Does this surprise anyone?

[JAYDEEJOHN]

http://www.brightsideofnews.com/news/2009/8/28/nvidia-plans-gt300-demos-for-late-september.aspx

 

[daedalus685]

Interesting.

I would have been worried that the 300 series may be years away if they didnt ahve something to tease with by the 5000 series launch. I am not very familiar with the history of Nvidia releases, does this stack up to the past? Anyone with an educated guess as to what this means so far as how long to market after these events? Given ATI had hardware to show off months ago, I'd still have to assume late november.

 

[JAYDEEJOHN]

Early June for the DX11 demos.
Several things I find interesting here is, the ongoing fight between theo and Charlie, where Charlie has claimed nVidia would be doing this very thing, but with a handful of cards, where we find Theo somewhat backing nVidia here, even making up crap about its drivers
Historically, nVidia has done many things to cause a halt to reality when they werent ready, ala DX10.1, not just releases
I wonder now how unimportant DX11 and W7 will be now?
Oh, and how great GDDR5 will be as well

 

[jennyh]

3 months since ATI demo'd at Computex. If you recall, TSMC also presented them with a wafer of 40nm dx11 cpu's.

That's 3 months of churning these things out for the September 10th launch. Nvidia on the other hand, will have a few working dx11 parts...probably hack jobs at that. If Nvidia had working dx11 demo's we'd have seen them by now btw.

If you believe 5x is out on Sept 10th, the very fastest Nvidia will be able to release the g300 is December. They are really praying it is ready by then, but I'm sticking to January/Feb release which is what I said in June.

Nvidia are *at least* 3 months behind ATI this round, and possibly a lot more.

 

[daedalus685]

Theo and Charlie can both stop writing articles any time now... Baffling how they still ahve jobs.. at any rate..

Though, if that article is true at all, we are a month away from even getting engineering samples of nvidia cards. That certainly is bad news for those of us wanting to buy some cheep GPU's around christmas . I hope this doesnt spell the begining of a 100 million dollar "dx11 is not worth it so don't make games for it" tirade.. I'm sick of playing the same games I've had for years..

 

[JAYDEEJOHN]

What I find interesting also, is the coming of LRB. Since LRB will be touting the very things nVidia has been since CUDA was more than a car or a fish, and Ive also noticed more people praising LRB that own nVidia cards.
This will hurt their marketshare more than anything, as I think ATI actually has a more dedicated base of customers.
Coming late, and coming late and missing on a new OS launch, the renaming, the high pricing, AC, Batman, all these things are pushing nVidia in a very bad direction.
Its mo wonder Theo was pumping Tesla so much in the article.
I think people are tired of this, and some will show it in their purchases
Some say, if nVidia hadnt pursued CUDA, there may have never been a LRB

If this turns out to be nothing but viral marketing, its just another huge nail in their coffin

If not kudos to them, but the G300 better be a gamer too

 

[JAYDEEJOHN]

Lets wait n see what ATIs pricing will be, even without competition. Yes, theyll likely be somewhat higher priced without it, but its not what theyre known for, and it wasnt high prices and high perf that has gained them marketshare.
I dont see them swaying much away from this approach

 

[jennyh]

Here is what is important.

The first dx11 cards in the hands of games devs will be ATI's. They will be getting to work on them asap. For at least 3 months, the only dx11 parts that games devs can work with will be ATI's. That automatically means the initial dx11 games will favour ATI strongly. Nvidia can't be bribing them with TWIMTPB, they can't afford it now anyway but they can't do that period until they actually have dx11 cards.

When Dirt 2 is out in full dx11, any other racing game that isn't dx11 is going to be compared unfavourably to it. What game dev that is developing a racing game would want that? None of them, once dirt 2 is released, all racing games will have to be dx11 if they are to compete.

There will be an uptake on it for that reason. It's not just dirt 2 either, simple fact is every games developer will be using an ATI dx11 card for the next few months at least. Anything else would be suicide. Being first out is incredibly important, that's why Nvidia held the lead for so long and why they have now lost it.

