Watch AMD's Vega livestream here

Or NVIDIA won't be releasing a 1080Ti and it's all been a huge red herring, lets face it they don't need to release a Ti variant of the 1080. All this could just be a ploy to see what AMD would do with the VEGA announcement..

That'd be cheeky alright, but yeah could happen. They indeed don't need a Ti perse; the 1080 is that good!
 
Perhaps, however I remain skeptical; both AMD and Nvidia haven't abandoned them for nothing and I don't think DX12 can change that. Running two cards simply provides a far smoother experience.

It could theoretically encourage the game developers to include support for multi-GPU setups if they knew that the horse power was available to them and the API was suited to it. I dunno.
 
It could theoretically encourage the game developers to include support for multi-GPU setups if they knew that the horse power was available to them and the API was suited to it. I dunno.

Oh absolutely! Previously the problem always lay with the designs of the dual chip cards and their drivers. So expensive to develop for only a meagre outcome, a Ti is a far more satisfying product.
 
Oh absolutely! Previously the problem always lay with the designs of the dual chip cards and their drivers. So expensive to develop for only a meagre outcome, a Ti is a far more satisfying product.

Yeah, I doubt Nvidia will skip the Titanium card. The 980ti sold incredibly well. It even outsold some lower end models. It was just a well placed GPU. If the prices hadn't shot up in Europe due to the currency exchange taking a crap on us, I would have bought a 980ti and a G-Sync panel over a Fury and a Freesync panel. Such a brilliant card, and it was a beast under water. Absolute beast.
 
They may skip the TI.

As has been said the 1080 is a very fast card. It's also only mid range silicon which is much cheaper to make (and has a higher success rate per wafer) than larger dies. So the profit they are making on the 1080 will be far higher than they could make on the TI.

There is no logical reason to release a TI either. All it would do is harm the sales of the 1080 which they are making an absolute mint on given it was a free launch* and the Titan XP which they are claiming £1100 or so for. Why would you sell a GPU with the same die as the Titan XP for cheaper and dethrone your own 1080?

Nah, they won't release a TI unless they absolutely have to. So unless AMD can force their arm they will just stay put with their astronomically priced mid range silicon.

* Pascal never existed before. It was Maxwell then Volta. As has been established Pascal is simply Maxwell with far higher clock speeds, or as it has been called "Maxwell on speed". This means "pascal" as a tech was nothing but a die shrink of Maxwell and thus R&D costs would have been significantly low. It will also be why they won't rush on to Volta because again, why spend money on something they don't need out there yet?
 
They may skip the TI.

As has been said the 1080 is a very fast card. It's also only mid range silicon which is much cheaper to make (and has a higher success rate per wafer) than larger dies. So the profit they are making on the 1080 will be far higher than they could make on the TI.

There is no logical reason to release a TI either. All it would do is harm the sales of the 1080 which they are making an absolute mint on given it was a free launch* and the Titan XP which they are claiming £1100 or so for. Why would you sell a GPU with the same die as the Titan XP for cheaper and dethrone your own 1080?

Nah, they won't release a TI unless they absolutely have to. So unless AMD can force their arm they will just stay put with their astronomically priced mid range silicon.

* Pascal never existed before. It was Maxwell then Volta. As has been established Pascal is simply Maxwell with far higher clock speeds, or as it has been called "Maxwell on speed". This means "pascal" as a tech was nothing but a die shrink of Maxwell and thus R&D costs would have been significantly low. It will also be why they won't rush on to Volta because again, why spend money on something they don't need out there yet?

This makes a lot of sense, yeah. Theoretically, when Vega 10 is released and if it doesn't beat a Titan XP and Nvidia feel no desire to release a Ti version, the sales of the Titan XP could suddenly jump as those that waited for the Ti now no longer feel the need to wait. And you'll also have a price drop of the 1080 and those that previously could not afford one jump on that.
 
This makes a lot of sense, yeah. Theoretically, when Vega 10 is released and if it doesn't beat a Titan XP and Nvidia feel no desire to release a Ti version, the sales of the Titan XP could suddenly jump as those that waited for the Ti now no longer feel the need to wait. And you'll also have a price drop of the 1080 and those that previously could not afford one jump on that.

Nvidia screwed a lot of people with Titan X(M). It was nearly a grand (£960 for an Asus card like I have) and then within a couple of months they launched the 980ti (to knock out the Fury X) but all of a sudden you have a card costing over £400 less (they were selling for £549 same as Fury X) and then all of a sudden absolutely nobody buys Titan XM before and you're selling the same die basically for £400 less. So that's £400 gone from each sale and you just screwed your users over for £400 for 6gb VRAM.

