NVIDIA Responds To AMD's Free Sync Demo

cyclone3d said:
And then when somebody came up with a way to run PhysX on an AMD card, Nvidia immediately shut them up.
Click to expand...
That's one view of what happened. The other is that Regeneration is simply a liar and never had PhysX running on AMD GPUs. Most people (who care about it) seem to have accepted the latter option, including <banned_site>'s own posters. Regen never showed anyone his work, despite promising several release dates to a couple of people, or even showed one shred of evidence he had it working. It was simply his word, a string of broken promises, then a claim he was shut down.

There are several reasons why he likely didn't get it running, particularly due to the significant architecture differences between NV's CUDA-enabled GPUs (GPU PhysX requires a CUDA compatible GPU) and AMD's GPUs at the time. Specifically, the code running on NV's GPUs didn't have a hell's chance of working on AMD GPUs and translation would have been a massive undertaking (turning scratchpad heavy code to register-based code and emulating other video driver services is not a trivial undertaking).

I'm pretty sure we've gone over all this before.
 
Activate: AMD said:
This is in complete contrast to CUDA, G-Sync and PhysX, all of which are explicitly proprietary. AMD couldn't implement even if they wanted to because nV isn't coughing up the goods.
Click to expand...
lol

So Mantle is not open (yet, but may be some time in the future), and I could also list a dozen other proprietary AMD GPU initiatives if you'd like, and that is somehow different from Nvidia having proprietary technologies too. Okay...

Seriously, GPU fans are just plain weird and way too wrapped up emotionally to even think straight.
 
Activate: AMD said:
AMD have explicitly said that its not tied to GCN
Click to expand...

AMD's official statements and website say you are wrong. It has been assumed it will work on other GPU's but I have yet to read an AMD released statement saying it is not designed for GCN.

AMD Mantle Site said:
Mantle is a software component that makes it easy to apply programming techniques and optimizations written for consoles to PC’s containing GCN architecture.
Click to expand...

AMD 10/4/2013 Press Release said:
Games enabled with Mantle speak the language of the Graphics Core Next architecture to unlock unprecedented performance and graphics quality.
Click to expand...

AMD R7 & R9 pages said:
Mantle:
There’s optimization, and then there’s Mantle. Games enabled with Mantle speak the language of GCN architecture to unlock revolutionary performance and image quality.
Click to expand...

And I already pointed out on the Mantle page where AMD stated Mantle requires GCN.

SO again if I missed it please let me know where AMD released a statement saying Mantle does not require GCN. All info provided on their site and in official statements say Mantle does require GCN.


Activate: AMD said:
So no, its not "completely proprietary" as long as the software standard is open. It may be up to the hardware vendors to implement in drivers and hardware, but thats not the same thing as "proprietary".
Click to expand...

Again, works only on AMD tech for the foreseeable future than it is proprietary.

If we use your logic then PhysX is not "completely proprietary" since NV once offered it to AMD who said no or G-Sync is not "completely proprietary" since NV COULD one day license the tech to AMD.

Like I said earlier, I really interested to read where AMD has officially offered the tech to any other GPU maker or that Mantle doesn't need GCN. I think Mantle is an awesome idea but I don't see AMD allowing green to have it.

After all Mantle's catch phrase is "Powered by AMD" and not "Powered by ALL!"
 
pxc said:
lol

So Mantle is not open (yet, but may be some time in the future), and I could also list a dozen other proprietary AMD GPU initiatives if you'd like, and that is somehow different from Nvidia having proprietary technologies too. Okay...

Seriously, GPU fans are just plain weird and way too wrapped up emotionally to even think straight.
Click to expand...

^this
 
I think this is more about repackaging something that's already an open standard as proprietary tech that will only work with proprietary hardware and charging a premium.
 
collegeboy69us said:
oh look -- nvidia being a whiney greedy piece of crap again, what's new?

they can keep their hardware and their software locked down tight as they want, it won't ever make me want to change where I spend my money.

