Reality is that they have blocked it on their all in wonder parts if you have a nV GPU in your system. Don't take my word for it, ask them. They have done exactly what the extreme loyalists are getting upset about here. Do I bash them for it? Of course not, the additional workload that must place on the driver team has to be beyond the realm of reasonable for such a fringe market, particularly considering that it would be counterproductive in helping them gain anythinig. That is a realistic business decission. Those that require drool buckets may see things differently, much like AMD pulling the chipset license from nVidia. It was a smart business decission on their end, despite the fact that nV gave them their best shot at Intel during the early Athlon days when the original nForce provided a superior overall platform to the Intel alternatives, at the end of the day their is no loyalty in the business world and they made the choice that best protects their interests- that is more of an issue that is directly comparable to not supporting Crossfire, ATi breaking the all in wonder parts when using nV GPUs is what is directly comparable to this situation.
Reality is that they have blocked it on their all in wonder parts if you have a nV GPU in your system. Don't take my word for it, ask them. They have done exactly what the extreme loyalists are getting upset about here. Do I bash them for it? Of course not, the additional workload that must place on the driver team has to be beyond the realm of reasonable for such a fringe market, particularly considering that it would be counterproductive in helping them gain anythinig. That is a realistic business decission. Those that require drool buckets may see things differently, much like AMD pulling the chipset license from nVidia. It was a smart business decission on their end, despite the fact that nV gave them their best shot at Intel during the early Athlon days when the original nForce provided a superior overall platform to the Intel alternatives, at the end of the day their is no loyalty in the business world and they made the choice that best protects their interests- that is more of an issue that is directly comparable to not supporting Crossfire, ATi breaking the all in wonder parts when using nV GPUs is what is directly comparable to this situation.
I would like proof of both of those things.
Particularly AMD pulling the license. AFAIK, AMD can't pull Nvidia's license, nvidia is a member of the hypertransport consortium and has just as much right to build technologies on it as AMD. It always seemed to me what happened was nvidia diverted resources to try to shore up the now superior intel platform, and at the same time AMD bought ATI and flooded the market with low cost chipsets, whereas nvidia focused on high margin parts.
No problem, just as soon as someone posts proof that nV is breaking PhysX nV/ATi tandems on purpose and they are breaking CF support intentionally on nForce parts.
AMD didn't pull the chipset license from nVidia:
http://www.pcper.com/article.p...=734&type=expert&pid=4
They just sort of stopped making chipsets for AMD since Intel is much more popular now.
Straight from the horse's mouth (Mr. Huang):
"We used to build chip sets completely top to bottom for every AMD microprocessor. When AMD bought ATI, it became a less logical decision to make. And so we decided that we would change our strategy from investing on AMD chip sets top to bottom to focusing on the Intel platform. "
As for the Crossfire disabling...if they won't even allow SLI on anything but their own chipsets (changed with the X58 since they sort of had to and we were able to see that it didn't require anything special)...why would they allow XFire on their chipsets? It's pretty obvious it's blocked by them after seeing what happened with SLI on X58.
akugami
Diamond Member
- Feb 14, 2005
- 6,210
- 2,552
- 136
Originally posted by: Qbah
Originally posted by: Keysplayr
Hehe, I see Intels business practices are cool with you. Nice i7 you have there.:evil:
What has Intel got to do with a topic about PhysX? You can run SLI on Intel and CF too - if anything, nVidia should learn from them! This was such blatant trolling and bad try to rebound it's sad.
I think it's more with Elfear's comments on the business practices of nVidia. Though to be fair, Elfear did say that everything else being roughly equal between ATI & nVidia he'd go nVidia. Elfear didn't say, "OMFG, burn nVidia! I'm going ATI only cause your business practices suck rocks!" I'd take that to mean that if nVidia provided a superior product (i7 vs whatever AMD has now for example) he'd go with that.
Some of us remember when you could buy any motherboard and drop in a CPU from a few different vendors. We all remember how expensive CPU prices got after Intel locked out the competition. Excepting AMD, where are the competition now? Regardless of industry, lock-ins rarely (if ever) benefit consumers. Consumers are the ones that lose in most of these cases.
Well, nVidia gets a larger potential market for PhysX as I've posted previously. This would make good business sense. Make your products more enticing and customers are more likely to buy it. Simple concept.
In the first scenario, we have a person who bought an ATI card. He'd like to enable PhysX but he can't cause nVidia doesn't allow it. In this case, nVidia makes no money. nVidia loses.
In the second scenario, we have an ATI card along with an nVidia card for PhysX. nVidia makes money, probably not as much as if the person bought an nVidia card as the primary GPU but they still made something and it reinforces PhysX as a standard. Not the best case scenario but at least they still made money.
In the third scenario, we have an nVidia + nVidia combo. nVidia just plain wins in this case.
Obviously the 3rd scenario is what nVidia would love but if I was a business, the second scenario sure beats out the first. And the second scenario is what they'd get.