Yeah, that was a dirty launch IMO. Obviously they don't care for their users at all but when they are also losing money it was kinda stupid.

They won't repeat that unless AMD have something pretty serious. Any slower than the 1080 and they can just leave everything in place and tweak the price of the 1070 to suit (probably £40 or so more than the AMD 'cause Nvidia tax).
 
* Pascal never existed before. It was Maxwell then Volta. As has been established Pascal is simply Maxwell with far higher clock speeds, or as it has been called "Maxwell on speed". This means "pascal" as a tech was nothing but a die shrink of Maxwell and thus R&D costs would have been significantly low. It will also be why they won't rush on to Volta because again, why spend money on something they don't need out there yet?

That's incorrect actually: Pascal isn't tweaked Maxwell (please do correct me with a trustworthy link if I'm wrong (eg no Wikipedia dung he he)). And Pascal existed for a very long time - you can look up old slides. At first the order was: Maxwell - Volta - Pascal. The latter two got reversed right before the former came out.

On a side note; throughout the years I see many mentioning the high prices by Nvidia (and they are). I'd like to think that it's because they have the better product and thus can ask what they want (leaving fair out of the equation; its a money making company, not a charity to please the common mab). More interestingly is that I'm pretty sure that AMD would do the same were the roles reversed, if only to cover their losses ;)
 
Last edited:
That's incorrect actually: Pascal isn't tweaked Maxwell (please do correct me with a trustworthy link if I'm wrong (eg no Wikipedia dung he he)). And Pascal existed for a very long time - you can look up old slides. At first the order was: Maxwell - Volta - Pascal. The latter two got reversed right before the former came out.

It's a tweaked Maxwell chip. Just look at them no need for a link?
Maxwell is a tweaked Kepler chip, Kepler is a tweaked Fermi chip. It's been this way since then. Pascal is just a very small tweak with added features that the 14nm shrink allowed room for. In addition it allowed higher clock speed due to FinFet.

Order doesn't matter. It's just a codename
 
That's incorrect actually: Pascal isn't tweaked Maxwell (please do correct me with a trustworthy link if I'm wrong (eg no Wikipedia dung he he)). And Pascal existed for a very long time - you can look up old slides. At first the order was: Maxwell - Volta - Pascal. The latter two got reversed right before the former came out.

On a side note; throughout the years I see many mentioning the high prices by Nvidia (and they are). I'd like to think that it's because they have the better product and thus can ask what they want (leaving fair out of the equation; its a money making company, not a charity to please the common mab). More interestingly is that I'm pretty sure that AMD would do the same were the roles reversed, if only to cover their losses ;)

I also remember seeing Pascal on an early roadmap leak. I kind of adopted the whole 'Pascal was never planned' thing that I hear on this site because so many were so confident about it. But I remember seeing it in an early leak.

Regarding your second point, I don't know if I agree. A product does not need to cost so much. If the product or service does cost so much, in my opinion it should be because the manufacturer/seller believes wholeheartedly that what they're offering is truly special and should only be available to the elite who see the greatness of what they're doing and are willing to set aside a good chunk of change for it.
I'm a musician and I seen it all the time with guitar manufacturers. Some guy in Germany deliberately charges twice his competitors because he only wants to make 15 guitars per year and only wants to sell to the few who cherish his work deeply. He doesn't care about the bottom line or marketing or mass appeal; he believes his product or service is prestigious and deserves a hefty premium, despite being equal to his significantly cheaper competitors. But his competitors produce 50 guitars per year and can get away with charging less without devaluing their business. That's a distinction I believe in.

The same could be applied to fine dining. A Michelin 2-star fine dining restaurant might upcharge their customers 20% and have a three-year waiting list, while their no-Michelin-star competing restaurant offers an equal dining experience with only a six-month waiting list. Why? Because they thoroughly believe that what they're doing is special and deserves to be recognised in a very specific way by a very specific clientele. Snobbery, in other words.

But that's not necessarily Nvidia. Nvidia seem to be charging so much both because they can (no competition) and because they are confident in their products. The 1080 is great, but it's not a revolution. They don't appear to have higher overheads and the chips shouldn't be that much to produce, but they have no current competitor and they like their product so they'll sell them at a premium. The problem, people feed into this; they continue to support such arrogance despite being intellectual enthusiasts.

