[TT] Pascal rumored to use GDDR5X..

[thesmokingman]

Higher density GDDR5 chips should use less power since you need less of them?

I'm not sure if 20nm is smaller than current GDDR5 lithography, but if it is, that should reduce power as well?

I don't think there's a free lunch, otherwise why would both NV and AMD have put in the research into stacked dram? There are other issues that are problems as well.

The company said it observed that command address protocoling and array speed were the two limiting factors, while the interface had plenty of additional headroom.

Second, however, it only addresses the bandwidth part of GDDR5’s problem. There’s always a power cost associated with increasing memory bandwidth, and simply doubling prefetch isn’t going to magically reduce GDDR5’s power consumption. One of the advantages of HBM that we covered long before AMD began talking it up this year is that HBM significantly improves GPU power efficiency. Micron will be pushing its technology in the opposite direction.

http://www.extremetech.com/gaming/2...-doubling-gddr5x-for-next-generation-hardware

 

[MrTeal]

Higher density GDDR5 chips should use less power since you need less of them?

I'm not sure if 20nm is smaller than current GDDR5 lithography, but if it is, that should reduce power as well?

They would also provide less bandwidth if you use less of them in most cases. The obvious exception to that is something like the TitanX vs 980Ti, where the TitanX uses 24 4Gb chips in x16 mode vs the 12 4Gb chips operating in x32 mode in the 980Ti. Overall bandwidth stays the same.

 

[xthetenth]

Minor point regarding stacked RAM and shorter PCBs: I'm pretty sure that it's not a coincidence that the Fury Tri-X has such amazing cooler performance and it has a large chunk of the cooler where air is free to go through the heatsink entirely rather than getting blocked by the PCB. So I'd expect shorter PCBs, especially shorter PCBs on cards that still fit high end power circuitry to allow some really stellar aftermarket coolers and benefit people with full sized cases as well as people in the nano use case.

 

[Erenhardt]

I was pondering...
What sort of things come through peoples mind when they talk 'power efficiency', I mean, look.
Kepler - Nvidia's Impressive, Efficient New Graphics Card... That Draws 20 Watts Less.,
Or even maxwell, where the difference is around 80 watts.
Sure it is more efficient (in gaming). Green tech. A lot of people make their/ their friend's/stranger on forum buying decision with CO2/watt calculations in mind etc, etc.

It's all good and noble, but here comes the kicker. It is now not important anymore. I mean, the efficiency was never important to those people.

What matters to them is a difference between their brand and competition in performance/watt.

If they really cared about efficiency, they would ask for HBM in every card possible - in the name of efficiency.

But they ask for something different. They don't want Expensive (assumption) HBM if it is not needed for their brand to have better efficiency than competitor. Already many posts in this thread say exactly this thing, but I will not quote anyone to avoid getting infraction.

So it is not really about watts, CO2, fossil fuels etc. but a made up metric that was lobbied as the most important of all metrics just to have something to show on marketing slides.

Shame to see so many people here are using power efficiency just to sell products (full stop optional), while not giving even a glimpse of thought into actual ecology.

/my_little_rant

 

[Blitzvogel]

Feel how you will about efficiency and performance/watt in desktop parts, but it's a bit step in converging mobile GPUs with desktop, and it has huge implications for Nvidia's SoCs. Even if the Tegra X1 is 20nm, it's under-10W power draw is incredibly impressive considering what it is capable of.

 

[railven]

I was pondering...
What sort of things come through peoples mind when they talk 'power efficiency', I mean, look.
Kepler - Nvidia's Impressive, Efficient New Graphics Card... That Draws 20 Watts Less.,
Or even maxwell, where the difference is around 80 watts.
Sure it is more efficient (in gaming). Green tech. A lot of people make their/ their friend's/stranger on forum buying decision with CO2/watt calculations in mind etc, etc.

It's all good and noble, but here comes the kicker. It is now not important anymore. I mean, the efficiency was never important to those people.

What matters to them is a difference between their brand and competition in performance/watt.

If they really cared about efficiency, they would ask for HBM in every card possible - in the name of efficiency.

But they ask for something different. They don't want Expensive (assumption) HBM if it is not needed for their brand to have better efficiency than competitor. Already many posts in this thread say exactly this thing, but I will not quote anyone to avoid getting infraction.

So it is not really about watts, CO2, fossil fuels etc. but a made up metric that was lobbied as the most important of all metrics just to have something to show on marketing slides.

Shame to see so many people here are using power efficiency just to sell products (full stop optional), while not giving even a glimpse of thought into actual ecology.

