nVidia X80, X80ti and X80 Titan rumoured Specs

[Starelementpoke]

2 minutes ago, Valentyn said:

Yup, HBM2 is going to be for Vega, and their engineer said it would cost too much and isn't ready for mainstream users just yet ( in relation to Polaris ).

Vega doesn´t look to far off from Polaris, doesn´t seem like it´s that bad.

[ThreePinkApples]

So, X80 Titanium and X80 Titan? Makes perfect sense

[-BirdiE-]

1 hour ago, TrigrH said:

if they seriously call it Geforce X im going polaris,

Seems rational...

 

/s

[Valentyn]

Just now, Starelementpoke said:

Vega doesn´t look to far off from Polaris, doesn´t seem like it´s that bad.

Polaris is going to be the "mid range series" to get the ball rolling with 16nm, Vega is their big boys. It's due very late this year, to mid next year.

Think of it as NV Mid Pascal to Big Pascal( next Titan )

[Starelementpoke]

Just now, Valentyn said:

Polaris is going to be the "mid range series" to get the ball rolling with 16nm, Vega is their big boys. It's due very late this year, to mid next year.

Think of it as NV Mid Pascal to Big Pascal( next Titan )

So it´s going from the middleweight title to the heavyweight championships?

[Starelementpoke]

2 minutes ago, -BirdiE- said:

Seems rational...

Sarcasm? 

[Valentyn]

9 minutes ago, Starelementpoke said:

So it´s going from the middleweight title to the heavyweight championships?

Essentially. They can get out the smaller chips quicker since they're much easier to make and take up the majority of sales.

Just like NV's 970 and 980 launched first last time, and some times after that they dropped the Big Champ the titan.

With a new manufacturing process it's even more important to get the mainstream cards rolling out first, especially since they usually slap around the older Champs anyway; but at a better price point.

[LAwLz]

1 hour ago, CommandMan7 said:

So what i'm getting from this is that only their $1000 flagship will even come close to the performance of the Radeon Pro Duo, but even then fall significantly short? IMHO, if you're going to be spending $1000 or more on a graphics card you may as well go for the better performance of the Radeon Pro Duo for the higher cost.

There is more to a card than just FLOPS.

 

There are also a lot of other things you have to take into consideration. Heat output, power usage, features, single GPU vs multi-GPU (so less issues when games does not properly support SLI/Crossfire), the Pro Duo is 1500 dollars and this card might be less (I sure hope it is less), and a bunch of other things.

 

Personally, I would not recommend anyone getting the Pro Duo. I would be very surprised if you can't get a better deal when than it when AMD and Nvidia launch their new cards in just a few months. The previous gen dual GPU card has never been as good of a buy as the current gen high end single GPU card (from a price:performance and user experience standpoint).

[shermantanker]

2 hours ago, CommandMan7 said:

So what i'm getting from this is that only their $1000 flagship will even come close to the performance of the Radeon Pro Duo, but even then fall significantly short? IMHO, if you're going to be spending $1000 or more on a graphics card you may as well go for the better performance of the Radeon Pro Duo for the higher cost.

 

I would rather have a single high powered card.

[MEC-777]

One thing that might be wrong on that first chart is the memory on the X80 will probably be GDDR5X, not GDDR5. And the X80Ti might actually be HBM2 as it will be the same GP100 chip (cut down).

 

Can't wait to see Polaris vs Pascal. Let the battle begin!!!

[Coaxialgamer]

I call BS. The gp 100 titan and X80ti should use the same the same memory , as using gddr5 on the cut down gp100 would mean they would have to use two memory controllers in the gp 100 chip , which costs more , lowers yields and taked more work to develop .

 

Not to mention a 512 bit, 8000mhz memory setup would cause power and heat to go through the roof in terms of memory .

 

 

[Prysin]

2 hours ago, Master Disaster said:

What? The X80 has 1.5 times the Tflop output of the 980TI, its no where even close to "on par" with it, based on Tflops alone the X80 obliterates the 980Ti.

 

The worry about these cards is the shockingly bad amounts of RAM Nvidia decided to go with, 6GB of RAM on the X80 just isn't enough, it should have been at least 8GB.

the more you post, the less seriously i am able to take you.

 

 

[Notional]

3 hours ago, Samfisher said:

I haven't had a single instance of ever exceeding the VRAM of my 970.  Anything that uses more VRAM than that, the GPU core is the limiting factor anyway.

This is and always will be a fallacy. Textures take up a lot of VRAM, but requires very little GPU power. Rise of the Tomb Raider requires more than 4GB of VRAM for very high. Doom seems to need that too for highest setting. So the more the merrier.

