Nvidia Walked Away From PS4 Hardware Negotiations
An anonymous reader writes "Tony Tamsai, Nvidia's senior vice president of content and technology, has said that providing hardware for use in the PlayStation 4 was on the table, but they walked away. Having provided chips for use in both the PS3 and the original Xbox, that decision doesn't come without experience. Nvidia didn't want to commit to producing hardware at the cost Sony was willing to pay. They also considered that by accepting a PS4 contract, they wouldn't have the resources to do something else in another sector. In other words, the PS4 is not a lucrative enough platform to consider when high-end graphics cards and the Tegra line of chips hold so much more revenue potential."
You can bet MS has approached them on providing chips for Durango too. I wonder if they told *them* to piss off.
Your political party doesn't care about your rights and only represents corporate interests.
You have to provide lots of parts at low cost and they will surely write in a lower price for each continued year of the console. That means you are tying up fab time on something is on an outdated process a few years down the road.
On the other hand AMD had to do this, they need the money so any margin is likely acceptable.
Sounds to me like they made a pretty good move there. My friend asked me a while ago why NVIDIA wasn't inside the PS4. I told him its probably because they didn't want to get locked into a contract with anyone, they wanted to stay independent.
Just sayin'...
Who needs a
I'm guessing there was a non-competitive clause in there which would have stopped nvidia from selling to Valve for the ValveBox.
The only thing I can take from this is that the potential growth in mobile platforms far outstrips the costs associated with developing hardware for another game console platform. Like a previous comment asked, I wonder if they told Microsoft to go away as well. If they did, what does this mean in the bigger picture? Is the future of gaming on tablets?
I wonder how much of the 'opportunity cost/things we could have been working on instead' factor has to do with the fact that AMD is simply in a tighter spot than Nvidia, and how much it has to do with the fact that AMD already makes CPU/GPU combination packages(and seems interested in making more), while Nvidia has nothing of that sort except their 'Tegra', which might be a snappy mobile part; but is fundamentally punching in a different weight class(if nothing else, Sony's plans for 8GBs of RAM get a lot uglier on a 32-bit architecture. Yes, ARM also has something PAE-like; but PAE is mostly a hack that makes running multiple independent programs on a 32 bit system with more than 4GB of RAM palatable, not something you'd want to design a game engine around.)
This isn't to say that Nvidia couldn't have done it(heck, what would buying VIA cost these days?); but Nvidia would need, essentially, an entire new flavor of product line for this job, while AMD, whether they call it this or not, is punching out a modestly customized APU, which almost certainly shares substantially with the ones that they sell for PCs.
Or they saw they were going to go x86, and most likely AMD because lol Intel, and weren't up to giving their main competitor any money at all.
I guess it will work out for both Sony and Microsoft in the end. They will be the test-bed for their new processor ideas, which sound very interesting.
Why are people running a blatent self-serving PR story?
We lost but... we didn't really want to win it anyway!
They, Allegedly, walked away.
Without video proof, we can't be sure they didn't strolled, strutted or even rambled away.
Considering AMD are producing the CPU chips for both platforms, and the the GPU as well, it isn't surprising that nVidia "walked" away. This is the eventual benefit of AMD buying ATI, in that they can produce both now. I have no doubt that AMD either have special consideration or simply could offer a better bid than nVidia could.
Regardless of the profit, this would be a big feather in AMD's cap. AKA "We produce both the CPU and GPU of all modern game consoles, don't you want to buy our chips?". Also in the bigger scheme of things, if you get game developers in such numbers making games for YOUR video card on millions and millions of consoles for all games, which are ported to say PC, what do you think those games will be optimized for? AMD. Which will look better? AMD. This is something that is going to change things in a pretty large way over the next 10 years.
nVidia should have paid money to be a part of this if only to prevent their rival in AMD from doing so. Perhaps they didn't have the money. More likely they think they have something that will make a difference. I doubt it.
I'm not fired, I quit is the sentiment I feel about nVidia's statement...
Or maybe it's because of the exact reason stated in the summary - Sony didn't want to pay them enough money.
+1 IDisagreeSoHeMustBeATrollOrAnAstroturferOrAShill
> "they wouldn't have the resources to do something else in another sector."
Implying that the resources for a PS4 would be radically different from their PC offerings.
60million units doesn't have revenue potential?
Not only that, the tech they came up with could likely be used for new laptops and set top boxes.
I suspect it was more likely because they didn't have the level of tech needed. ATI had their APU systems lined up already and with tweaking, they're perfect for a console. I'm not sure that NVidia had anything approaching the power of these APUs drawn up (their focus has been on desktop graphics and tablet).
Rumours suggest that the 3DS was going to use NVidia tegra based tech but they couldn't keep the heat down so Nintendo went with the as-seen-in-every-bargain-bucket-chinese-tablet Mali+arm combination.
Sony wanted to install rootkits on nVidia's computers.