We have all seen what stingy and proprietary does in the electronics industry. AMD is making cool shit, and then saying "here ya go, run with it, figure it out and make it work for your application" Nvidia is basically saying "fuck you pay me"
Click to expand...

Apparently you believe R&D is free and when a company spends millions developing something it should be public domain.

Neither AMD nor NVIDIA are running a charity. Both will attempt to compete through exclusive features, just likenp pretty much any consumer electronics company. Get off the delusional high horse.
 
LeninGHOLA said:
I think this is more about repackaging something that's already an open standard as proprietary tech that will only work with proprietary hardware and charging a premium.
Click to expand...

You speak the truth, Ulyanov.
 
pxc said:
That's one view of what happened. The other is that Regeneration is simply a liar and never had PhysX running on AMD GPUs. Most people (who care about it) seem to have accepted the latter option, including <banned_site>'s own posters. Regen never showed anyone his work, despite promising several release dates to a couple of people, or even showed one shred of evidence he had it working. It was simply his word, a string of broken promises, then a claim he was shut down.

There are several reasons why he likely didn't get it running, particularly due to the significant architecture differences between NV's CUDA-enabled GPUs (GPU PhysX requires a CUDA compatible GPU) and AMD's GPUs at the time. Specifically, the code running on NV's GPUs didn't have a hell's chance of working on AMD GPUs and translation would have been a massive undertaking (turning scratchpad heavy code to register-based code and emulating other video driver services is not a trivial undertaking).

I'm pretty sure we've gone over all this before.
Click to expand...

Whether that story is true or not, it's a undisputable fact that plenty of people have publicly exhibited working examples of using an NVidia card as a dedicated PhysX card when their primary card is an AMD card. And it's undisputable that NVidia did (does) everything it can to nip it in the bud including lawyering up.
 
inglewood78 said:
If you invented something that is unique and could make you money, would you give it away for the good of the people or would you patent it and make money off of it? What's the difference with this? Lots of hypocrisy in this thread.

Now you if want to argue that the tech is a pos and not worth the cost, well, that's a completely different topic.
Click to expand...

I don't think anyone is saying that Nvidia shouldn't be able to patent the g-sync module and make money off selling it, but instead are saying that they shouldn't be able to block other companies from using that module.

Big difference.
 
Cr4ckm0nk3y said:
AMD's official statements and website say you are wrong. It has been assumed it will work on other GPU's but I have yet to read an AMD released statement saying it is not designed for GCN.
Click to expand...

And I already pointed out on the Mantle page where AMD stated Mantle requires GCN.

SO again if I missed it please let me know where AMD released a statement saying Mantle does not require GCN. All info provided on their site and in official statements say Mantle does require GCN. [/quote]
multivendor.jpg

http://techreport.com/news/25651/ma...ite-games-dice-calls-for-multi-vendor-support


Cr4ckm0nk3y said:
Again, works only on AMD tech for the foreseeable future than it is proprietary.
Click to expand...
Revisit the definition of "Proprietary" and rethink that statement. If AMD is allowing anyone to implement it (aka if nvidia came up with a way to shoehorn it into their GPU's) then that isn't proprietary. The difference is that if AMD made PhysX or CUDA run on their hardware, nVidia would be banging on their door with a Cease and Desist

Cr4ckm0nk3y said:
If we use your logic then PhysX is not "completely proprietary" since NV once offered it to AMD who said no or G-Sync is not "completely proprietary" since NV COULD one day license the tech to AMD.
Click to expand...
:rolleyes: You really don't understand the word "proprietary" do you? Offering to license is not the same as publishing the standard and saying "go do it"
 
Spazturtle said:
USB was designed to replace around 7 proprietary Serial Bus connectors.

If was bloody horrible before USB.

Their is a good reason why people hate proprietary stuff.
Click to expand...

Um what?

EIA-232 and EIA-530 are not closed standards, and this is what we used before USB when referring to the common DB9 / DB25 ports on the back of computers.

It was horrible due to the way these ancient communications standards worked.
 
Activate: AMD said:
multivendor.jpg

http://techreport.com/news/25651/ma...ite-games-dice-calls-for-multi-vendor-support
Click to expand...

