Will Eidos patch Batman:AA to remove the Vendor ID check?

[Corum Jhaelen Irsei]

Originally Posted by SIrPauly

This is how I feel personally:



Do I think it was a low blow from nVidia? Yes.

Do I like it? No

Do I think nVidia gets questionable at times -- especially when they don't have the hardware advantage? Yes

Do I think it is leverage? Yes

Do I think nVidia is disingenuous at times? Yes

How much clearer can someone be? But I don't offer emotional knee-jerk views and try to be some-what constructive.

Try to think a bit:

How does airing out dirty laundry to the community help things? To offer the illusion how good ATI is morally and nvidia is a bunch of low blow scoundrels -- and why certain titles shine on nVidia's developer relations. Attack the selling point.

Why say nVidia doesn't really care about gamers?

Because take advantage of product launched that it is obvious that ATI cares because our products are geared for gamers and not scientific applications that are boring. They actually attacked double precision and ECC in their own marketing documents as a bad thing to sell product now to the gamer.

What do you think these guys do? It's about selling product right now -- for ATI and nVidia -- and you're just a number -- just like me, imho.

After all that, you will buy their product without a moment of remorse. Nv is doing things to harm the gamer. As long as there are no repercusions, they will keep on doing it.

Edit: Here is a link to them harming the gamer. http://www.rage3d.com/board/showthread.php?t=33955180 There is no way to spin this other than to say that they are screwing a company that has made a better way to use multiple video cards. As for me, I want to be able to use the fastest method to connect the cards, to hell with Nv's profit.

[makomk]

Originally Posted by Exposed

- Nvidia worked with Eidos to provide an AA solution in DX9. Now, this is more than just a few lines of code. Apparently, the in-game AA isn't applied via brute force...it's applied intelligently, to maximize performance and anti-alias only where its needed.

- AMD didn't bother to submit anything remotely equivalent. In fact, their "solution" was too late, and I bet it wasn't as comprehensive as Nvidia's.

From what I can tell, AMD's solution was basically the same as NVidia's, actually - except it doesn't come with a restriction to NVidia hardware. See BSN's report - apparently, AMD have analysed the code in question and it uses a common, publicly-known technique recommended by both companies. The code itself also works fine on AMD hardware if the vendor check is disabled.

[SIrPauly]

Remorse? Oh, the gaming gods. One must remorse to buy a nvidia GPU!

[SIrPauly]

Originally Posted by Corum Jhaelen Irsei

After all that, you will buy their product without a moment of remorse. Nv is doing things to harm the gamer. As long as there are no repercusions, they will keep on doing it.

Edit: Here is a link to them harming the gamer. http://www.rage3d.com/board/showthread.php?t=33955180 There is no way to spin this other than to say that they are screwing a company that has made a better way to use multiple video cards. As for me, I want to be able to use the fastest method to connect the cards, to hell with Nv's profit.

Corum,

In that thread you're offering -- look down and there is a response from CJ -- a manager from MSI.

[SIrPauly]

Originally Posted by makomk

From what I can tell, AMD's solution was basically the same as NVidia's, actually - except it doesn't come with a restriction to NVidia hardware. See BSN's report - apparently, AMD have analysed the code in question and it uses a common, publicly-known technique recommended by both companies. The code itself also works fine on AMD hardware if the vendor check is disabled.

If it was so trivial and easy -- just write the code again and test it on ATI products and be done with it. Spend the little resources for the in-game feature even though it is a silly way of spending them.

This way you could still make nVidia look bad -- show them how it was a low blow but was a bigger person by spending resources for their customer base - by getting the job done -- instead of bitching about how bad nVidia is.

So nVidia may be a bully -- you don't go wining to the principle. Beat the bully by doing more and offering more than what the bully can. By solving the problem instead of bitching about it -- would be a huge positive step in the right direction for perception.

[Corum Jhaelen Irsei]

Originally Posted by SIrPauly

Corum,

In that thread your offering -- look down and there is a response from CJ -- a manager from MSI.

I suspect you are wrong about him being a manager. If he were, he would not be posting like that. I think we can disregard his post, until someone from either MSI or Nv release a press statement, it looks like Charlie is right. Hell if he works for MSI, he would know if Charlie is right about th VGA stuff.

