AMD: Speculation, Rumors, and Discussion (Archive)

[Kaotik]

Many mistakes.

GP100 has already been announced yet article shows Q4-2016.

Volta is 2017 not 2018/2019.

http://techfrag.com/2016/01/25/nvid...t-in-april-gtx-1080-in-june-and-volta-in-2017

GP106 is July 2016 not Sept 2016.

The one you linked is just as much speculation as 3DCenters.
GP100 might refer to actual production for anything but the DXG-1-machines, in PCIe form.
GP106 has been said to be july, august and september by different sources, any of them is just as likely as the next.
Volta is around late 2017 / early 2018 according to NVIDIAs latest roadmaps. Based on their past schedules, 2018 is more likely, and probably not even too early in the year (I'd put my money on march/april and GTC 2018)

 

[xEx]

http://www.3dcenter.org/news/amd-nvidia-grafikkarten-roadmap-201617

That wouldn't make sense since P10 = big part P11 = small part but then V10 = small part V11 = big part?

 

[Razor1]

That wouldn't make sense since P10 = big part P11 = small part but then V10 = small part V11 = big part?

That what its always been. There is no "size" correlation between the numbers, only a time correlation, 10 was on paper before 11 .

 

[CSI PC]

The one you linked is just as much speculation as 3DCenters.
GP100 might refer to actual production for anything but the DXG-1-machines, in PCIe form.
GP106 has been said to be july, august and september by different sources, any of them is just as likely as the next.
Volta is around late 2017 / early 2018 according to NVIDIAs latest roadmaps. Based on their past schedules, 2018 is more likely, and probably not even too early in the year (I'd put my money on march/april and GTC 2018)

Yeah generally you can only buy the P100 now as part of the DXG-1 solution.
But Nvidia's closest clients (big multi-million $ solution) can buy the P100 directly as well, although such supercomputer projects take months of planning to replace older K20s and K40s, these clients are buying P100 in the thousands for a single project.

Volta HAS to be 2017 to meet their contractual obligations as part of 2 new supercomputers.
Cheers

Edit:
This from the Oak Ridge National Labatory:

Summit is the next leap in leadership-class computing systems for open science. With Summit we will be able to address, with greater complexity and higher fidelity, questions concerning who we are, our place on earth, and in our universe.

Summit will deliver more than five times the computational performance of Titan’s 18,688 nodes, using only approximately 3,400 nodes when it arrives in 2017. Like Titan, Summit will have a hybrid architecture, and each node will contain multiple IBM POWER9 CPUs and NVIDIA Volta GPUs all connected together with NVIDIA’s high-speed NVLink. Each node will have over half a terabyte of coherent memory (high bandwidth memory + DDR4) addressable by all CPUs and GPUs plus 800GB of non-volatile RAM that can be used as a burst buffer or as extended memory. To provide a high rate of I/O throughput, the nodes will be connected in a non-blocking fat-tree using a dual-rail Mellanox EDR InfiniBand interconnect.

Upon completion, Summit will allow researchers in all fields of science unprecedented access to solving some of the world’s most pressing challenges.

I have been following the planning schedule and milestones and it is pretty tight, meaning the confidence has to be their to meet this commitment.
Caveat would be this is a Tesla part.

 

[TARTOFCP]

I heard some rumors.
That may or may not be true.
a certain chip have 48 cu full
consumer chip have 40-44 cu
v...

 

[xEx]

That what its always been. There is no "size" correlation between the numbers, only a time correlation, 10 was on paper before 11 .

So the 10 are the first to come out and the 11 the second? I don't know how I feel about that nomenclature lol. they surely had to have another one they use internally to difference their product.

 

[trinibwoy]

Love it.

http://videocardz.com/61007/amd-wants-gamers-to-start-the-uprising

 

[xEx]

OMG! marketing doing right for the first time in AMD history

 

[Razor1]

They have done this stuff before, I would rather them spend money and time and actual hardware and better presentation of games than this type of fluff.

If they have to have a marketing campaign for gamers, which gamers kinda have an idea of what they are getting anyways, its not like the average idiot on the corner, they have another thing coming.

Seriously, would you click on an online ad like this when you see it?

I'm going to take a quote out of the forbes article

Chris Hook interjects: “What was cool about Raja’s vision was that it was very much ‘if we focus on community first, and we listen carefully to what we get tweeted at everyday, and we make the customer first, it may take us some time to build the business but we’re going to be on the right track.”

Doesn't this sound exactly like what they stated about the rv680?

Sorry AMD, too much BS........

 

[trinibwoy]

Yeah, it's still a gamble though. If their cards are cheaper and slower they haven't really accomplished anything except cement AMD's reputation as the "budget" option that can't run with the big boys.

For this to work they need to make nVidia's cards look overpriced and out of touch with the 99%.

