything and that should be understandable.

Yup...just like always. And yet, despite there being revamps of revamps in the past...PowerVR was not found to compete.

Albeit a fact it´s still irrelevant to my point.

I'm more than happy to wait. You just don't seem to be happy that I'm not holding my breath at the same time.

My own personal perspective: up until now I have good reason to hold my breath more than I did in the past. Kyro3 was supposed according to that Fu persona from ST M to launch a couple of months past that idiotic K2se in late summer, which for everyone would have meant late fall or at least in the winter. After there wasn´t a sign of it when the summer was over I started getting those funny itches again that there´s something not right there. Personally I´ll start worrying, if they won´t announce anything this year, until then I have headroom for optimism.

Not because of anything else, but because they have undergone some minor yet significant changes in their workflow plans that justify said optimism.

Finally since I´d be lying to not admit that I´m completely blown out of my socks with the R300, I´m just as anxiously awaiting ATI´s next effort; and yes this time it is rather NV (between the two) that has to convince me.

 

[Joe DeFuria]

I had public company announcements in mind.

Then don't go and EMPHASIZE "NO ONE SAID." Say what you mean. PowerVR never said.

Though again, it can be interpreted that PowerVR was implying an actual product in '03, but I agree it's open to interpretation.

And who guarantees as of yet that it won´t? Who says that an announcement can´t be done with a full technology presentation? No one but PowerVR/ImgTec employees can possibly know such details at this stage anyway.

Agreed.

Officially we´ve been made aware that they won´t develop for the time being another graphics part, after Axe was dropped. On the other hand we´ve been also repeatedly made aware that PowerVR is in fact developing a new part. What am I missing here?

Sigh.

Work with me here. If BB comes out with a PR tomorrow saying "we've got a killer chip planned for debut in 2004"....what would your reaction be?

Would it be "ALRIGHT! CAN'T WAIT!!!"

Or would it be more like "I'll believe it when I see it."

In short: I don't care WHAT THEIR PLANS ARE. I've already witnessed their "plans" not pan out quite as planned much more often than not.

Advantages as disadvantages like with about everything.

Yes.

The advantage in that case would be what John mentioned a page or so ago; they haven´t had neither the necessary "compatibility" transitions from former parts, nor anything else to distress them from but just concentrate on that project.

They also don't have the experience of having built a mass produced pixel shader and drivers to go with it.

Plus you seem to forget that a rather crucial manufacturing process has been ironed out by a 3rd party and that others can always learn from their design mistakes too. You don´t believe that IHVs do not in fact put competing products under the "microscope" do you?

They can't learn as much from "other's mistakes" as the others themselves. You do realize that, don't you?

I wouldn´t be surprised to see three high end PS/VS3.0 parts at about the same timeframe next year,

I would.

yet if you´d ask for guarantess or bets I´m telling you out front that I´ll pass. I can´t be sure about anything and that should be understandable.

Yes, it is.

My own personal perspective: up until now I have good reason to hold my breath more than I did in the past. Kyro3 was supposed according to that Fu persona from ST M to launch a couple of months past that idiotic K2se in late summer, which for everyone would have meant late fall or at least in the winter. After there wasn´t a sign of it when the summer was over I started getting those funny itches again that there´s something not right there. Personally I´ll start worrying, if they won´t announce anything this year, until then I have headroom for optimism.

I take the opposite approach: I worry until I see some announcement, then I re-evaluate if I should still be worried based on the announcement.

Not because of anything else, but because they have undergone some minor yet significant changes in their workflow plans that justify said optimism.

My reasoning is that whatever changes they've made, they still aren't calling their own shots.

Finally since I´d be lying to not admit that I´m completely blown out of my socks with the R300, I´m just as anxiously awaiting ATI´s next effort; and yes this time it is rather NV (between the two) that has to convince me.

For me, I am very "uninterested" in PS 3.0...from any vendor. DX9 PS 2.0 is and will be IMO the DX9 standard. PS 3.0 will be the Ps 1.4 for DX9. Nice to have, fun to play with for developers and hobbyists, but won't mean much for the consumer. That'll have to wait until DX10.

I'm very interested in the spring '04 products not because of any new shading capability, but on the hopes for another big jump in performance, and perhaps increases in filtering / AA options and quality.

 

[Dave H]

Metcalfe stated in October 2002 that Series will debut in 2003.

Yes he did. And specifically, debut on 0.13 micron...which implies, imo, the debut of a part, not just tech.

To me this implies only that the IC is sampling, not volume availability of end-user cards.

Anyways, carry on...

 

[Joe DeFuria]

To me this implies only that the IC is sampling, not volume availability of end-user cards.

Anyways, carry on...

I didn't claim high-volume of end-user cards.

My main point was that it implies more than a press release with paper specs. If we see a product launch with working silicon in '03, I'd consider the prediction of a "debut in '03" to be met.

 

[Ailuros]

Work with me here. If BB comes out with a PR tomorrow saying "we've got a killer chip planned for debut in 2004"....what would your reaction be?

Would it be "ALRIGHT! CAN'T WAIT!!!"

Or would it be more like "I'll believe it when I see it."

I didn't say or think that either when I heard about Acceleon. If I would have any doubts than it would concern everything they would release, since there hasn't been anything purchasable at all to date. I know what your answer to that would be, but as I said times and times before, I've learned never to say never again.

Frankly albeit just a paradigm, I'm not very fond of multichip sollutions either, so the interest would be limited anyway.

Believe it or not I AM looking forward to see DeltaChrome as a more realistic example, albeit S3 didn't exactly have exeptional hard- or software in the past and their absence from the market is longer than everyone else's. Mind you change the line to "I'll judge it when I see it" and we'll have an agreement.