[Corum Jhaelen Irsei]

Originally Posted by SIrPauly

If it was so trivial and easy -- just write the code again and test it on ATI products and be done with it. Spend the little resources for the in-game feature even though it is a silly way of spending them.

This way you could still make nVidia look bad -- show them how it was a low blow but was a bigger person by spending resources for their customer base - by getting the job done -- instead of bitching about how bad nVidia is.

So nVidia may be a bully -- you don't go wining to the principle. Beat the bully by doing more and offering more than what the bully can. By solving the problem instead of bitching about it -- would be a huge positive step in the right direction for perception.

Ati already did. They offered it and were refused. How many times has that been said in this and other threads so far? You are being obtuse on the issue.

The last part really pisses me off about your post. You have been reading this, you know that ATi did solve the problem, then had their solution thrown back in their face. What do you do than, when the bully owns the teachers?

[caveman-jim]

Originally Posted by makomk

From what I can tell, AMD's solution was basically the same as NVidia's, actually - except it doesn't come with a restriction to NVidia hardware. See BSN's report - apparently, AMD have analysed the code in question and it uses a common, publicly-known technique recommended by both companies. The code itself also works fine on AMD hardware if the vendor check is disabled.

What got AMD seriously aggravated was the fact that the first step of this code is done on all AMD hardware: "'Amusingly', it turns out that the first step is done for all hardware (even ours) whether AA is enabled or not! So it turns out that NVidia's code for adding support for AA is running on our hardware all the time - even though we're not being allowed to run the resolve code!
So… They've not just tied a very ordinary implementation of AA to their h/w, but they've done it in a way which ends up slowing our hardware down (because we're forced to write useless depth values to alpha most of the time...)!"

http://www.brightsideofnews.com/news....aspx?pageid=1

So Batman performs less than optimally on AMD cards because the 'nVidia IP' MSAA implementation (which is perfectly standard, and is how both companies recommend it be implemented) is half-performed on AMD card- they are just locked out of the render path to actually apply it.

The nVidia implementation of MSAA in Batman Arkham Asylum forces AMD cards to perform the first part of the calculations, which are never used and the result never shown (MSAA).

Wouldn't it have been better to stop the code execution before the first path? Why position the vendor lock out after beginning code execution of implementing MSAA?

I'm willing to concede there might be a valid technical reason that all cards, no matter of final render path, need to execute the nVidia MSAA alpha operations.... but that's really suspicious.

[makomk]

Originally Posted by SIrPauly

If it was so trivial and easy -- just write the code again and test it on ATI products and be done with it. Spend the little resources for the in-game feature even though it is a silly way of spending them.

Except that ATI only found out they were expected to do this after the game was released, by which point it was kinda late. Until then, Eidos strung them along with a bunch of promises that were never met. Have you read the BSN report? Aside from anything else, it makes it clear that the NVidia statement was a bunch of lying lies - contrary to NVidia's statement, ATI was quite happy to QA the code in question during the development process and contribute back, except for the small issue of it refusing to run on their hardware.

[SIrPauly]

Originally Posted by Corum Jhaelen Irsei

Ati already did. They offered it and were refused. How many times has that been said in this and other threads so far? You are being obtuse on the issue.

The last part really pisses me off about your post. You have been reading this, you know that ATi did solve the problem, then had their solution thrown back in their face. What do you do than, when the bully owns the teachers?

I agree with Kombatant:


http://www.rage3d.com/board/showpost...7&postcount=91

[SIrPauly]

Originally Posted by makomk

Except that ATI only found out they were expected to do this after the game was released, by which point it was kinda late. Until then, Eidos strung them along with a bunch of promises that were never met. Have you read the BSN report? Aside from anything else, it makes it clear that the NVidia statement was a bunch of lying lies - contrary to NVidia's statement, ATI was quite happy to QA the code in question during the development process and contribute back, except for the small issue of it refusing to run on their hardware.

I read it -- and they found out in August. When did Eidos offer them the same opportunity and do the same thing for in-game as nVidia?