 

 

Quote

Those with 4GB GPUs are recommended to use High as VRAM stuttering can be observed on Very High, especially when swapping between gameplay, cutscenes and cinematics, and between gameplay zone transitions.

For a smooth experience with those max-quality, 4K x 4K textures, a 6GB GPU is instead recommended. And to crank things up to 4K (3840x2160), with Very High textures and max settings, we'd recommended GeForce GTX TITAN X GPUs with 12GB of VRAM, as usage can near 10GB over prolonged sessions.

 

[Samfisher]

1 minute ago, Notional said:

This is and always will be a fallacy. Textures take up a lot of VRAM, but requires very little GPU power. Rise of the Tomb Raider requires more than 4GB of VRAM for very high. Doom seems to need that too for highest setting. So the more the merrier.

 

 

 

 

It was more regarding 1080p games ever use more than 4GB VRAM, and if you ever exceed 4GB it's at a higher resolution that the GPU would struggle with anyway.  No mention of texture causing the performance drop, which I know doesn't.

[byalexandr]

3 hours ago, TrigrH said:

lets add "nvidia confidential" to the end of our excel table so it seems more legit....

 

if they seriously call it Geforce X im going polaris,

I was going with Polaris anyways

[Bensemus]

3 hours ago, VagabondWraith said:

If this comes out q4/q1 2017 that falls inline perfectly for my upgrade path.

It falls inline with my upgrade path too. Unfortunately my wallet runs on a different, much slower, schedule

[Notional]

6 minutes ago, Samfisher said:

It was more regarding 1080p games ever use more than 4GB VRAM, and if you ever exceed 4GB it's at a higher resolution that the GPU would struggle with anyway.  No mention of texture causing the performance drop, which I know doesn't.

Remember that textures are spread out on a Z axis too, and that you can zoom in on them. Even 1080p can benefit from 4K textures quite massively. No matter your screen resolution (or virtual super resolution), you should always go for the highest texture setting your VRAM allows.

[jaggysnake57]

ill make my mind up when benchmarks come out

[awesomeness10120]

4 hours ago, Master Disaster said:

I think I heard that GDDR5X was in fairly short supply and production wasn't expected to increase significantly until the summer so although it will certainly be cheaper I dont think it'll be as cheap as we all think it will/should be.

Correct me if I'm wrong, but some Polaris cards will have HBM1. I think it was PcPer that said that.

[totjup5]

5 hours ago, TrigrH said:

lets add "nvidia confidential" to the end of our excel table so it seems more legit....

 

if they seriously call it Geforce X im going polaris,

At least it's better than gtx 1080.....

[SteveGrabowski0]

5 hours ago, Sakkura said:

It says the X80 would be capable of 8.192 TFLOPS where the 980 Ti does 5.632 TFLOPS. Shit just got holy hand grenaded.

hahahahahaha

 

[Sakkura]

2 hours ago, MEC-777 said:

One thing that might be wrong on that first chart is the memory on the X80 will probably be GDDR5X, not GDDR5. And the X80Ti might actually be HBM2 as it will be the same GP100 chip (cut down).

 

Can't wait to see Polaris vs Pascal. Let the battle begin!!!

It says it runs at 8 GT/s, which should be possible with GDDR5. GDDR5X is supposed to enable speeds up in the 14 GT/s range, if you're not pushing that high it's probably better to just stick with good old GDDR5.

[Prysin]

5 hours ago, Dabombinable said:

With HBM, 4GB of vRAM is fine. Its not as effective as the on die cache of a Celeron when compared to a Pentium II with twice the off die cache (same perforamnce at the same clock speeds despite the Celeron having half the cache), but its a lot better than having 4GB GDDR5/GDDR5X.

And there will still be some issues and the fact that not all games have Crossfire profiles, and that some games don't even launch with them (or ones that work as they should).

there are plenty workarounds for missing or broken profiles. Both on consumer and AMDs side.

With DX12 games however, multi adapter is PURELY on game dev side...

 

If you think slow CF/SLI fixes was bad before, just you wait and see how long game devs are gonna wait before solving it.

[Prysin]

9 minutes ago, Sakkura said:

It says it runs at 8 GT/s, which should be possible with GDDR5. GDDR5X is supposed to enable speeds up in the 14 GT/s range, if you're not pushing that high it's probably better to just stick with good old GDDR5.

didnt micro say that 14GT/s wasnt going to happen before 2017 or 2018?

[-BirdiE-]

For all those saying 4GB just isn't enough VRAM.... I believed that too!

 

Until I saw this...

 

http://www.google.com/url?q=http%3A%2F%2Fwww.digitaltrends.com%2Fcomputing%2Famd-fury-x-crossfire-sli-titans%2F&sa=D&sntz=1&usg=AFQjCNF5xwOCBg7ON2VCQY-T9vmzE_KE9A