There's two views on this. You can lock your customers in early and set it so that anyone who wishes your products or technologies must use them on your terms. The potential pitfall is the market finds your terms restrictive and decide not to adopt your products. You lose. An example of it working is the iPhone, but the iPhone built on the success of the wildly popular iPod. An example of it not working is Mac computers.
Alternatively you can try to get as much people to use your products/tech as possible before you find ways to lock them in. Example is iPods after they released a version with connectivity to a PC. Yes, the iPods have iTunes lock-in but that wasn't a huge issue with the iPod's uptake. You can get music from pretty much anywhere into an iPod. Apple has now used the iPod/iTunes dominance as a vehicle to their other products.
I disagree. If one was to say nVidia has strong leverage on GPGPU then I wouldn't be able to argue. ATI's GPGPU offerings have been anemic compared to nVidia's. However, in the physics acceleration space, everything is in such an early state that things can go either way. Some of the Havok tech demos (for what it's worth) have showed great potential, certainly no less than that of PhysX.
Good games typically take 2-3 years to develop. That means that it could potentially be another 3-5 years before we see a real winner start to emerge in the Havok vs PhysX wars. Yes, I know that games using PhysX are being developed even now but I think for physics acceleration to truly shine it'll take a grounds up approach where a game is specifically made with physics acceleration in mind and not just as an added feature for some added fluff as in current games.
The whole X-factor in all this is how much resources Intel will bring to push Havok. After all, nVidia has provided resources to developers to help push their own technology (The Way It's Meant To Be Played program) as have ATI (Get in the Game). Intel has a vested interest in pushing Larrabee and Havok. While Intel has a really tough battle ahead, they do have a lot of money on their side.
These are poor analogies as none worked with the other to begin with. This is a case were functionality has been taken away from the customers. Its as if Seagate said b"From now on our HDs will no longer function in a RAID setup if they detect a non Seagate drive in the array."
The funny part of this whole thing is this is supposed to be a shot at ATI. But its not, its a shot at consumers, specifically consumers who paid for and own nVidia video cards.
AyashiKaibutsu
Diamond Member
- Jan 24, 2004
- 9,306
- 4
- 81
I think this is being blown out of proportion... Maybe just maybe some new stuff in the phyx API breaks with an ATI card. It would be a huge added cost to nvidia to try to support physx in such a setup. Hired goons sitting at a telephone cubical 24/7 aren't cheap. The guy that said he was using ati/nvidia said win7 worked but some tech demos didn't work right (presumable new games using the effects the demos did wouldn't work right either). Because all you want in a 3d/physics setup is to run windows amirite? Dealing with nontech savy people trying to deal with mixed cards certainly has the potential to cause massive headaches.
Right, those analogies were piss poor. I know that I'd be pissed if I spent ~$100 for a card to run Physx than Nvidia decided to pull that functionality on me. The pro-Nvidia crowd can try and color this however they want, but the bottom line is they hurt some of their customers with this one. The good news is I doubt there is a very large population of users with AMD cards and an Nvidia part for Physx. But those who do have that set up are getting screwed now.
As others pointed out there is a difference between not supporting and disabling.
I its not like hard drive manufactures have to sit on the phone and guide people through setting up complicated network drives or raid setups and figure 'ah we better just disable this'.
Also If nV really wanted to push physX, then I would think they should try to put it in the chip set (nforce). Might not be possible, but would greatly increase the install base.
Keysplayr
Elite Member
- Jan 16, 2003
- 21,219
- 55
- 91
You feel Nvidia should not have disabled something that was never intended to work or be supported in the first place. Am I right? The fact that it worked at all, and not in every situation mind you, was probably sheer luck. Happenstance. Never intended to work nor was it ever advertised to work with an ATI card as a primary in the system. It's not how it was designed or intended to work.
SirPauly
Diamond Member
- Apr 28, 2009
- 5,187
- 1
- 0
Not recommended/not supported is different than something that did work (even if not suported) then is turned off by a later driver.
How about this secario? You buy a lower end Radeon 4x00 card for it's 7.1 surrond support and HD accerleration to use in your Intel rig that has integrated graphics. After you buy the card and have been been using it for a while, it's been working just fine, on your Intel i7 rig AMD turns off those features because they say you need an AMD chipset for those features. I would imagine people wouldn't be too happy with AMD.
To me this is pretty similar, people used this set up and paid money for an Nvidia card to run Physx, now Nvidia is removing that ability if you have a non-Nvidia card coupled with the Nvidia card (though it DID work before, even if not supported). The good news for Nvidia from a PR stand point is my guess that the amount of people that are using that set up is very small.
SirPauly
Diamond Member
- Apr 28, 2009
- 5,187
- 1
- 0
Working doesn't mean working and stable in all the content intended -- even future PhysX content that may be much more complex than it is now. Some may need to see what nVidia's data is first before jumping to conclusions that this is just for leverage to lock out intentionally ATI -- end of discussion -- that's it. There may be other reasons as well that some dismiss. It's not coloring and making excuses; it is a desire to know more data; to understand more with clarity.
