Hopefully Volta will be out middle of next year at the latest but start of the year would be better as it would be easier to skip this current gen for me
What if it costs £1180 though
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Nvidia Potential Roadmap Update for 2017: Volta Architecture Could Be Landing As Early As 2H 2017
- Thread starter Gregster
- Start date
What if it costs £1180 though
Apparently they're working on screens up to 2250 PPI at the moment http://www.androidauthority.com/samsung-11k-display-624185/
That would mean 23.6x the current VR resolution. Or 10496x5831 (5248x5831 per eye) at the same physical size, and FOV, as the current headsets.
So as long as they don't drop this research, it's looking good
I think this is very sensible.
I did get a Vive 1, but it's partly because I'm an enthusiast and partly because I saw it as essentially a kickstarter for the second generation. Hoping that if enough people gave them money for what is essentially a very very refined tech demo, they will get serious with gen2 for everyone to enjoy.
Soldato
All these guys going on about async compute, if Nvidia do implement this, won't it just make their lead that much further?
Atm a RX480 gets a huge boost over the GTX 1060 for example in DX12, Vulkan, Async compute. If Nvidia also go this route with their next line we could see AMD beaten in every metric, which I am not sure is a good thing. We need AMD to compete at whatever level, because if you think current prices are bad. Imagine a world where AMD exits from PC GPU's, that new GTX X6 card will probably cost £400.
Atm a RX480 gets a huge boost over the GTX 1060 for example in DX12, Vulkan, Async compute. If Nvidia also go this route with their next line we could see AMD beaten in every metric, which I am not sure is a good thing. We need AMD to compete at whatever level, because if you think current prices are bad. Imagine a world where AMD exits from PC GPU's, that new GTX X6 card will probably cost £400.
I mean, sure. AMD need to get their DX11 performance in order. Or games need to rapidly transition to DX12 and Vulkan.
It looks like Volta will be Nvidia's true 'full' architecture for all these new features. Pascal is just very good at coping, rather than being fully designed for it from the ground up.
It all depends on how DX12/Vulkan support goes. It may be that AMD maintains an edge there (but much smaller), but if Volta matches/exceeds Vega in DX12 even by a little, then AMD will also have a reson to hope customers 'forget'.
On AMD side, they will hope that users 'forget' DX11 games and accept that AMD's performance is 'good enough'. (This is not unlikely. For example, I already think the 480 has 'good enough' DX11 speed compared to the 1060).
On NVidia side, they're already hoping users forget that they ever promised good DX12 performance for Maxwell with 'drivers' which everyone will still be waiting for.
Permabanned
Navi is where things will get really, really interesting. I have little doubt that Vega will compete - neither does nVidia which is why, if true, they are moving Volta forward.
Navi on the other hand will be a whole new playing field due to being scaleable - nVidia Volta performing too well? OK stack another few dies on an interposer and shut them down. Given that Vega looks to be the last "big silicon" die AMD release on the GPU side Navi will make bucket loads of profit also being a small die, scaleable and modular.
Fury was the "proof of concept" if you will for the process, but Navi will (I think) have a single interface that treats multiple cores as one GPU, effectively rendering Xfire / SLI redundant, especially given the future direction of API's being explicit & implicit MGPU enabled.
Exciting times ahead in 12 - 18mths, moreso than in the last 20yrs of GPU development I think.
Navi on the other hand will be a whole new playing field due to being scaleable - nVidia Volta performing too well? OK stack another few dies on an interposer and shut them down. Given that Vega looks to be the last "big silicon" die AMD release on the GPU side Navi will make bucket loads of profit also being a small die, scaleable and modular.
Fury was the "proof of concept" if you will for the process, but Navi will (I think) have a single interface that treats multiple cores as one GPU, effectively rendering Xfire / SLI redundant, especially given the future direction of API's being explicit & implicit MGPU enabled.
Exciting times ahead in 12 - 18mths, moreso than in the last 20yrs of GPU development I think.
Associate
looking forward to that as well.
Nvidia would not benefit from raising low/mid-priced cards further even under a monopoly. They can get away with high prices for enthusiast cards because enthusiasts will spend the money. Low/mid range users, not so much. Pricing them out will only hurt the market and damage PC gaming, which would be the last thing they want to do.
I'm not saying an Nvidia monopoly would be a good thing, but some people get way too dramatic about what it would mean. Either way, it's up to AMD to compete for our money, not for us to support the underdog with an inferior product.
Also, I hear talk about 4k VR headsets - we definitely aren't there yet. We dont even have a viable 4k/60fps single card right now and you want something that can do 4k/90fps+ in VR? Aint happening for another 2-3 years. Only thing that'll save us is if we get foveated rendering very soon.......which isn't impossible. Some good progress on it right now. Seems like the hardware is nearly there, we just need the software to implement it in an optimized and adaptable manner, which isn't so easy.
And in terms of Volta, at first I dismissed this article as low level rumor BS, but if Nvidia are already putting out a Titan, they are going to need something good for later next year. Launching Volta could well be just what the doctor ordered. Problem is, last roadmap we saw had Volta around 2018 and how often does a new architecture come in sooner than anticipated?
Do we even know what AMD means by 'scalable' specifically?
When I saw that, I took it to mean they might go for an architecture where they could get more variants per die and have them all hit good levels of TDP and performance.
Also not sure what you mean by Fury being proof of concept. They used stacked memory, not stacked graphics processors. Not the same thing by any stretch.
And there's nothing to suggest that future API's are looking to adopt inherent multi-GPU capability. Especially not when we're talking about low level API's.
I'm not sure things are quite as exciting as you're thinking.