/my_little_rant

Odd. Long time AMD fan, you know how it was during Fermi? Every AMD user and their mother were hollering and hooting about power efficiency. That flew out the window the day HD 7970 got taken out by GTX 680. ANd it was a dead talking point when HD 7970 Ghz basically traded it all in to beat GTX 680.


My point is, all these talking points go back and forth ad nauseam. No one here is explicitly asking for GDDR5X. All I see is a bunch of "if HBM is really delayed, this seems like a reasonable stop gap." YOu know what else I don't see? "If HBM is delayed, that's fine, I hope they delay all the cards until HBM is readily available."

I won't be surprised if roles swap and AMD is back to efficient over Nvidia. But I'll laugh internally when now the AMD camp is bragging efficiency again like there is no tomorrow.

 

[Keysplayr]

Whatever each respective GPU company comes up with regarding memory technology used, I'm sure it will be just fine. At least for normal folks.

 

[AtenRa]

Odd. Long time AMD fan, you know how it was during Fermi? Every AMD user and their mother were hollering and hooting about power efficiency. That flew out the window the day HD 7970 got taken out by GTX 680. ANd it was a dead talking point when HD 7970 Ghz basically traded it all in to beat GTX 680.

Sorry but that was not the same as GTX480.

GTX480 came 6 months later and had 100W more power consumption than HD5870.

GTX680 came 3 months after HD7970 and it had only 30W less power consumption.

Then HD7970 GHz came after 3 months of GTX680 and only had 67W higher consumption.

 

[railven]

Sorry but that was not the same as GTX480.

GTX480 came 6 months later and had 100W more power consumption than HD5870.

GTX680 came 3 months after HD7970 and it had only 30W less power consumption.

Then HD7970 GHz came after 3 months of GTX680 and only had 67W higher consumption.

Which cards were faster?

 

[Azix]

7970GHZ was faster. 480 Was faster

 

[AtenRa]

Which cards were faster?

The point is that GTX480 came 6 months after HD5870. It was faster but at the time it was already DOA and 100W higher consumption.

Also, HD5870 was faster and had lower consumption than GTX470. And it was available 6 months earlier.

edit. For a few watts more than the GTX480 (single card) you could get the HD5970 that was way faster in the majority of the games.

 

[Goatsecks]

To my mind the primary reason efficient GPUs are desirable is not because of energy savings but because they are cooler and/or more quiet.

If you are that worried about the environment or saving money PC gaming, on any hardware, is a terrible hobby.

 

[tential]

To my mind the primary reason efficient GPUs are desirable is not because of energy savings but because they are cooler and/or more quiet.

If you are that worried about the environment or saving money PC gaming, on any hardware, is a terrible hobby.

Pretty much the only reason I care about it.
That and to put gpus into smaller cases as I am interested in fitting the hedt platform into as compact of a case as possible water-cooled and combined with a dual gpu aio cooler since amd is insistent on providing this product and since it will be shorter than full length gpus now why not?

Other than that... I'm not sitting there debating 30 watts between 2 cards..

 

[Keysplayr]

Sorry but that was not the same as GTX480.

GTX480 came 6 months later and had 100W more power consumption than HD5870.

GTX680 came 3 months after HD7970 and it had only 30W less power consumption.

Then HD7970 GHz came after 3 months of GTX680 and only had 67W higher consumption.

Really?

You're going to base your thunderous difference in your argument over 30W.
Man, if you can't keep it real and say, "Yeah, I agree. Both the GTX 480 and the 7970 were very thirsty GPUs in comparison to their direct competition." Then say nothing.

 

[Keysplayr]

7970GHZ was faster. 480 Was faster

Selective memory. When 7970GHz launched it was trading blows with the 680. Over time, the 7970 pulled away with driver improvements as a top end card should in comparison to mid range. Rightly so.

 

[Paratus]

Odd. Long time AMD fan, you know how it was during Fermi? Every AMD user and their mother were hollering and hooting about power efficiency. That flew out the window the day HD 7970 got taken out by GTX 680. ANd it was a dead talking point when HD 7970 Ghz basically traded it all in to beat GTX 680.


My point is, all these talking points go back and forth ad nauseam. No one here is explicitly asking for GDDR5X. All I see is a bunch of "if HBM is really delayed, this seems like a reasonable stop gap." YOu know what else I don't see? "If HBM is delayed, that's fine, I hope they delay all the cards until HBM is readily available."

I won't be surprised if roles swap and AMD is back to efficient over Nvidia. But I'll laugh internally when now the AMD camp is bragging efficiency again like there is no tomorrow.

