NVIDIA Pascal Thread

[Silverforce11]

Tell me when Pascal launches and tell me when Polaris launches. Exact dates. And then tell me when we should see all kinds of Polaris leaks.

I don't know, see, we're discussing rumors.

AMD have said they plan a June release for Polaris. NVIDIA has not made the date public. Rumors of a May release at this point is bull, unless its a paper launch.

 

[ShintaiDK]

I don't know, see, we're discussing rumors.

AMD have said they plan a June release for Polaris. NVIDIA has not made the date public. Rumors of a May release at this point is bull, unless its a paper launch.

So you base your assumption on nothing. Must be awful that Nvidia could release first

Where is all this Polaris information and leaks if its just 2 months or less away? According to you it must be paper launch.

 

[R0H1T]

The ASP has gone up because Nvidia has started selling mainstream GPUs for enthusiast prices, you know the thing that half this thread has been complaining about. And so what if they've taken away market share from AMD, doesn't change the fact that their strategy of increasing prices in a declining market has lead to record earnings.

Blame the buyers then, who buy these dGPU's on day one, when they're rewarding Nvidia for such tactics but I'm sure that won't happen. Free market, my choice & all that. The notebook market is also the reason why they get this premium because some buyers associate high end plus notebooks with Nvidia, a lot of the times because of their mobile dGPU.

Point is that airfathaaaaa couldn't comprehend how a company could go for higher margin and low volume vs lower margin and high volume, and still remain profitable, even though there are tons of examples of companies doing so and being very successful (Nvidia themselves being one of them).

The point is, rationally speaking, in a market where all buyers are neutral, most of them will go for more value for their money whilst the rest would be inclined to get the GPU whose (certain) features, or performance level, they prefer like Physx & full Async compute et al. This is where you get an equilibrium with fairly competent solutions sharing the market fairly equally. The marketing, mindshare & brand loyalty disrupts all this. The perfect capitalist markets do not exist, they never did, & we only have ourselves to blame for it.

At the end of the day it's simply about finding the optimal inflection point between price and volume (volume being driven by demand at the given price), all companies try to do this and Nvidia's strategy of raising prices has clearly succeeded and lead to higher total profits.

Yes it has but again not just because they have better performance & higher efficiency, the impact of brand awareness & loyalty can never be understated & is, for me, one of the main reasons why Nvidia's marketshare took off on 28nm.

This upcoming 14nm could be different & we may see AMD getting back to their "Evergreen" days.

 

[Silverforce11]

So you base your assumption on nothing.

If you have some hardware retail exp, you'll know which timeframes are important. Especially produced and shipped from China to the world.

All I am saying is 1 month out from a retail release, leaks are unstoppable due to the volumes of the products as well as the numbers of people whom have access to it already.

Feel free to disagree.

If it's late May and there's no detailed Polaris leaks, I would be worried about it being a paper launch too. This is certain since AMD have said June.

NV have not given a date, so it simply could be that rumors are wrong, rather than a paper launch.

 

[Sweepr]

I think some people are interpreting the rumours wrong. Latest info points to a mid-May press launch, probably specs reveal and maybe some reference model reviews from select websites. When they say May launch it's because Computex starts in late May and that's where formal launch should happen, with different custom models being shown by NVIDIA's partners. I expect retail availability for GDDR5 GP104 cards in late June or July.

 

[antihelten]

Blame the buyers then, who buy these dGPU's on day one, when they're rewarding Nvidia for such tactics but I'm sure that won't happen. Free market, my choice & all that. The notebook market is also the reason why they get this premium because some buyers associate high end notebooks with Nvidia, a lot of the times because of their mobile dGPU.

Why would I want to blame anyone? Nvidia is simply doing what they're supposed to do as a company, and I don't care what other people choose to do with their cash.

The point is, rationally speaking, in a market where all buyers are neutral, most of them will go for more value for their money whilst the rest would be inclined to get the GPU whose (certain) features, or performance level, they prefer like Physx & async et al. This is where you get an equilibrium with fairly competent solutions sharing the market fairly equally. The marketing, mindshare & brand loyalty disrupts all this. The perfect capitalist do not exist, they never did, & we only have ourselves to blame for it.

So your point is that in a market that has never existed, nor ever will exist, for any product anywhere in the world, Nvidias strategy is like "commiting suicide". Eerm ok but I don't see how that is even remotely relevant for the discussion at hand.

Ever heard of the joke about the physicist and the spherical cow?

Yes it has but again not just because they have better performance & higher efficiency, the impact of brand awareness & loyalty can never be understated & is, for me, one of the reasons why Nvidia's marketshare took off on 28nm.

This upcoming 14nm could be different & we may see AMD getting back to their "Evergreen" days.

No one ever said that the it was due to better performance/efficiency, quite the contrary, people have been complaining about the higher prices not being commensurate with higher performance (the whole selling mainstream GPUs at enthusiast prices thing).

 

[raghu78]

Tell me when Pascal launches and tell me when Polaris launches. Exact dates. And then tell me when we should see all kinds of Polaris leaks.

