2080ti super

littlepuppy · 6 Dec 2019 at 21:37

Anyone know the official release date. A friend at work who is rarely wrong is saying q1 2020... I was under impression there was not going to be one but he is adamant it’s coming... anyone here heard any rumours?

robfosters · 6 Dec 2019 at 21:39

No one knows if there’s even gonna be one yet. It’s all rumours and pie in the sky atm.

Kaapstad · 7 Dec 2019 at 06:59

RTX Titan and its been available for ages.

robfosters · 7 Dec 2019 at 09:26

Kaapstad said:
RTX Titan and its been available for ages.

Not really. The idea of super cards is faster versions at the price of the originals or just below.

Titan is much more expensive than the 2080Ti. If they rebadged the Titan as a 2080Ti super and dropped it to around a grand, I’d be well tempted.

IT Troll · 7 Dec 2019 at 09:54

At this stage you have to make your own. Just requires a 2080Ti and two donor 2080 Supers for the faster memory.

Kaapstad · 7 Dec 2019 at 10:58

robfosters said:
Not really. The idea of super cards is faster versions at the price of the originals or just below.

Titan is much more expensive than the 2080Ti. If they rebadged the Titan as a 2080Ti super and dropped it to around a grand, I’d be well tempted.

The 2080 Ti Super exists in the form of the RTX Titan.

You are just arguing about the price.

I am not sure why NVidia even make a big thing about memory clock speeds on the Super edition cards.

An RTX Titan can run all 24gb of its Samsung memory at better than Super speeds all day.

FlyingScotsman · 7 Dec 2019 at 14:47

Kaapstad said:
The 2080 Ti Super exists in the form of the RTX Titan.

You are just arguing about the price.

I am not sure why NVidia even make a big thing about memory clock speeds on the Super edition cards.

An RTX Titan can run all 24gb of its Samsung memory at better than Super speeds all day.

Why is your pcie bus running at x16 2.0 Kaap?

Kaapstad · 7 Dec 2019 at 20:43

FlyingScotsman said:
Why is your pcie bus running at x16 2.0 Kaap?

I forgot to set it to PCI-E 3.0 in the bios but fixed it shortly after that pic was taken.

4K8KW10 · 7 Dec 2019 at 21:03

littlepuppy said:
Anyone know the official release date. A friend at work who is rarely wrong is saying q1 2020... I was under impression there was not going to be one but he is adamant it’s coming... anyone here heard any rumours?

So theoretically, we could see this 2080 Ti Super launched in the first few months of 2020 – perhaps as a pre-emptive strike before AMD releases its supposed ‘Nvidia killer’ high-end Navi GPU which is rumored to be due mid-2020 (and could be revealed as soon as CES).

https://www.techradar.com/news/nvidia-rtx-2080-ti-super-is-coming-and-it-could-be-a-monster-gpu

...there is still continuous improvement that we can do, both at the current node or in terms of the near future.
...

Interestingly, the company also mentioned that there is work that they can do on the existing 12NFF front.

https://wccftech.com/nvidia-we-would-like-to-surprise-everyone-with-7nm-gpu-announcement

Kaapstad · 8 Dec 2019 at 06:22

The difference in performance between the 2080 Ti and RTX Titan is about 7%.

How on earth are NVidia supposed to fit a new Super 2080 Ti into that space and what would be the point.

Next year will be all about NVidia on 7nm, why would they waste time and resources on anything else.

4K8KW10 · 8 Dec 2019 at 09:28

Kaapstad said:
The difference in performance between the 2080 Ti and RTX Titan is about 7%.

How on earth are NVidia supposed to fit a new Super 2080 Ti into that space and what would be the point.

Next year will be all about NVidia on 7nm, why would they waste time and resources on anything else.

Well, the current TU chips are very large - one is 754 mm^2, another is 545 mm^2.
If they want to at least equal their performance, they must go very big on the new N7 node, which could prove counter-productive and not efficient.

I have been saying for months that the problem is that the N12 chips are maxed out, so they would need to max the N7 node out to gain some advantage.
Otherwise, the move onto N7 will be waste of time.

TNA · 8 Dec 2019 at 10:05

4K8KW10 said:
Well, the current TU chips are very large - one is 754 mm^2, another is 545 mm^2.
If they want to at least equal their performance, they must go very big on the new N7 node, which could prove counter-productive and not efficient.

I have been saying for months that the problem is that the N12 chips are maxed out, so they would need to max the N7 node out to gain some advantage.
Otherwise, the move onto N7 will be waste of time.

I'll quote this one to so you can see you are wrong again soon.

You are basically saying there is no such thing as performance gains from new architecture. Maybe you have heard of Kepler and Maxwell architectures which were both on 28nm and there was a very nice improvement in performance...

Saying moving on to N7 is a waste of time in any fashion shows a severe lack of knowledge once again.

koooowweeee · 8 Dec 2019 at 12:01

Oh the claws are out.

bru · 8 Dec 2019 at 12:36

Look at Fuji, look at Vega.
Look at polaris, look at RDNA.

Well who'd of thought it, performance improvements with new architecture, combined with a die shrink.

To think that NVIDIA won't do the same, just as they have pretty much with every generation, is certainly short sighted.

FlyingScotsman · 8 Dec 2019 at 13:00

Kaapstad said:
I forgot to set it to PCI-E 3.0 in the bios but fixed it shortly after that pic was taken.

Cool mate.

Kaapstad · 8 Dec 2019 at 13:19

4K8KW10 said:
Well, the current TU chips are very large - one is 754 mm^2, another is 545 mm^2.
If they want to at least equal their performance, they must go very big on the new N7 node, which could prove counter-productive and not efficient.

I have been saying for months that the problem is that the N12 chips are maxed out, so they would need to max the N7 node out to gain some advantage.
Otherwise, the move onto N7 will be waste of time.

Indeed the TU102 chips are huge, this is precisely why NVidia need to get onto 7nm ASAP.

7nm will give the same performance as 12nm for a lot less silicon real estate.

On 7nm I don't think we will see chips as big as the TU102 but they will be more powerful due to the higher transistor count that will be possible.

4K8KW10 · 8 Dec 2019 at 14:10

Kaapstad said:
Indeed the TU102 chips are huge, this is precisely why NVidia need to get onto 7nm ASAP.

7nm will give the same performance as 12nm for a lot less silicon real estate.

On 7nm I don't think we will see chips as big as the TU102 but they will be more powerful due to the higher transistor count that will be possible.

The production should also be economically reasonable. N7 wafers are pretty expensive now, by the looks of it:

https://www.techdesignforums.com/practice/technique/assessing-the-true-cost-of-node-transitions/

Woodsta888 · 8 Dec 2019 at 16:55

OK google

TNA · 8 Dec 2019 at 17:19

Woodsta888 said:
OK google

Lol.

Dicehunter · 8 Dec 2019 at 18:15

IF a 2080 Ti Super launched, First it would be a pointless release, And secondly it would be a pointless purchase as Ampere would be out a few months later with increased performance at or under the price of current gen.