Of course the flip side is also true. With Fermi it was, who cares about efficiency it's the fastest! Then with Kepler the green team was suddenly all about how great efficiency is.

Selective memory. When 7970GHz launched it was trading blows with the 680. Over time, the 7970 pulled away with driver improvements as a top end card should in comparison to mid range. Rightly so.

You make a good point. AMD top end cards age more gracefully then their NV peers. Even when the original prices are similar the AMD cards last longer and are better bang for the buck.

If you have access to a new card each generation then NV makes more sense. If you tend to skip generations then AMD makes more sense, which isn't as good for AMD's bottom line.

Back OT.

If NV is using Gddr5x instead of HBM for the next gen then their first concern is protecting themselves from something, be it supply constraints, prices, etc because from a technical stand point HBM is superior.

I'm sure GDDR5X will be sufficient for NVs next generation if required but by history probably not the one after that.

 

[Azix]

Selective memory. When 7970GHz launched it was trading blows with the 680. Over time, the 7970 pulled away with driver improvements as a top end card should in comparison to mid range. Rightly so.

shhh

trading and faster now = faster anyway

 

[Keysplayr]

Of course the flip side is also true. With Fermi it was, who cares about efficiency it's the fastest! Then with Kepler the green team was suddenly all about how great efficiency is.




You make a good point. AMD top end cards age more gracefully then their NV peers. Even when the original prices are similar the AMD cards last longer and are better bang for the buck.

If you have access to a new card each generation then NV makes more sense. If you tend to skip generations then AMD makes more sense, which isn't as good for AMD's bottom line.

Back OT.

If NV is using Gddr5x instead of HBM for the next gen then their first concern is protecting themselves from something, be it supply constraints, prices, etc because from a technical stand point HBM is superior.

I'm sure GDDR5X will be sufficient for NVs next generation if required but by history probably not the one after that.

What are you talking about?

 

[Keysplayr]

shhh

trading and faster now = faster anyway

I guess you have to put it in a way that you're comfortable with.

 

[AtenRa]

Really?

You're going to base your thunderous difference in your argument over 30W.

So now 30W is not a lot ?? it was huge like mount Olympus when people comparing GTX 680 vs HD7970

You know what ?? I didnt care about GTX480 power usage back in 2010. I still have my GTX480 but HD7970 was nothing like it.

 

[Keysplayr]

So now 30W is not a lot ?? it was huge like mount Olympus when people comparing GTX 680 vs HD7970

You know what ?? I didnt care about GTX480 power usage back in 2010. I still have my GTX480 but HD7970 was nothing like it.

It's not nothing man. But don't act like both those cards weren't thirsty.
It's just important to you that Nvidia used more. That's all.

Be refreshing for a change.

 

[DooKey]

Yet another thread ruined by the video card company fan hordes.

Who gives a crap what kind of VRAM is used in either brand as long as you get top tier performance in games? IF NV can = AMD or vice versa with differing VRAM types what is the difference?

Neither one of these companies cares about you guys. All they want is your $$$$ and being a white knight for either is foolish.

Carry on with the video card company pimping.

 

[dark zero]

Yet another thread ruined by the video card company fan hordes.

Who gives a crap what kind of VRAM is used in either brand as long as you get top tier performance in games? IF NV can = AMD or vice versa with differing VRAM types what is the difference?

Neither one of these companies cares about you guys. All they want is your $$$$ and being a white knight for either is foolish.

Carry on with the video card company pimping.

The problem is that the natural evolution (HBM) is being interrupted and screws the fans very hard

 

[Paratus]

What are you talking about?

You pointed out that the 7970ghz eventually pulled away from the 680 as befitted a high end card vs a midrange card.

Both cards retailed for $499. But the 680 which took the crown at release has failed to keep up.

ergo the 7970ghz which was basically equivalent when released was a better value over time than the 680 per your own comment.

This has been repeated with the 780 when it was the high end gaming card and to a lesser extent the 980.

My other point was if NV is going to use GDDR5X for Pascal instead of HBM2 then in my opinion it's due to a strong business case as opposed to a technical reason.

There's no reason for NV to skip HBM's superior power efficieny, form factor and bandwidth for pascal unless there is a significantly limited supply or a significant risk of delaying Pascal.

AMD doesn't have the luxury of waiting another round for HBM2, NV does.

 

[Azix]

The problem is that the natural evolution (HBM) is being interrupted and screws the fans very hard

It's still early times for HBM. There's bound to be a mix of VRAM tech next year. Then 2017 everything should be HBM