Correct me if I am wrong but those slides are from DICE at APU 13 and Johan Andersson's thoughts on Mantle. That is not AMD.

Techreport Article said:
He also brought up other Frostbite games that will support the API, and he shared his own wish list for Mantle's future.
Click to expand...

Again your backup is a wishlist from DICE? The article is even titled "DICE calls for multi-vendor support." Maybe if it was called "AMD announces multi-vendor support" and the slides were AMD's you could/should have linked it.

Techreport Article said:
But the "pink elephant in the room," as he called it, is multi-vendor support. Andersson made it clear that, while it only supports GCN-based GPUs right now, Mantle provides enough abstraction to support other hardware—i.e. future AMD GPUs and competing offerings.
Click to expand...

This was Johan Andersson's thoughts and wish list for Mantle's future. Not an official AMD statement.

Show me where AMD has said it doesn't require GCN or can be used by any GPU.

The last paragraph of the article sums it up best

Techreport Article said:
There's no telling yet whether Mantle will ever become a cross-vendor, cross-platform standard, or whether the future holds something different, such as a competing Nvidia API or a future version of DirectX with some of the same perks.
Click to expand...


Activate: AMD said:
Revisit the definition of "Proprietary" and rethink that statement. If AMD is allowing anyone to implement it (aka if nvidia came up with a way to shoehorn it into their GPU's) then that isn't proprietary. The difference is that if AMD made PhysX or CUDA run on their hardware, nVidia would be banging on their door with a Cease and Desist
Click to expand...

Again show me where AMD has offered it to NV. That is all I am asking you.

ATI had an opportunity to use PhysX and said no. Right after independent Radeon devs found a way to use PhysX and CUDA on AMD hardware, with NV's support, and AMD refused to help.

http://www.techpowerup.com/64787/ra...ffered-to-help-us-expected-more-from-amd.html

After ATI's negative responses to PhysX, GPU PhysX was locked down to NV only. Believe it was 2009.

Activate: AMD said:
:rolleyes: You really don't understand the word "proprietary" do you? Offering to license is not the same as publishing the standard and saying "go do it"
Click to expand...

I think you are confused on proprietary. And now Mantle is a published standard that anyone can use? Show me one presser or one example of this "standard" being used by NV or Intel. NV can't "go do it" because Mantle as of now requires GCN to operate.

Does the API have abilities to allow other GPUs? Maybe according to one developer but has it officially been opened up to other GPU's? Not that I have seen.

So again, how is Mantle not proprietary? Mantle only works on AMD hardware and is still exclusively owned by AMD.
 
Last I checked nvidia wasn't a charity. If they put the work in to create a product they have every right to sell it for whatever they want. The market will either accept it or reject it. Simple as that. AMD isn't entitled to anything and neither is the community to put it bluntly. You want it, then buy it. If not, keep walking.
 
Zarathustra[H];1040529408 said:
Pretty similar to what they did with PhysX, and CUDA, both things that later were able to be done through OpenCL, but now we have fragmented proprietary standards instead.
Click to expand...

I think there's still a lot of confusion out there regarding what CUDA and OpenCL are.

OpenCL code is nothing more than CUDA kernels with a little bit of extra code for setting up the device. So if you have CUDA code, you're basically 1-2 days away from having a working OpenCL solution.

Code written for CUDA isn't proprietary or anything like that. It all comes down to the what the author wants to do with it.