Nobody knows the exact dates. But I think we will hear from Nvidia and AMD on the respective launches in Q2 2016 with retail availability in late Q2 / early Q3 2016. Right now I would not bet who would launch first. AMD might want to see Nvidia pricing before launching as they are the one far behind in market share and we might see both launches spread apart by a couple of weeks.

 

[tviceman]

Nobody knows the exact dates. But I think we will hear from Nvidia and AMD on the respective launches in Q2 2016 with retail availability in late Q2 / early Q3 2016. Right now I would not bet who would launch first. AMD might want to see Nvidia pricing before launching as they are the one far behind in market share and we might see both launches spread apart by a couple of weeks.

It makes sense for AMD to launch a week or two later (if they're ready of course) to steal Nvidia's thunder. Then again AMD has been telegraphing their plans, timelines, chip strategy, and targeted performance improvements for quite awhile now so if Nvidia is confident launching whenever and however they want, then that IMO is telling in and of itself.

 

[R0H1T]

Why would I want to blame anyone? Nvidia is simply doing what they're supposed to do as a company, and I don't care what other people choose to do with their cash.

Then why did you bring that point up?

I don't either but if you're saying that their "enthusiast level pricing for mid range parts" strategy works, then there is more than one factor at play for that to be successful.

So your point is that in a market that has never existed, nor ever will exist, for any product anywhere in the world, Nvidias strategy is like "commiting suicide". Eerm ok but I don't see how that is even remotely relevant for the discussion at hand.

It may in a perfect world, but we don't live in it. Then again the original comment was in relation to smartphones & Apple, the only correlation to be drawn wrt Nvidia is the luxury brand image & their early mover advantage, especially with Maxwell. The dGPU market is just a duopoly & as such cannot be really compared with the much more competitive disposable electronics market, including the smartphone industry.

No one ever said that the it was due to better performance/efficiency, quite the contrary, people have been complaining about the higher prices not being commensurate with higher performance (the whole selling mainstream GPUs at enthusiast prices thing).

The strategy works as a novelty, it won't last forever. Eventually people will want more for their money & that will reflect in Nvidia's marketshare &/or profits, the latter can also be controlled through their cost structure.

 

[PPB]

Polaris is due in June.

The rumors are that Pascal will beat it to the release in May, no?

Basically, if we're a month from release and there's no detailed leaks, it's either a paper launch (press seeding/briefing, no retail til later), the rumors are wrong, or this is the most secretive mission impossible release ever.

Pro tip: check the mfg date of the gp104 die shot lying around. Its like someone at nvidia rushed to the interwebs with a picture of a die fresh from the fab to say "we areeee nooot lateeeee, see see!!!"

Sent from my XT1040 using Tapatalk

 

[Sweepr]

I think some people are interpreting the rumours wrong. Latest info points to a mid-May press launch, probably specs reveal and maybe some reference model reviews from select websites. When they say May launch it's because Computex starts in late May and that's where formal launch should happen, with different custom models being shown by NVIDIA's partners. I expect retail availability for GDDR5 GP104 cards in late June or July.

Quoting before the trolls say NVIDIA promised a hard launch in May. Anyway, if we were to believe the experts here there would be no Pascal till Q4.

 

[antihelten]

Then why did you bring that point up?

I never said anything about blaming anyone, nor have I ever complained about their prices in this thread, so I'm really not sure what point it is you think I brought up?

I don't either but if you're saying that their "enthusiast level pricing for mid range parts" strategy works, then there is more than one factor at play for that.

That obviously goes without saying. If their midrange parts had enthusiast level performance, then people wouldn't be calling them midrange parts, and since the strategy works (sales wise), it must obviously be because of something other than performance.

It may in a perfect world, that we don't live in. But again the original comment was in relation to smartphones & Apple, the only correlation to be drawn wrt Nvidia is the luxury brand image & their early mover advantage, especially with Maxwell. The dGPU market is just a duopoly & as such cannot be really compared with the much more competitive disposable electronics market, including the smartphone industry.

Apple has far higher margins than Samsung, but lower volume, and yet somehow despite them focusing on margins over volume they are nowhere near "committing suicide".

The fact that the smartphone market is more competitive than the GPU market just shows that going for margin over volume doesn't just work in a duopoly like the GPU market, but also in highly competitive markets. In other words it simply further reinforces the argument that prioritizing margin over volume is a perfectly viable business strategy (which again is the thing airfathaaaaa couldn't understand).

The strategy works as a novelty, it won't last forever. Eventually people will want more for their money & that will reflect in Nvidia's marketshare &/or profits, the latter can also be controlled through their cost structure.

Nothing lasts forever, but their pricing strategy is definitely not a novelty. The only novelty was AMDs small die strategy, before this strategy Nvidia was selling GPUs for $650+ and after it ended they went straight back to selling GPUs for $650+

It is far more likely that we will see prices continue to gradually climb upwards from here on out, expecting another period like AMDs small die strategy is optimistic to put it mildly.