[razor1]

Originally Posted by Corum Jhaelen Irsei

I suspect you are wrong about him being a manager. If he were, he would not be posting like that. I think we can disregard his post, until someone from either MSI or Nv release a press statement, it looks like Charlie is right. Hell if he works for MSI, he would know if Charlie is right about th VGA stuff.

Hmm he didn't give any information in this specific post so he can post things like that, you see AMD reps do it here and on B3D from time to time, even nV guys too at B3D. Charlie is wrong as usaul.

Originally Posted by Corum Jhaelen Irsei

Ati already did. They offered it and were refused. How many times has that been said in this and other threads so far? You are being obtuse on the issue.

The last part really pisses me off about your post. You have been reading this, you know that ATi did solve the problem, then had their solution thrown back in their face. What do you do than, when the bully owns the teachers?

My understanding ATi wanted the nV code to be changed and they weren't allowed that. ATi could have wrote thier own code for AA, which they didn't want to do. There could be licensing issues in there for the engine/game but those are easy to overcome if ATi programmers are sent to the developer's location.

[makomk]

Originally Posted by SIrPauly

I read it -- and they found out in August. When did Eidos offer them the same opportunity and do the same thing for in-game as nVidia?

They found out about the vendor lock in August and were told it would be removed. They only discovered that wasn't true at around the same time as the rest of us did - and the same with finding out they were suddenly expected to provide their own AA implementation. Reading between the lines, NVidia may have pulled a bit of a con on Eidos.

[Corum Jhaelen Irsei]

Originally Posted by razor1

Hmm he didn't give any information in this specific post so he can post things like that, you see AMD reps do it here and on B3D from time to time, even nV guys too at B3D. Charlie is wrong as usaul.



My understanding ATi wanted the nV code to be changed and they weren't allowed that. ATi could have wrote thier own code for AA, which they didn't want to do. There could be licensing issues in there for the engine/game but those are easy to overcome if ATi programmers are sent to the developer's location.

Your veracity is, as always suspect. How do you know Charlie is wrong? Because some random forum rat said so? That's not going to cut it. We don't need your "understanding" There are facts that are not in dispute, and they don't match your "understanding" ATi did make code, but were not allowed to use it because Edios had a better deal from their "partner" (pay off).

[SIrPauly]

Originally Posted by makomk

They found out about the vendor lock in August and were told it would be removed. They only discovered that wasn't true at around the same time as the rest of us did - and the same with finding out they were suddenly expected to provide their own AA implementation. Reading between the lines, NVidia may have pulled a bit of a con on Eidos.

That data would be welcomed but the Eidos data in the PC perspective piece clearly offers the ATI was offered the same opportunity to add in-game AA and mentioned sending an engineer in house.

This would be kind of insane resources to spend on a nVidia sponsored title considering ATI is probably using their resources to evangelize DirectX 11.

I can understand both parties. nvidia trying to do what is best for their customers and may go too far with leverage and being a bully. It's their investments and code work -- doesn't matter if one agrees or not -- it is what it is.

I can understand ATI's point as well.

Still lean with ATI and would like to see nVidia bend on this one -- as one may like to see ATI more pro-active -- one may like to see these kinds of things go away on something like in-game AA in this example.

Hopefully the two companies can improve on their developer relations moving forward from this episode.

[razor1]

Originally Posted by Corum Jhaelen Irsei

Your veracity is, as always suspect. How do you know Charlie is wrong? Because some random forum rat said so? That's not going to cut it. We don't need your "understanding" There are facts that are not in dispute, and they don't match your "understanding" ATi did make code, but were not allowed to use it because Edios had a better deal from their "partner" (pay off).

How do I know, ask MSI directly, ask CJ, or even better, Ask Charlie for his sources, he has pretty much has none or they are just not capable at what they do at thier jobs, you should see how he gets berated at B3D when he posts there to support his articles, they have very little facts, most of them are based on hypothesis based on poor knowledge of the industry.

Keep reading the emails back and forth nV, did give permission to Eidos, Eidos inturn asked ATi dev rel to send them code, which they haven't. Not only that, the emails were pretty specific what Eidos was able to give to ATi and what they weren't. If you were to lets say create some software and you had ownership of that IP, would you just give it out to anyone who asks? I wouldn't it.