It is true that CUDA-compiled code can not run on other GPU architectures or with OpenCL libraries.
 
everyones amd this , nvidia that. The real culprit are the monitor makers sitting on their asses.
amd should become a monitor maker now or force them to implement dp1.3?
the day we start having monitors tied to gpus is the day we all lose.

sorry your monitor is not supported on X brand buy Y brand instead. :rolleyes:
 
The issue isn't lack of dp1.3, it's that few monitor/panel makers are implementing the VESA standard for controlling VBLANK intervals. It's not a requirement and there has been little demand since while vendors like AMD have had it in hardware for multiple generations, it's not been enabled. The reason is likely due to the (software) issues Nvidia worked through to create G-Sync, which apparently requires more work than just showing a carefully controlled demo on a specially configured laptop in order to be useful in actual real world games. :p
 
Tych-0 said:
So armed with a little more knowledge now, am I at all right in saying that nVidia has basically released a Display Port 1.3 function a head of time and put a brand name (G-Sync) to it? That's really what I'm walking away with here.
Click to expand...
Pretty much, also the features used in free sync aren't part of DP 1.3. AMD hopes it will be, and it still needs the proper controller card to work in the monitor which is not part of VESA standards so even if it's in DP 1.3 it still needs a special add in card(not unlike g sync) to work. Basically AMD hopes it will be in both standards but has no guarantee, nvidia said fuck it and made their own. AMD's free sync wont hit the market unless those standards become reality, ie 1-2+ years down the roads, Nvidia's stuff will hit the real market in a short time.

Nvidia has a real product, AMD basically has a tech demo that rips off G-sync for PR purposes why else name it free sync.
 
Terpfen said:
It's hard to say what it does and does not support; 1.3 won't be finalized until the second half of 2014.

Not sure why people are getting aggravated over Nvidia's statement. They invested R&D time and money and deserve a return on their investment. This is how improvements are made. Notice that AMD did nothing about graphical tearing until Nvidia did. If it's as easy as FreeSync, why wait until after Nvidia announced G-Sync? Because lack of competition.

Anyway, let Nvidia and AMD fight. The end result is improved technology for us.
Click to expand...

No, no they don't. They deserve exactly what people choose to give them, people lose money on investments all the time. Nvidia isn't some special flower that deserves gumdrops and ponies because it spent some cash designing proprietary hardware.

It's in most people's best interests that G-Sync fails and Freesync wins. The alternatives are either the technology as a whole failing or people being forced to pay a Nvidia tax when they buy a monitor because it includes something they can't use. Neither of those outcomes would be good.
 
Seriously, fans of anything or anyone are just plain weird and way too wrapped up emotionally to even think straight.

opps :)
 
Forgetting all the heated debate about what is proprietary and whether Mantle is open and whether green or red has the bigger e-penis, the real point is that I cannot imagine anyone being happy about having a specific monitor attached to a GPU in order to enjoy functionality. That is much different than a GPU vendor offering a platform or piece of hardware that allows programmers the choice to implement.

For example, even PhysX doesn't bother me. If you want "that" technology, buy the Nvidia card. Your choice. But proprietary integration of hardware is another matter altogether. That smacks of a closed eco system a la Apple. No thanks!
 
Terpfen said:
And also does not solve the fundamental problems addressed by G-Sync.
Click to expand...

Current gsync monitors don't solve the problem of not being shitty tn panels.
Which have some serious fundamental problems in my eyes.
 
Spazturtle said:
G-Sync adds $275 to the monitors price, who wants to pay an extra $275 for a feature they cannot use?

So their will either have to keep multiple SKUs or they will lose customers.
Click to expand...

Aside from taking EU prices and converting it to USD (because you would need to forget to take out VAT for such a price difference) where do you see this for American prices?
 
NoOther said:
I think it's funny how many people try to make it seem like AMD is some entirely benevolent organization and Nvidia is some completely evil corporation. They are both for profit companies trying to find more ways to gain an edge over one another. After all, AMD is the company that stole technology to even get started.
Click to expand...

It's part of AMDs marketing strategy on social networks, such as forums. You will see the same handful of people posting misinformation, attacking other members, trying to get people banned and threads locked. Not just here but on pretty much every tech forum. Anandtech forums are pretty much over run with this kind of viral marketing.
 
PRIME1 said:
It's part of AMDs marketing strategy on social networks, such as forums. You will see the same handful of people posting misinformation, attacking other members, trying to get people banned and threads locked. Not just here but on pretty much every tech forum. Anandtech forums are pretty much over run with this kind of viral marketing.
Click to expand...