 

[raghu78]

Quoting before the trolls say NVIDIA promised a hard launch in May. Anyway, if we were to believe the experts here there would be no Pascal till Q4.

Nvidia did not promise anything publicly. Its just a few websites who are speculating based on rumours and a few leaked die shots. Anyway we should have some confirmation this quarter from both Nvidia and AMD on the respective launches.

 

[Mopetar]

Nvidia did not promise anything publicly. Its just a few websites who are speculating based on rumours and a few leaked die shots. Anyway we should have some confirmation this quarter from both Nvidia and AMD on the respective launches.

All indications so far suggest AMD will launch first, but Nvidia won't be too far behind and will have the advantage of having a bigger GPU for 4 - 7 months depending on how long it takes AMD to release Vega.

Nvidia can afford to keep their cards closer to their chest and if they know that AMD will beat them to market, they have no reason to tip their hand to AMD in any way or it removes the benefit of being able to react to what AMD does based on performance and pricing of the cards that are launched.

They know that they have a loyal customer base so they don't need to release any early information, and even if AMD launches first a lot of people will wait to see what Nvidia's response is before buying a Radeon. AMD, however, knows that they lost badly during this current round so releasing information isn't going to hurt their sales all that badly at this moment or they don't care if it does.

 

[airfathaaaaa]

http://vr-zone.com/articles/msi-geforce-gtx-1080-gaming-8g-leaked/108755.html

 

[Glo.]

Erm... does the GTX 1080 has dual 8 pin connector?

 

[Krteq]

http://vr-zone.com/articles/msi-geforce-gtx-1080-gaming-8g-leaked/108755.html

Wait, are they trying to say that GP104 is pin compatible to GM204?

//... and GDDR5X are using the same signal traces?

WTF?

Are those GP104 photos fakes or what?

 

[antihelten]

Wait, are they trying to say that GP104 is pin compatible to GM204?

//... and GDDR5X are using the same signal traces?

WTF?

Are those GP104 photos fakes or what?

So sort of a woodscrews moment via leaks (i.e. a real chip, but mounted in a non-functional card). Feels a bit too conspiracy theory'ish imho, although it would explain the lack of significant spec/performance leaks that Silverforce11 mentioned earlier.

 

[airfathaaaaa]

or its just one more reason not to believe ANYTHING from any site from both sides...

they are basicly playing heads/tails just to be sure that in the end they will be on the right side

meanwhile the flood with missinformation about literally everything and we spend days here arguing about things that most likely might or might not be true

 

[antihelten]

or its just one more reason not to believe ANYTHING from any site from both sides...

they are basicly playing heads/tails just to be sure that in the end they will be on the right side

meanwhile the flood with missinformation about literally everything and we spend days here arguing about things that most likely might or might not be true

This isn't just a he said she said situation though. We have the actual picture of a GP104 die, which as far as I can tell looks real (real as in not photoshopped), and clearly it would appear that the PCBs matches up to the older card.

 

[Mopetar]

or its just one more reason not to believe ANYTHING from any site from both sides...

they are basicly playing heads/tails just to be sure that in the end they will be on the right side

meanwhile the flood with missinformation about literally everything and we spend days here arguing about things that most likely might or might not be true

With the Polaris speculation there is at least some (albeit very brief and limited) demonstrations of actual performance from AMD. It's very likely that they chose scenarios that show off the card in the best possible light rather than an average case, but between that and the patents that have been posted there's some concrete room for speculation.

As far as Nvidia goes there, there's a lack of credible information and all we really have at this point is official information on P100 which can tell us a little bit about what to expect with the rest of the Pascal lineup, but not a whole lot in general. Beyond that, the only real information we have is that they'll have a larger die than AMD and will probably launch within a similar time frame, but will likely launch after AMD.

 

[airfathaaaaa]

This isn't just a he said she said situation though. We have the actual picture of a GP104 die, which as far as I can tell looks real (real as in not photoshopped), and clearly it would appear that the PCBs matches up to the older card.

you know this is the mess.. you dont know what to believe anymore
is it real? maybe its a photoshop maybe it was an early prototype that has nothing to do with what we will see from nvidia who the knows anymore with so much BS from the sites

 

[antihelten]

Meh false alarm, Videocardz were just being lazy and didn't look at the PCBs properly. The PCB in the GP104 is not the same as the MSI GTX 980 Gaming 4G:

 

[airfathaaaaa]

Meh false alarm, Videocardz were just being lazy and didn't look at the PCBs properly. The PCB in the GP104 is not the same as the MSI GTX 980 Gaming 4G:

i think they said the layout is the same not that they used the very same pcb

 

[antihelten]

i think they said the layout is the same not that they used the very same pcb

from the videocardz article in question:

It’s a perfect match

Either way it's not even the same layout except for some very overarching features, seeing as there are a bunch of components on the GP104 pic that are not present on the 980 PCB. Basically the GP104 PCB just has the same mounting holes layout and it's also an increased height like the 980 PCB, but that's pretty much as far as it goes (aside from the obvious reference layout stuff).

Either way this is all fairly irrelevant.