Now on nV's side if you were to create hardware and then software that runs on top of that, why spend your time, your resources to insure the software runs on your hardware, and then test on another competitor's hardware? If you were forced which is what AMD/Ati wants nV to do that, guess what that isn't capitalism anymore. Its closer to Marxism, or Communism because now nV is sharing its wealth to help out a company that isn't capable or doesn't want to do its own work for what ever reason.

This whole problem comes down to $$$$$$$$$$$$$$$$$$$$$$$ who has it who doesn't, who is willing to spend it, who isn't.

[SIrPauly]

There is some discussions about licenses and the need to send an engineer in house at the developers to do the work.

So for ATI to code an in-game AA for this title -- ATI had to do the work in house at the developer to work around the license agreements?

Is this accurate?

It seems to make some sense why it would be easier just to have the lock out pulled away from ATI's point-of-view and just use nVidia code.

[SIrPauly]

http://forum.beyond3d.com/showpost.p...postcount=1116

Originally Posted by MFA

That makes sense, on location help probably gets around the license restrictions ... on the other that just shows how ridiculous the situation is. That's a huge ****ing expense to take people off their regular jobs and fly them over.

[caveman-jim]

posts moved to new thread on b3d - http://forum.beyond3d.com/showpost.p...&postcount=458

[SIrPauly]

I think most of this is semantic:

Eidos has control of the AA -- it is their title and using an engine that is licensed.

Imho,

Eidos wasn't going to offer AA for the PC port, from my understanding, and allowed nVidia to code for their customer base -- this specific in-game AA coding belongs to nVidia -- but Eidos still controls AA, and it may be allowed, meaning AMD would have to code for it themselves --and have to do the same thing nVidia did -- so all the parties may be satisfied.

AMD didn't desire to do it the same way nVidia did -- because the code would be very similar anyway and probably would have to send engineers to the developer because of license agreements -- which in turn would of been a heavier expense and AMD is evangelizing DirectX 11.

So, Amd tried to work around the lock out and try to use nVidia's code - or offer an AA solution using the current code , but the developer will not allow it because it is nVidia's IP - and probably need nVidias go ahead.

[caveman-jim]

@ SirPauly: Your statements indicate you haven't read the BSN article fully.

The AA implementation from nVidia is standard code. This is good, btw. It is the same in design as AMD would offer.

The vendor lock out is not for the whole nV supplied code. There is a performance penalty for ATi cards because the alpha operations code still runs on the ATi card, the resolve path doesn't. So the ATi card does half the work and no result.

The vendor ID might be legally enforcable, but is it the best thing for the game, and for consumers? This is debatable, but my opinion is no.

What are AMD's options here? To provide code thats already included in the game? To provide an in-game option that can be provided by their own management tool? Both seem to be lacking significant returns for the time and money invested, especially as they were asking to address this issue before the game was released (see time line in BSN article).

nVidia could be getting major props from the community by 'magnanimously' allowing their AA implementation code to be used by ATi cards. They are choosing to protect their brand instead. This is detrimental to non-nVidia customers. Why should nVidia care about non-nVidia customers? Because one day they might want them to become nVidia customers, and this might leave a bad taste in their mouth.

Eidos should not have signed whatever agreement they did to include nVidia exclusive functions, if they wanted the best experience for all gamers. Obviously this title was aimed at nVidia customers, once the GPU accelerated only PhysX features were included, why not also lock out AA? It's not like ATi customers are getting the full game anyway, does it matter?

AMD have a genuine beef here with the performanc penalty. For the rest, they are demonstrating nVidia's focus on their gaming customers.

To some this will be seen as being at the expense of all gamers, and potential customers. It's possible that some will only buy nVidia hardware from now on. It's possible that some will refuse to. This polarization is bad for the industry as it creates shcisms and shrinks markets, capping growth for all involved.

nVidia and Eidos have picked a poor road to travel down, and are making decisions based on cross-licensing and revenue bottom lines instead of consumer experience and value. This is a perfectly acceptable business practice, but if you have the choice to use a solution that doesn't restrict your market (especially in Eidos' case) why not go the most open way? Become a leader and driving force and the mind-share and goodwill will repay itself in sales.