Apple is in a similar situation, but far worse. They charge a premium almost solely because of mindshare. Some of their products were in the past phenomenal, but now they have so much competition that they're not the best any more. Still they remain ever popular, betraying their exclusivity pricing structure. It's why people associate hipsters with Apple. Hipsters are comically known for being purposefully different and going against the grain, but they still choose to buy from Apple, a company based around forgotten greatness and marketing. Beats by Dre, Alienware, all mediocre products sold as exorbitant prices to those that are deluded enough to feed into the hype. Nvidia aren't that bad, but they could theoretically become that bad if people don't support AMD with their wallets.
 
Last edited:
That's incorrect actually: Pascal isn't tweaked Maxwell (please do correct me with a trustworthy link if I'm wrong (eg no Wikipedia dung he he)). And Pascal existed for a very long time - you can look up old slides. At first the order was: Maxwell - Volta - Pascal. The latter two got reversed right before the former came out.

On a side note; throughout the years I see many mentioning the high prices by Nvidia (and they are). I'd like to think that it's because they have the better product and thus can ask what they want (leaving fair out of the equation; its a money making company, not a charity to please the common mab). More interestingly is that I'm pretty sure that AMD would do the same were the roles reversed, if only to cover their losses ;)

Pascal is Maxwell on speed. It's as simple as that.

CljWQzc.jpg


That is the original slide that was released when Kepler released. There was no Pascal at all, just Maxwell straight to Volta.

https://www.youtube.com/watch?v=nDaekpMBYUA

At the same clocks Maxwell is pretty much Pascal.

Nvidia got a lucky die shrink with enormous clocks (thanks to a small die without a kitchen sink)

The original video is here.

https://www.youtube.com/watch?v=7jIDbZG1VxY
 
Pascal is Maxwell on speed. It's as simple as that.

CljWQzc.jpg


That is the original slide that was released when Kepler released. There was no Pascal at all, just Maxwell straight to Volta.

https://www.youtube.com/watch?v=nDaekpMBYUA

At the same clocks Maxwell is pretty much Pascal.

Nvidia got a lucky die shrink with enormous clocks (thanks to a small die without a kitchen sink)

The original video is here.

https://www.youtube.com/watch?v=7jIDbZG1VxY

Good post. Thanks. That makes it pretty definitive, IMO: Pascal was just a refreshed Maxwell.
 
I tried looking for the reversed slide but can't find it. It did exist and I remember it being posted on another forum in a discussion I was a part of. It doesn't matter.

Don't get me wrong; I too believe that Nvidia is way overcharging but I don't see my buying their stuff as supporting them. I buy it because I'm looking for specific performance and features and it so happens to be them that qualify. The same applies to a brand of peanut butter I may buy, if that makes sense.

Musician - cool! What guitars do you play? I used to play bass.
 
I tried looking for the reversed slide but can't find it. It did exist and I remember it being posted on another forum in a discussion I was a part of. It doesn't matter.

Don't get me wrong; I too believe that Nvidia is way overcharging but I don't see my buying their stuff as supporting them. I buy it because I'm looking for specific performance and features and it so happens to be them that qualify. The same applies to a brand of peanut butter I may buy, if that makes sense.

Musician - cool! What guitars do you play? I used to play bass.

I'm a guitarist. I'm a teacher, actually. Well, I'm trying to be. :D I've been playing since I was around 14 and I'm 28 now; so half my life. I lubz it. If it weren't for being a musician I would probably invest more in computers and do things like water cooling and have a dual GPU setup. I can't afford too expensive passions. #firstworldproblem

If you're looking for a specific performance metric and Nvidia provides it, that's totally fine. In fact, that's great. But buying it just because it's 'OMG SO MUCH POWERZ!!1!', that's different. I wasn't suggesting that's why you or anyone in particular is doing that, but it does happen very often.
 
I'm a guitarist. I'm a teacher, actually. Well, I'm trying to be. :D I've been playing since I was around 14 and I'm 28 now; so half my life. I lubz it. If it weren't for being a musician I would probably invest more in computers and do things like water cooling and have a dual GPU setup. I can't afford too expensive passions. #firstworldproblem

If you're looking for a specific performance metric and Nvidia provides it, that's totally fine. In fact, that's great. But buying it just because it's 'OMG SO MUCH POWERZ!!1!', that's different. I wasn't suggesting that's why you or anyone in particular is doing that, but it does happen very often.

Nice man! What's your favourite guitar? For guitars (played a little next to my main bass playing) it was an SG, for bass a Fender Aerodyne. Lol @ #firstworldproblem (been there - am there, I'm fact.

Oh I knew you weren't! Just wanted to clarify that in general. At 34 I'm way too old to pick sides, let alone make purchases for the so much power reason (but yeah that happens) IMHO.
 
Back
Top