Get free satoshi (Bitcoin) and Dogecoins
Forget about AMD GPU optimization. Specific shaders optimization accounts for maybe 10%.
Think about future games being natively written with 8 cores in mind. No more buy 2 Nvidia cards to see some PhysX sparks, all games will use Havoc, Bullet or some other physics library computing on AMD GPUs.
Who logs in to gdm? Not I, said the duck.
The plan to use x86 processors is almost definitely the reason, Nvidia would have to spend a lot of time and effort to create a new architecture to fit what Sony and MS are planning for these new consoles. AMD already has all the technology, rights, and knowledge to do it with minor tweaking.
There seems to be a lot of similarities to the Snapper Lawnmower story.
"Jim Wier believed that Snapper's health -- indeed, its very long-term survival -- required that it not do business with Wal-Mart. "
then maybe the PS4 won't be so expensive?
I guess that's just not the way it's meant to be played.
As he said Nvidia has a lot going on and the decision to develop for the ps4 and next xbox were just to small of a margin in profits compared to what they are currently doing.
It makes sense really. If NVidia currently has their plate full with projects why would they take on a new project with less profits and sacrifice the ones they stand to make more money off of?
Besides ATI may suck at making desktop/laptop gpus, desktop/laptop cpus and such but they have a lot of experience with console gpus.
It wasn't just sales that were dissappointing. The console was dissappointing. It sits in a box almost never used. I loved the PS2, the PS3 was the last straw.
you should mention what OS version, what video card, etc you were running such that their... CRAP... blew up on you.
I've got a Radeon HD4770 and everything since 11.4 or so worked great (Until they dropped the 4000 series cards in 12.6 then added them back in I guess 12.8).
I have OpenCL 1.0 support, and mined bitcoins on that particular hardware for the whole of winter '12-'13 and had no problems. And that was on linux too. The windows drivers were a breeze to install and operate in comparison.
Having all games (and thus their ports) on million and millions of xbox and PS consoles designed and optimized for your specific hardware for the next 10 years is worth money. Any profit they actually get is just icing.
Profit is never "just icing". Profit is the entire purpose for a for-profit company. No (sane) business exists to just make revenue. They have to do more than break even on the job.
60million units doesn't have revenue potential?
Wrong question. The correct question is whether it has profit potential. Revenue is just how much you sell. Profits are how much you keep. Profits = Revenue - Expenses. The revenue is not sufficiently larger than the costs then there is no point in being in that line of business.
I'd say Sony already decided to use AMD before talking to Nvidia. The Nvidia meeting was only a way to make AMD reduce their asking price.
Nvidia do not sell APUs, so they had no chance really. Combining the GPU and CPU on same silicon reduces the price, the power consumption and also both CPU and GPU share the same memory, giving better performance (http://vr-zone.com/articles/amd-afds-seattle-trinity-follow-on-kaveri-to-have-true-shared-memory/16258.html)
Go to the original interview. Read the key quote.
"I'm sure there was a negotiation that went on,"
Does ANYONE in a company that is aware of a thing phrase their knowledge like this? "I'm sure" means "I don't know, but assume a company like ours must have been in the running, so I'll talk as if that is fact."
Here's the truth, for Slashdot readers so out of the loop, they are completely clueless.
Nvidia does not make console/PC-class CPUs. It struggles badly to even produce the Tegra ARM SoC parts- parts that are currently 2 years and falling against the original roadmap Nvidia had at the start of the Tegra project.
Nvidia does not make console/PC-class 'chipsets' - the 'glue' that interconnects the major system components. Once upon a time it produced chipsets for the PC market, but was pushed out entirely, first by Intel, and then by AMD.
All Nvidia could have provided was a VERY non-integrated GPU. That is a completely useless technology for this round of consoles from Nintendo, Sony, Microsoft and Valve. Nvidia was never in the running because Nvidia never had the ability to provide the required chips. Neither did Intel, by the way.
PS things get even worse for Nvidia when you consider that AMD fabricates many of its own chips via 'Global Foundries' (the 'independence' of the two is an investing and accounting con-job).
I think Nvidia realizes that this time around, consoles have a crapton more competition than they did 6-7 years ago. Phones and tablets are going to prevent a bunch of people from even thinking about buying a console. Likely the whole cycle sells less than half of the PS3/360 one.
Because the Dual Shock 3 doesn't speak the special "Made for iPad" protocol. Even on Android, I don't see a lot of people buying a $59.99 Dual Shock 3 for a $2.99 game, especially when Android system updates (such as 4.1 to 4.2) tend to change the Bluetooth stack to break the driver applications that these controllers use. It's like using a PC with a TV and multiple gamepads: possible, but in practice never done outside hardcore geekdom.
Casual gamers have a lot more options with tablets and smartphones than they did when the last generation of consoles came out with really only PCs and older consoles to compete against.