Permabanned
Something forced their hand Seanspeed. They are speeding up their development & release cycle to compensate.
In response to your other comment regarding monopoly, have you enjoyed the tick - tock Intel performance increases with high prices you've enjoyed for the last 10 years?
Looks like relief is coming on that front with Zen as well. My AMD shares are returning a 47% profit as I speak if I bailed out but I see over 10USD per share by new year.
In response to your other comment regarding monopoly, have you enjoyed the tick - tock Intel performance increases with high prices you've enjoyed for the last 10 years?
Looks like relief is coming on that front with Zen as well. My AMD shares are returning a 47% profit as I speak if I bailed out but I see over 10USD per share by new year.
Permabanned
Scaleable means scaleable as in you can increase performance by modual stacking, it's not die cutting for profit. Fury proved for the first time that a GPU could be mated to an interposer with HBM and then mounted to a card. Proof of concept. It's going that way.
All forward API's are looking at explicit and implicit MGPU - it's the new paradigm don't ya know?
Please keep up.
Well we have no idea if this rumor is even true.
And if it is, we do not know it is because their hand is being forced. They could simply be ahead-of-schedule. It would be unusual, but it would also be unusual for them to have the ability to have room to push things forward significantly. Not like Nvidia are back in the engineering labs leaning back and sipping coffee. This is a breakneck speed industry by nature.
I have no problem with Intel. People who think they are also sitting back, sipping coffee have no idea what they're talking about. Intel are pushing forward relentlessly. Margins for gains just aren't what they were before. Especially for gaming.
I also have no problem with Intel CPU pricing. For something that will easily last five years without getting into "I really need to upgrade" territory, spending $250-350 on a good CPU is great value, if you ask me.
Not sure how this applies to Nvidia, though. I do agree that stagnation and impetus to push the tech would be my bigger concern in an Nvidia monopoly, but not as much as you, I dont think. There'd still be plenty of reason to push things, largely because continued progress supports the PC gaming market. Stagnation on the GPU front, unlike the CPU side, would kill people's impetus to want to buy a new GPU. Imagine if GPU's were like CPU's and you only needed to buy one every five years to keep up with the latest games with great performance? Nvidia would suffer in that situation. It is in their best interest to push forward as much as they can, regardless of competition. They want to drive the graphics in games to greater heights, which drives demand for new GPU's. It's an ever constant cycle that supports Nvidia's business from a very fundamental level.
Again, is that you guessing or do you have a source for that?
Like I said, stacking a GPU is not the same thing as stacking memory.
No need for the condescending snippy attitude bud.
Anyways, no, I've not seen any paradigm shift towards multi-GPU with the new API's. Quite the opposite.
Permabanned
1st point, nope, got nothing but scaleable doesn't mean and never has been about die cutting, feel free to put this assertion in the spank bank for future correction if I'm proven wrong, I'm on the record.
Your second point is moot, I said putting GPU & memory on an interposer was the proof of concept. Nobody else ever did it with a GPU and if you are going to make a modular, scaleable one you best use this method. You are splitting hairs now. dismissed.
As far as API's go I'm not even going to respond, do your own research and prove me wrong. It's that retarded an argument.
I'll get snippy when people make a dumb argument about the final detals of technology that hasn't been released when asked to prove it - prove Volta will be released before the formal time frame. Prove what Volta is? At least we have an indication on what Navi is.
So go on and tell us exactly what Volta is about again Mr Seanspeed?
No, it hasn't so far because current architectures can only scale so much on a given die. If they can improve that with an architecture that works well in both TDP and performance at larger ranges, they can get far more products out of a single die.
This would indeed be a 'scalable' architecture.
Anyways, thanks for answering my question - you are simply guessing and have no source.
It was a proof of concept of what? Using an interposer? That does not mean that stackable graphics processors are do-able. Maybe they are, I'd just like a stronger argument than pointing to a product that stacked memory. This is not splitting hairs by any means. It is crucial as to whether or not it's something we can expect.
You made the claim, it's you who needs to support your assertion.
I've been keeping up plenty. Multi-GPU in the new API's is made harder due to not having the handy abstraction handling this in on the driver level. Notice how few DX12/Vulkan games support it? Obviously it can be done, but there is absolutely nothing to suggest it is going to be more supported than before, much less become an outright paradigm shift.
What are you talking about? I never claimed to have the answers to these things. :/
Anyways, I've learned very quickly you are not worth ever responding to again. You're rude, condescending, have no intention of supporting your arguments with anything meaningful, and also admitted you are financially invested in AMD. A cocktail for an absolutely insufferable discussion...
I really hope you dont talk like this to people in real life man.
Associate
A lot of truth in this statement.
Quite a few people on this forum need to seriously look in the mirror and consider their behaviour. Unfortunately a forum seems to encourage a certain behaviour which sometimes is not even a sign of poor social skills. It's just that people hide behind a random name and feel they can post whatever and treat others in whatever way they feel like it.
It's sad really, especially when people come here looking for help and meaningful technical discussions and instead they get crap all.
Caporegime
So that means the new Titan X will probably only have a lifespan of a year?? That is about the same lifespan the GTX980TI had too!!
You say that as if you are surprised
Graphics card get out of date very quickly.
Caporegime
You say that as if you are surprised
Graphics card get out of date very quickly.
Well for a £1100+ card to be outdated that quickly lol.
It makes me wonder if people with a GTX980TI should just skip the new Titan X especially considering that Nvidia is supposedly supplying agressive rebates to clear Maxwell stock.
Volta is meant to be a new uarch which is more compute focussed.
At least the sub £300 cards are more like 18 months to 24 months.