 

[Razor1]

Yeah, it's still a gamble though. If their cards are cheaper and slower they haven't really accomplished anything except cement AMD's reputation as the "budget" option that can't run with the big boys.

For this to work they need to make nVidia's cards look overpriced and out of touch with the 99%.

Like they did with the 4xxx line, but this isn't going to happen, performance isn't close enough to create that kind of backlash.

 

[Razor1]

BTW this advertising campaign, reminds me of a creative director that I worked with at Publicis many years back. I certainly hope they didn't go through an advertising company, cause they are paying a ton through the nose for crap.

 

[Esrever]

Nvidia just has to counter with "better dead than red"

 

[renderstate]

LOL

 

[Anarchist4000]

So the 10 are the first to come out and the 11 the second? I don't know how I feel about that nomenclature lol. they surely had to have another one they use internally to difference their product.

AMD just refer to "Polaris" and not the 10/11 even though they have associated parts. So they designed 10, then realized they needed a smaller one so made 11. The next step up from Polaris 10 was Vega, so 10 there. Then an even bigger Vega, 11.

Yeah, it's still a gamble though. If their cards are cheaper and slower they haven't really accomplished anything except cement AMD's reputation as the "budget" option that can't run with the big boys.

For this to work they need to make nVidia's cards look overpriced and out of touch with the 99%.

Do we even know they're slower/cheaper though? Polaris 10 is the competition for GP106 which isn't even out yet.

 

[xEx]

AMD just refer to "Polaris" and not the 10/11 even though they have associated parts. So they designed 10, then realized they needed a smaller one so made 11. The next step up from Polaris 10 was Vega, so 10 there. Then an even bigger Vega, 11.


Do we even know they're slower/cheaper though? Polaris 10 is the competition for GP106 which isn't even out yet.

That makes sense.

Also this is a tech forums so people sometimes doesn't understand the think of normal people. Marketing is as important as the product itself, and yes you gonna cross me for saying that but i still have to say it. If AMD wants to make money again they need to change their image of the bad cheap company. yes having good products its the starts but that alone wont change your imagine.

 

[Razor1]

That makes sense.

Also this is a tech forums so people sometimes doesn't understand the think of normal people. Marketing is as important as the product itself, and yes you gonna cross me for saying that but i still have to say it. If AMD wants to make money again they need to change their image of the bad cheap company. yes having good products its the starts but that alone wont change your imagine.

Advertising doesn't change the image over night either. Advertising works well to get your name out there, people that are targeted don't know much about the products in that market, and when there are many other alternatives to what you have. In this case, there is only one other alternative, people buying the products are not fully sheep lol, and the AMD/Radeon name is pretty well know.

You are saying they have a bad image, or bad brand name which is associated with a cheap alternative, this means its well know already. The only way to change the perception is not to go on an advertising campaign but to show the products and what they can do and why they aren't a cheap alternative.

I don't know but do you remember the first Infinity commercials (many years ago) they never showed the car, but they marketed it as being as good as lexus, MB, BMW, Audi? Yeah it was a flop, not only did it hurt the Infinity sales, it hurt their image to the point for many years to come and still haven't been able to compete in the marketplace very well.

 

[xEx]

Advertising doesn't change the image over night either. Advertising works well to get your name out there, people that are targeted don't know much about the products in that market, and when there are many other alternatives to what you have. In this case, there is only one other alternative, people buying the products are not fully sheep lol, and the AMD/Radeon name is pretty well know.

You are saying they have a bad image, or bad brand name which is associated with a cheap alternative, this means its well know already. The only way to change the perception is not to go on an advertising campaign but to show the products and what they can do and why they aren't a cheap alternative.

I don't know but do you remember the first Infinity commercials (many years ago) they never showed the car, but they marketed it as being as good as lexus, MB, BMW, Audi? Yeah it was a flop, not only did it hurt the Infinity sales, it hurt their image to the point for many years to come and still haven't been able to compete in the marketplace very well.

Im not saying that you could just spend your budget in marketing and get away with it. Im saying that with good products alone you cant, you need a good marketing in order to get the people know and get interesting, excited, hyped about your product. Also its easier to sell a bad product with good marketing than a good product with bad marketing(HTC vs Samsung for example)

 

[Razor1]

Difficult to do when you look at the target audience. They aren't going to be easily fooled with good marketing of a lesser product. If you have a comparable product with good marketing yeah that would work.

people buying smart phones aren't the same people that buy AIB's. You can't "fool" these people into thinking the product is about the same or better if its not.

 

[Pressure]

The marketing seems spot on though.

They will give you a better VR experience than the minimum VR spec (GeForce GTX 970) at an affordable $199. They don't say they are the fastest (for obvious reasons) but the sensible choice for a much larger market than the 1% market.

Data from the steam survey may back this up.