 

[JAYDEEJOHN]

We wont have long to wait. DIRT2 could be seen as a killer must have game, plus DX11. Then, itd go hand n hand

 

[jennyh]

Dinosaurs with win xp and dx9 can keep it for all I care. I'll be enjoying dx11, we'll see how long you can hold out stranger.

 

[JAYDEEJOHN]

Demo looked real good. It may be a good seller, who knows, not a big racing fan.
Point is, any DX11 game could be the next have to have game, and the DX11 association goes with it

 

[JAYDEEJOHN]

Have you read what the devs using the new frostbite DX10 engine have to say about it?
Its not a DX9 port.

 

[JAYDEEJOHN]

"DX 10 was designed in a way to reduce the CPU overhead per draw call. But this doesn’t help if the engine GPU abstraction layer was build for a DX9 like API. This way you end mapping a DX9 like interface to DX 10 and waste all the CPU power that you can win by using DX10.

You can even make bigger mistakes that would make DX10 slower then 9."
http://forum.beyond3d.com/showpost.php?p=1326798&postcount=102
Thing is, like I said in another thread, once those DX10 engines are out, party over for DX9

"Porting from 10 to 11 is an easy job but with a simple port you only enable your game to use DX11 features. To get some benefits you need to use these features.

Porting from 9 to 11 is the same pain as 9 to 10. If you do it the naïve (fast) way you will end with the same bad performances we already know from bad 9 to 10 ports."
http://forum.beyond3d.com/showpost.php?p=1323210&postcount=66

 

[gamerk316]

jenny, Dirt2 isn't "full DX11", it simply supports it as an extra feature. Probably just Shaders and a few other effects. There won't be any "full" DX11 cames for some time.

Heck, a lot of engines still use DX8 features...

 

[JAYDEEJOHN]

I think TGGA explained this to you gamer, and about this type of reasoning.
You could use DX8 to make something look like raytracing, given the time and effort, whats your point?

 

[JAYDEEJOHN]

My point is, as each DX10 engine comes, so does DX10 and 11, and the porting becomes more expensive to go the other way, down to DX9.
Now, if I were a dev working with a DX9 engine today, Id be wondering how my next job would be, as I would know nothing about DX10/11 and its features in game, and since I want exposure to it, since its the future, I may have to seek it out.

Now, if I owned and ran the gaming house, Id know that consoles are coming due in 2 years. Knowing this, Id have a DX10/11 capable engine readied for it, and also have some of my games and devs using it for familiarity, its only good business practice.
Any of these happening puts an early end to DX9, or a portion of it, part of its remaining market, and it wont ever recover it either, thats my point

 

[JAYDEEJOHN]

It just seems SS, were actually saying the same thing, tho Im not so sure about gamer.
Its as tho magically, 2 years from now, thats when DX11 will fully appear, and the transition will occur quickly, and if youre like me, and it sounds like it to me, there will be a gradual transition to full DX10/11, but its already happening, and will only pick up steam from here (pun intended), and once DX9 is gone, its gone, and only non AAA games will be using it 2 years from now. Correct?

 

[Mousemonkey]

I remember that it was Doom 3 that got me (and a few others) to move to XP from ME simply by refusing to be installed on anything other than XP, bring on Doom 4 I say.

 

[The_Blood_Raven]

A full DX11 game will make sense in the coming months. If it is visually stunning it will receive high marks regardless of gameplay and people will buy it and DX11 cards so quickly they will be flying off shelves. You wonder where I get this? Does Crysis ring any bells? People are sheep, and some times that can be good, besides doesn't history always repeat itself stranger?

 

[jennyh]

If you think dx11 is more hyped than dx10 then frankly, you have a terrible memory.

Dx10 was gonna be wonderful beyond compare, remember Crysis? Then Crysis looking almost exactly the same in dx9?

Microsoft will not make that mistake again.

 

[jennyh]

Also

Dx11 will be up to 20% faster just because of multithreading. When people start seeing dx11 with 20% better fps in dx9 titles, that is what will make a huge number change over.