For every shill AMD has I'm willing to bet Nvidia has 2. These forums have no shortage of them.
 
pxc said:
Seriously, GPU fans are just plain weird and way too wrapped up emotionally to even think straight.
Click to expand...
This x100. The video card subforum is scary.

I don't see a downside here. Nvidia comes up with new tech that requires a more expensive monitor. Are you guys seriously afraid the prices of ALL monitors will go up $100 (or $275, lol) based off a feature that a tiny fraction PCs will use? I think monitor manufacturers will be able to handle another product line, don't lose sleep about it.

And Nvidia isn't making it open platform.. boo hoo. Welcome to real life, companies get paid for their work.

I'm pretty stoked about G-sync. Seems like other people are too considering how AMD guys have their panties in a bunch. Hint: sell your card for 50% more than your bought it for (thanks Litecoin!) and buy a Nvidia GPU. Problem solved.
 
You know AMD is on to something when PRIME1 starts thread crapping...pay attention folks, this technology is going to be a game changer!
 
david_ said:
This x100. The video card subforum is scary.

I don't see a downside here. Nvidia comes up with new tech that requires a more expensive monitor. Are you guys seriously afraid the prices of ALL monitors will go up $100 (or $275, lol) based off a feature that a tiny fraction PCs will use? I think monitor manufacturers will be able to handle another product line, don't lose sleep about it.

And Nvidia isn't making it open platform.. boo hoo. Welcome to real life, companies get paid for their work.

I'm pretty stoked about G-sync. Seems like other people are too considering how AMD guys have their panties in a bunch. Hint: sell your card for 50% more than your bought it for (thanks Litecoin!) and buy a Nvidia GPU. Problem solved.
Click to expand...

Nvidia did not come up with it. Free-sync capability will be part of of display port v1.3.

They just want to charge you for it before its free with new monitors. If you by a G-Sync monitor then you may be stuck with using Nvidia only cards to use the feature.



http://techreport.com/news/25878/nvidia-responds-to-amd-free-sync-demo
 
Get it now or wait until.. who knows. There's always a premium on that.

Article: Desktop monitors use other interfaces, like HDMI and DisplayPort, and typically have a scaler chip situated in the path between the GPU and the panel. As a result, a feature like variable refresh is nearly impossible to implement on a desktop monitor as things now stand.
 
When growing up nvidia was a fan of rambus. They took a play out of rambus playbook.
Join standard. Take stuff from the standard and make it proprietary :p

only difference between rambus and nvidia is that there are ppl willing to jump off bridges for them.
 
Technikal said:
Nvidia did not come up with it. Free-sync capability will be part of of display port v1.3.
Click to expand...

Freesync lacks a scaler ASIC with variable refresh capability. So it will still rely on vsync and it's problems. Also AMD has no plans currently to release this to the public so don't hold your breath.

If AMD had a product they could charge you money for they would. They are not a charity.
 
mdzcpa said:
Forgetting all the heated debate about what is proprietary and whether Mantle is open and whether green or red has the bigger e-penis, the real point is that I cannot imagine anyone being happy about having a specific monitor attached to a GPU in order to enjoy functionality. That is much different than a GPU vendor offering a platform or piece of hardware that allows programmers the choice to implement.

For example, even PhysX doesn't bother me. If you want "that" technology, buy the Nvidia card. Your choice. But proprietary integration of hardware is another matter altogether. That smacks of a closed eco system a la Apple. No thanks!
Click to expand...

I'm sure that if you get a G-Sync monitor it will work fine with AMD cards. And if you want the dynamic refresh rate functionality then you can just use FreeSync that AMD is promising you.
 
pcjunkie said:
You know AMD is on to something when PRIME1 starts thread crapping...pay attention folks, this technology is going to be a game changer!
Click to expand...

And thank-you for not quoting him!!!
 
Dear AMD,
Thanks for pissing in our cornflakes and showing people you can do something for free we were about to bend our customers over for. You screwed us out of millions. We will sick the fanboys on you now...
Sincerely yours, Nvidia
 
You must log in or register to reply here.
Back
Top