One may not rule out "leverage" becase the data is not too clear and one must trust nVidia on the surface because does anyone know what nVidia means by the technical aspects here?
Right now I am leaning to a cost/benefit decision more so than just locking out but may change my mind. Has nothing to do with coloring anything.
And it's sheer luck, happenstance, that memory made by different companies works together, hard drives, etc... I can even run printers from different companies on my pc at the same time. It is absolutely absurd to even try to defend Nvidia on this one. They clearly put code in on purpose to disable their card from working when it detected other hardware in the PC.
Nvidia might as well put an asterisk on the box of all of their products: *We may disable PhysX from working on your system at our leisure without warning.
PhysX working on a Geforce card is pure luck? Comeon now you make it seem as if it's running on the ATI card. I know what you're trying to say but the fact is, is that it's not anything extraordinary that it was running in the first place...since PhysX IS designed to run on nV cards. The Ageia cards worked fine regardless of the primary GPU so it's hard to believe the same functions couldn't be pulled off in this case. This is plain and simple a "business" move...whether it's actually good for business is debatable.
I can often see where you are coming from with your post, but on this one you are just plain wrong, it is a move that does nothing but harm legitimate consumers.
These people paid for and bought a nVidia GPU, knowing it can do physX as a secondary card.
nV disabled that feature for for business/selfish reasons only. (or there would have been some kind of note as to why).
Whether it is nVidias *responsibility* to make it work if ATI drivers are installed is debatable, but nVidia needs to put on their big boy pants on this one and let it work if it works.
Whats next, PhysX only works on certain monitors, or with Intel CPUs?
When physX is working as a secondary card, what card is rendering the graphics is as relevant information as what CPU is in the computer.
Again this is just proving the point that physX is not a feature worth basing any decisions on what card to buy on.
VirtualLarry
No Lifer
- Aug 25, 2001
- 56,587
- 10,225
- 126
That would be a hoot. Perhaps NVidia will come out with "game acceleration" NIC drivers for mobos with NV onboard NICs (using the same principles claimed by the Killer NIC, bypassing driver layers). And then they will tie them in to having NV video cards, chipsets, etc.
Such that if you installed a third-party NIC, you couldn't even play some top-tier games that used NV features.
I can see this happening. Games writting for the "NV gaming platform" on PC, that do a driver/system-hardware scan, and only play if you have the requisite NV drivers present on a PC. If NV had enough market share of those devices, you know that they would do it, transforming TWIMTBP into a real, serious, vendor lock-in.
VirtualLarry
No Lifer
- Aug 25, 2001
- 56,587
- 10,225
- 126
VirtualLarry
No Lifer
- Aug 25, 2001
- 56,587
- 10,225
- 126
And it's Intel's fault that you cannot. Back when bus interfaces were nearly open standards, back in the Socket7 days, you could get CPUs from Intel, Cyrix, NatSemi, Via, UMC, and AMD that worked with Socket7 mobos, using Intel chipsets.
What do you think QPI on an i7 is for? It's a re-worked PCI-E express design, but patented by Intel explictly such that there is vendor lock-in.
When on the other hand, AMD used HyperTransport, which is an open standard. Which is why they have chipsets from other companies that support their CPUs.
The market loses when there is vendor lock-in. Customers prefer choice. Choice begets competition, which promotes efficiency and lower prices and more features and technological progress.
That doesn't make it right nor desireable.
VirtualLarry
No Lifer
- Aug 25, 2001
- 56,587
- 10,225
- 126
Since when did AMD shut NV out of the chipset business? Proof?
My understanding was that NV's chipsets for AMD's CPUs were simply sub-par for the market, and failed, and they pulled out of the business.
Since HyperTransport is an open standard, I'm not even certain where AMD would have the legal means to even shut NV out of that market, unlike Intel with it's proprietary, patented, system busses.
VirtualLarry
No Lifer
- Aug 25, 2001
- 56,587
- 10,225
- 126
Are we talking about the original ATI All-In-Wonders, the AGP ones that barely worked right in a system even with the proper drivers and no additional video cards in the system, or are you talking about the more recently re-released series of All-In-Windows, that have a PCI-E interface? Are you saying that the tuner functionality is disabled if you have an NV card in the system?
VirtualLarry
No Lifer
- Aug 25, 2001
- 56,587
- 10,225
- 126
Uhm, Keys, just a little computer-science tip for you.
Nothing within a computer system happens due to "luck". It's a deterministic system. If something "works", then it was engineered that way.
Clearly, someone in marketing felt that the engineering teams were in error, and they reversed course.
Isn't that what you said? That Physx working (on an Nvidia GPU, since that's what it runs on) was 'probably sheer luck' when used in a system with an AMD GPU. That's how I took what you said.
Obviously it wasn't luck, it worked, than Nvidia decided to pull the rug out. Good thing they care about their customers who spent money on an Nvidia GPU to run Physx. :laugh:
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products- Started by poke01
- Replies: 411
Facebook
Twitter