[acroig]

Originally Posted by caveman-jim

Eidos should not have signed whatever agreement they did to include nVidia exclusive functions, if they wanted the best experience for all gamers.

nVidia and Eidos have picked a poor road to travel down, and are making decisions based on cross-licensing and revenue bottom lines instead of consumer experience and value. This is a perfectly acceptable business practice, but if you have the choice to use a solution that doesn't restrict your market (especially in Eidos' case) why not go the most open way? Become a leader and driving force and the mind-share and goodwill will repay itself in sales.

Agreed.

[NIGELG]

Originally Posted by caveman-jim

@ SirPauly: Your statements indicate you haven't read the BSN article fully.

The AA implementation from nVidia is standard code. This is good, btw. It is the same in design as AMD would offer.

The vendor lock out is not for the whole nV supplied code. There is a performance penalty for ATi cards because the alpha operations code still runs on the ATi card, the resolve path doesn't. So the ATi card does half the work and no result.

The vendor ID might be legally enforcable, but is it the best thing for the game, and for consumers? This is debatable, but my opinion is no.

What are AMD's options here? To provide code thats already included in the game? To provide an in-game option that can be provided by their own management tool? Both seem to be lacking significant returns for the time and money invested, especially as they were asking to address this issue before the game was released (see time line in BSN article).

nVidia could be getting major props from the community by 'magnanimously' allowing their AA implementation code to be used by ATi cards. They are choosing to protect their brand instead. This is detrimental to non-nVidia customers. Why should nVidia care about non-nVidia customers? Because one day they might want them to become nVidia customers, and this might leave a bad taste in their mouth.

Eidos should not have signed whatever agreement they did to include nVidia exclusive functions, if they wanted the best experience for all gamers. Obviously this title was aimed at nVidia customers, once the GPU accelerated only PhysX features were included, why not also lock out AA? It's not like ATi customers are getting the full game anyway, does it matter?

AMD have a genuine beef here with the performanc penalty. For the rest, they are demonstrating nVidia's focus on their gaming customers.

To some this will be seen as being at the expense of all gamers, and potential customers. It's possible that some will only buy nVidia hardware from now on. It's possible that some will refuse to. This polarization is bad for the industry as it creates shcisms and shrinks markets, capping growth for all involved.

nVidia and Eidos have picked a poor road to travel down, and are making decisions based on cross-licensing and revenue bottom lines instead of consumer experience and value. This is a perfectly acceptable business practice, but if you have the choice to use a solution that doesn't restrict your market (especially in Eidos' case) why not go the most open way? Become a leader and driving force and the mind-share and goodwill will repay itself in sales.

One of the best posts on this subject....especially the line about ATI users becoming Nvidia users some day.I don't know why Nvidia and their fans don't seem to see it this way.Like I said before it's better to do 'nothing' than to do something wrong or something that can turn off potential customers.

[SIrPauly]

Originally Posted by caveman-jim

@ SirPauly: Your statements indicate you haven't read the BSN article fully.

The AA implementation from nVidia is standard code. This is good, btw. It is the same in design as AMD would offer.

The vendor lock out is not for the whole nV supplied code. There is a performance penalty for ATi cards because the alpha operations code still runs on the ATi card, the resolve path doesn't. So the ATi card does half the work and no result.

The vendor ID might be legally enforcable, but is it the best thing for the game, and for consumers? This is debatable, but my opinion is no.

What are AMD's options here? To provide code thats already included in the game? To provide an in-game option that can be provided by their own management tool? Both seem to be lacking significant returns for the time and money invested, especially as they were asking to address this issue before the game was released (see time line in BSN article).

nVidia could be getting major props from the community by 'magnanimously' allowing their AA implementation code to be used by ATi cards. They are choosing to protect their brand instead. This is detrimental to non-nVidia customers. Why should nVidia care about non-nVidia customers? Because one day they might want them to become nVidia customers, and this might leave a bad taste in their mouth.