But do they actually make use of those options? For example, are tablet and smartphone users pairing Bluetooth gamepads to their devices? I'd love to see statistics saying so. But until then, consoles will dominate game genres that traditionally work better with the physical buttons of a gamepad than with a flat, feedbackless sheet of glass. The other option is to come up with some reliable control scheme for playing a platformer like Mega Man series on a tablet.
The first Steam box is the Xi3 Piston, which appears to be a rebadge of the X7A Modular computer.
The sad thing is, no one cares about the PC anymore.
"No one" is a strong phrase. What do you think developers are using to make mobile games other than a PC (or a Mac, which Apple builds from essentially PC parts)?
CPU technology will be in the quantum computer era within those 10 years as the nano scale limit is breached. It's basically a dead end for consoles. The consoles will have initial success and then ultimately fail within 4-5 years. They will not have a 10 year life span, 4-5 years is now the lifespan of consoles.
Quantum Computers? Really? I wouldn't hold your breath.
My money is we will have workable fusion power plants before we have a usable quantum computer that is able to play video games.
Low profit plus opportunity cost equals a bad decision. Nvidia made the right business choice. That capacity can now be used for more profitable products.
It's a crock of crap that the market isn't lucrative enough. They got beat out and when people started asking them how come not them, this is their answer.
No sane business holds grudges like that. If MS wants it, it'll be written into the next contract and either nVidia will agree or not get the contract.
Apple anyone? Rumor was Apple was going with Nvidia. Nvidia announced that had a deal with Apple and then Apple (well Jobs) killed the deal. Why? Apple has to announces things on Apple's schedule i.e. at some hyped Apple event.
Many businesses are run by what many would consider to be not sane people. Sometimes that helps the business and others it hurts the business.
I think you mangledthe rumor a little bit. I believe the alleged event you were alluding to was that Apple punishing ATI (not Nvidia) for a pre-announcement faux pax... Of course it probably actually didn't happen that way IRL, but it still makes a funny internet fable...
That's assuming that the games can be constructed to pass off the physics computing to another core in parallel to current processing. Some problems can't be solved by throwing another core onto it. It's like the old saying, just because it takes 9months for one woman to have a baby, you can't make a baby in a month with 9 women. If games don't change tremendously to be highly parallel then there isn't a major advantage.
Well, there's spam egg sausage and spam, that's not got much spam in it.
any PC that can run Steam and be connected to a TV
It's not whether it can be connected to a TV. Effectively all PCs for the past two and a half decades have a VGA or HDMI output or both, which works with an HDTV. It's whether it's marketed as being TV-ready, whether the case is designed not to look out of place next to a TV, and whether the software shipped on it is TV-ready with the option of USB gamepad control for everything. That might be enough to get over the traditional hurdles that hinder HTPC adoption by those who aren't hardcore geeks.
Does anyone remember NVIDIA's own console announced a couple months ago? There may be reasons NVIDIA didn't want to spend resources on PS4...
There is no IF here, they will have to adapt to 8 parallel threads at 1.6GHz. You cant write modern game running on one 1.6GHz core.
Who logs in to gdm? Not I, said the duck.
at least the 4600 line was. I wanted one really bad. Great performance and great visual quality. I just couldn't get the darn thing stable. I miss my 1600 (a slightly faster and cooler running 9800) but it can't keep up with Street Fighter 4 or Sonic Generations.
I've heard if you step up to the $300-$400 amd range the problems go away, but I don't want that much card. My GT 240 with DDR was $90 bucks....
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
Think about future games being natively written with 8 cores in mind.
Because game devs haven't spent the last 7 years writing games natively for 6 cores?
Again, some problems cannot be solved by throwing more cores on the problem. Using another core to handle the physics computation might be okay if the process can be done that way. If the game/physics process is interwoven it doesnt't gain any performance by using another core. If that isn't clear look at graphics cards. You can cross link them for better performance but that only gets you so much. Graphics cards from 5 or 6 years ago will not be able to handles the new Crysis at max settings no matter how many of them you throw at the problem. To play the new Crysis you are going to have to buy the newest and fastest graphics cards.
Well, there's spam egg sausage and spam, that's not got much spam in it.
Of course you can break almost every problem into smaller subproblems. Of course there are things you cant do (markov chains?), but games arent one of them, 90% of game code can run in parallel (including physics, how else do you think GPU would compute physics? GPU is a MASSIVE array of small CPUs).
Who logs in to gdm? Not I, said the duck.
Parallel computing relies on the assumption that the tasks can be executed concurrently. If the tasks are interdependent, parallelization fails to gain any advantage. Not all computing can be broken down in such a way. From Wikipedia:
"When a task cannot be partitioned because of sequential constraints, the application of more effort has no effect on the schedule. The bearing of a child takes nine months, no matter how many women are assigned."
In this particular case, throwing another core to run a physics engine may not provide enough benefit. By your logic, the only difference between the newest GPUs and older ones is that there are more cores in the new ones? Or did the manufacturers work on making the cores faster and changing the architecture?
Well, there's spam egg sausage and spam, that's not got much spam in it.