 

[jennyh]

Oh yeah that was a dizzy one.

 

[jennyh]

I did, ofc mean people will see a 20% difference in dx11 compared to dx9 games. Even without a dx11 card, simply adding multi-threading support to a dx9 game will make it run 20% faster or even more.

That's gonna make people want Windows 7. More people want 7, more reason to code dx11.

 

[techgeek]

You know that any games that are released in the next year or so have been in development for as much as two years already. So to think that DX11 games are coming out any time soon is wishful thinking. If it gets tacked on at the end of the development cycle, it's not going to be a show stopper and likely with look nearly the same in DX9 and DX10. In most cases, if you look carefully at the titles released from now to a year from now, you'd be surprised to find that development is still centered around DX9. Look at STEAM's stats, most people are still using DX9 capable cards and those using DX10 capable cards on XP still outnumber those using DX10 cards in Vista (though admittedly it's close). Game developers and publishers are going to want their games to run on the lion's share of the hardware out there. Sadly that is still DX9, whether on XP, Vista or Windows 7. If you look at Windows versions, over 50% of STEAM users are still using XP. I like the bleeding edge of technology as much as the next geek, but I live in this world and the market rules what technology is used in this world. The market says that making games for DX9 is going to make more money than gambling on hardware and OS uptake in the following year.

Now games beginning their development cycles now will likely focus primarily on DX10. This is good for those using DX11 hardware because DX11 isn't a huge leap from DX10. So it will be easier to support DX11 features from a DX10 code base than it was to support DX10 from a DX9 code base.

To use consoles to base an argument that DX11 is going to be pushed into the forefront in the very near future is baseless. This generations consoles are essentially DX9. When the next generation comes out they may use DX11 hardware, but we are still a few years off of the next generation console. In the mean time game development is going to be centered around DX10, so the likelihood is that the API's for the next generation consoles will be primarily DX10 when they first come out.

Now to whether or not nVidia releasing DX11 hardware later than ATI is going to hurt them, sure it is. From a practical point no, but from a perception point, yes. Anyone buying a graphics card immediately after ATI releases the 58XX cards would be think twice about buying a DX10 card. Whether this is practical or not doesn't matter. The DX11 features may never get used during the cards useful lifetime, but having those features makes the purchaser feel more future proof. This is the same thing that happened to ATI when nVidia released the 8800 series. It didn't matter that nVidia didn't have DX10 drivers out, the hardware supported DX10 and that perception drove it's sales. It didn't hurt that they also performed really well in DX9 either. Coming late to the DX10 party really hurt ATI's market share, not to mention the 2900XT's lackluster performance. Of course having DX10 hardware for so long without a functioning DX10 driver hurt nVidia in the long run too (class action suit). Hopefully nVidia won't make the blunder that ATI did with the 2900XT. I personally don't think that nVidia is going to release their DX11 hardware as late as ATI did with the 2900XT. This could help the nVidia faithful keep their resolve in holding off to see what nVidia has to offer. In the end, I think it will hurt nVidia's market share somewhat, how much depends on nVidia's execution. If they can release a DX11 part no more than a few months after the 5800, and if it can compete performance wise, the damage may not be too bad for them.

I think DX11 is a fairly big deal, but I don't think it's as big a deal as some zealot's make it out to be. To listen to them, when they get their 5800 and power it up, they expect nothing less than God staring back at them through their monitors. I think it will take a few years before most of us will really appreciate DX11 and by that time everyone will have forgotten about the first generation of DX11 cards (other than who was first to market with one).

 

[jennyh]

On the other hand, the dx10/Vista farce has made certain others far too sceptical.

Microsoft won't make the same mistake twice. Tesselation alone practically ensures that there will be a noticable gap between dx9 and dx11. The speed benefits are obvious and sure a lot of games will just have dx11 bolted on, but if that includes 20-50% better fps and better looks, what gamer is going to resist it?

Are you gonna buy a $140 260gtx or a $100 5670?