Eidos should not have signed whatever agreement they did to include nVidia exclusive functions, if they wanted the best experience for all gamers. Obviously this title was aimed at nVidia customers, once the GPU accelerated only PhysX features were included, why not also lock out AA? It's not like ATi customers are getting the full game anyway, does it matter?

AMD have a genuine beef here with the performanc penalty. For the rest, they are demonstrating nVidia's focus on their gaming customers.

To some this will be seen as being at the expense of all gamers, and potential customers. It's possible that some will only buy nVidia hardware from now on. It's possible that some will refuse to. This polarization is bad for the industry as it creates shcisms and shrinks markets, capping growth for all involved.

nVidia and Eidos have picked a poor road to travel down, and are making decisions based on cross-licensing and revenue bottom lines instead of consumer experience and value. This is a perfectly acceptable business practice, but if you have the choice to use a solution that doesn't restrict your market (especially in Eidos' case) why not go the most open way? Become a leader and driving force and the mind-share and goodwill will repay itself in sales.

Does ATI get any blame at all in your mind?

[caveman-jim]

Originally Posted by SIrPauly

Does ATI get any blame at all in your mind?

absolutely. In general terms ATI's dev rel is not as good as it needs to be; they need to invest more engineer time and money in reaching out to the community (people like Ray Adams, projects like F@H) and in engaging with software developers (providing solutions for including features like Physics via DirectCompute or OpenCL) as well as cross promoting.

IMO they are edging in the right direction but very slowly, and I understand that they are resource constrained, too.

In this particular instance, I think it was never going to be something ATi got involved in - Eidos used UE3 and that comes with an inherent slant towards nVidia (physX support).

In general terms, there's lots more AMD can be accused of not doing. In this specific case it's hard to see where the foul is on their part - they were locked out without prior negotiation and put at a performance penalty for arbitrary reasons.

[bobvodka]

Originally Posted by caveman-jim

Eidos used UE3 and that comes with an inherent slant towards nVidia (physX support).

And AMD need to get on the ball a bit more with UE3 and working with Epic simply because UE3 just went "free" for anyone to use for games (well, it has costs later depending on usage and sales but still), something which has peaked the intrest of ALOT of developers (myself included who is current a HD5870 owner) and will probably lead to a rise in UE powered games.

[caveman-jim]

Originally Posted by bobvodka

And AMD need to get on the ball a bit more with UE3 and working with Epic simply because UE3 just went "free" for anyone to use for games (well, it has costs later depending on usage and sales but still), something which has peaked the intrest of ALOT of developers (myself included who is current a HD5870 owner) and will probably lead to a rise in UE powered games.

It's not exactly free:

Licensing


Licensing FAQs

Purchase a Commercial License
Type of License:

$99 Royalty bearing
$2,500 Per seat

Coming Soon!
UDK Commercial Terms

If you use UDK in your business, sell a game or application created using UDK, sell services or training and in connection with that business, distribute an application you've created using UDK, or use UDK in any way to generate revenue, directly or indirectly, in addition to your agreeing to the UDK EULA, you are required to sign a UDK Commercial Use License Agreement. A summary of the current terms of this license are as follows (note that the below terms are only a summary – the actual terms appear in the UDK Amendment):

If you are using UDK internally within your business and the application created using UDK is not distributed to a third party (i.e., someone who is not your employee or subcontractor), you are required to pay Epic an annual license fee of $2,500 (US) per installed UDK developer seat per year. This license fee only applies to UDK seats used for development; no license fee is required for hardware where only the resulting applications are installed.

If you are creating a game or commercial application using UDK for sale or distribution to an end-user or client, or if you are providing services in connection with a game or application, the per-seat option does not apply. Instead the license terms for this arrangement are $0 (zero) up-front, and a 0% royalty on you or your company's first $5,000 (US) in UDK related revenue, and a 25% royalty on UDK related revenue above $5,000 (US). UDK related revenue includes, but is not limited to, monies earned from: sales, services, training, advertisements, sponsorships, endorsements, memberships, subscription fees, rentals and pay-to-play.

http://www.udk.com/licensing.html

But yes, there is an opportunity here.

[bobvodka]

Originally Posted by caveman-jim

It's not exactly free:

Well, yes, I just didn't feel like getting into the details