If I mention prejudism based on track record, you'll deny it again.

They also don't have the experience of having built a mass produced pixel shader and drivers to go with it.

They're neither begineers, nor a couple of students that woke up one morning and decided to build a high end chip in a garage. Don't take me there either; there I could add a couple of not so flattering remarks for a market leader with fairly the biggest amount of experienced professionals in it's portofolio.

They can't learn as much from "other's mistakes" as the others themselves. You do realize that, don't you?

Question is if the others have had actually enough time to correct all mistakes in the meantime; a promise that was made with a refresh product, but I haven't seen it yet happening.

My reasoning is that whatever changes they've made, they still aren't calling their own shots.

Relative again, but I'll agree, since it's not going to change anytime soon.

For me, I am very "uninterested" in PS 3.0...from any vendor. DX9 PS 2.0 is and will be IMO the DX9 standard. PS 3.0 will be the Ps 1.4 for DX9. Nice to have, fun to play with for developers and hobbyists, but won't mean much for the consumer. That'll have to wait until DX10.

Personal opinion and as that respected. Up to 2.0 you're bound (very oversimplyfied) to render all shader ops you're going to use. With 3.0 if your hardware has instruction slots up the wazoo you can effectively code a whole game based on shaders, just because in reality you'll end up rendering only a fraction of the instructions you're going to call for.

In my rather simplistic opinion I'd say that simple PS2.0 is much closer to PS1.4/DX8.1 and PS/VS3.0 much closer to whatever shape dx10 might take in 2005. Granted 3.0 could have made up for a new DX version entirely. If they end up concentrating just on PS2.0 we'll end up with another bunch of pity reflections in a few pits here and there.

I'm very interested in the spring '04 products not because of any new shading capability, but on the hopes for another big jump in performance, and perhaps increases in filtering / AA options and quality.

They could have easily dedicated a large amount of transistors for an insane amount of MSAA samples, but then again others would be bickering about it's lack-lustering compliance and hardly any developer will ever have a look at it again. AF won't be a problem this round, nor texture filtering in general; when it comes to MSAA IMRs have already cut off with fillrate/bandwidth saving techniques in that department a lot of a TBDRs advantages.

They apparently can't satisfy them all. But if we truly mean a high end contender to the other two, it has to be as feature complete and have comparable performance under all situations. I don't think I need to write a long winded drivel how much wizzbang features count in order to win impressions, recognition and deals.

Finally they've been trumpeting their architecture's superiority for ages; a true high end board is the best presupposition to finally prove it. No one will ever be convinced with a mainstream or budget part.

 

[Entropy]

The big questionmark is if they have a liscensee.
We know that TI, Intel and so on have liscenced the MBX design.
On the Series5 front however, not only have we seen no such announcement, but we haven't seen any mention of it in any financial reports or other sources, we haven't heard of anyone reserving fab space, nor anyone who would seem to be gearing up for a launch.

The PC market wasnt even mentioned anymore in the AGM presentation, this is getting ugly.

What we do have however is very clear statements to the effect that ImgTec will not bring the part to market themselves. So it seems simple enough. If they do not have a top-secret liscensee (in this industry? Hah!), the Series5 won't reach the market.

The only cheering factor here is the attitude the PowerVR people have. They aren't saying anything, but they are keeping silent in a pretty bullish way.

I'm with Ailuros on this, until 2003 is over and we still haven't seen anything, there is still space for hope that we will see the entrance of something a bit more interesting than another IMR with tricks up its sleeve. If the XGI parts and the DeltaChrome makes it to market, that's great, and good for everyone. But I must say that a PowerVR design would create interest in ways that these, or for that matter a 120 W brute force monster from ATI or nVidia, can't bring to the industry.

Entropy

 

[darkblu]

i gave my vote for deltaChrome: that part looks to be a very well ballanced dx9 desktop contender

If you say so, because we have no idea of the clockspeed, features, etc. All we know is that it's a 0.13 micron 8-pipeline GFX engine which supports DX9, a bit vague IMO.

actually we have an idea of the features http://www.s3graphics.com/DeltaChromeDX9.html

And Deltachrome is depicted by S3 themselves as mainstram/low end :

We will leave the hardcore minority to ATI and nVidia, we'll focus on the normal folks with normal budgets. http://www.homelanfed.com/index.php?id=12478.

that only says they won't be targeting the high margins market segment. we indeed know nothing about the performance. nevertheless, i find the featureset to be well ballanced.

I remember this :
Tuesday, January 7th, 2003 : S3 Graphics Launches All New DeltaChrome with Dazzling Hi-Def Graphics and Video with Microsoft DirectX 9.0 and HLSL support
http://www.s3graphics.com/pressrel/2003_01_07.html

Nine months later and still nothing...

again, check the s3graphics site.

 

[parhelia]

How many times do they have to state that they´re targeting the high end market until people actually realize it?

Hey, XGI stated too they were targeting the high end market, yet you don't believe them...

Why can´t we just finally wait until this month runs out, then make up our mind?

And while we´re at it PS/VS3.0 products all had to experience minor delays and that not alone because of chip complexity

True, though this goes against all those who were saying in another thread that any IHV was doomed if their soon to be released chip didn't support PS 3.0 ...

 

[Dave Baumann]

True, though this goes against all those who were saying in another thread that any IHV was doomed if their soon to be released chip didn't support PS 3.0 ...

Funny, I didn't see anyone saying they were "doomed".

Parhelia, come out with it now please - what is your relationship with XGI?