Leaked slide shows performance boost of Nvidia's RTX 2070/2080 Super mobile GPUs

midian182

Posts: 9,744   +121
Staff member
Something to look forward to: Fans of gaming laptops could soon have something to celebrate, with Nvidia’s RTX 2070 Super and RTX 2080 Super mobile cards seemingly on their way and offering a significant boost over their non-Super counterparts.

The news comes from PCGamesN, which got its hands on a confidential Nvidia slide that compares the performance of the RTX 2080 Super and 2070 Super laptop graphics cards against the rest of the company’s current mobile GPUs.

The slide comes from an anonymous source, and, assuming it’s genuine, shows the 2070 Super is faster than the standard RTX 2070, as one would expect, and around 5.8 times faster than the GTX 1050 with DLSS turned on.

The chart also shows the RTX 2080 Super sits comfortably at the top of the pile and is roughly 6.7 times faster than a 1050. There’s no standard RTX 2080 mobile card on the chart, which PCGamesN theorizes is due to it offering almost identical performance to the 2070 Super that uses the same TU104 GPU.

The chart’s performance levels are taken from three games: Control, Wolfenstein: Youngblood, and Deliver Us The Moon, all of which were running at 1080p with the setting at max. We also see that the benchmarks were run on an unreleased 10th-gen Intel Comet Lake-H CPU (reportedly an i7) along with 16GB of RAM and Windows 10.

The slide is reportedly part of a release kit for laptop manufacturers, so it looks as if machines packing these Super cards will arrive soon, possibly as early as this month.

Permalink to story.

 
How does that chart say the 2070 Super is 50% faster than the 2070 or the 2080 Super being 50% faster than the 2070 Super? The 2070 Super would have to be off the chart (over 7.5X) for it to be 50% faster than the 2070 for DLSS.
 
Very unclear article and explanation of what that chart is showing.

My thought would be these are performance of new 'Super Laptop' parts compared to the Max-Q laptop cards - surely not against the desktop parts????
 
All I know is my 1070 still kicks *** in my laptop, quite possibly the best performance to price ratio ive ever seen for 1080p gaming. I use it more than my desktop and I have a 2080 Ti in that thing.
 
Very unclear article and explanation of what that chart is showing.

My thought would be these are performance of new 'Super Laptop' parts compared to the Max-Q laptop cards - surely not against the desktop parts????

I agree. The highest 2080 "Super" TDP I am aware of is 115 watts, so yeah I would put my MSI GT76 Titan with a 205 watt 2080 up against one all day.
 
Very unclear article and explanation of what that chart is showing.

My thought would be these are performance of new 'Super Laptop' parts compared to the Max-Q laptop cards - surely not against the desktop parts????
It's showing the performance compared to the 1050 - that's why the graph shows "1X" for that part...

Wish they'd use FPS or a benchmark score instead, but I guess they want to keep that close to the vest.

I think it's pretty obvious that these are all laptop cards - this might be another reason there is no 2080, as there was quite the difference between 2080 Max-Q and "regular" laptop 2080s which Nvidia probably didn't want to highlight in print...
 
Very unclear article and explanation of what that chart is showing.

My thought would be these are performance of new 'Super Laptop' parts compared to the Max-Q laptop cards - surely not against the desktop parts????

It is indeed very unclear and misleading.

Let me explain

The X axis is "relative performance". This means performance against a control. In this case that is a GTX 1050. The values on the X axis are 1X, 2X, 3X, ect.

In other words 1X = GTX 1050 performance while 5X = 5 times GTX 1050 performance.

In addition, Nvidia also tacks on extra "DLSS On" performance to it's 20xx series (which is not supported on the 1050) from a list of games it cherry picked.

For starters, using a low end card like a 1050 as a base index is simply misleading. The 1050 launched at about $100, 1/4rth the price of the 2060.

Second, thanks to Nvidia's card naming, we have no idea which parts are mobile and which are not. The fact that mobile and desktop chips share the same name leaves things like this open to manipulation. I'm not saying this is happening, only that it's extremely confusing and disingenuous.

Third, Nvidia decided to show RTX cards in the best possible light by enabling DLSS on newer cards and only testing on the few DLSS supported games. It's not representative of the performance you will see at large nor is it apples to apples as the 1050 does not support DLSS.
 
It is indeed very unclear and misleading.

Let me explain

The X axis is "relative performance". This means performance against a control. In this case that is a GTX 1050. The values on the X axis are 1X, 2X, 3X, ect.

In other words 1X = GTX 1050 performance while 5X = 5 times GTX 1050 performance.

In addition, Nvidia also tacks on extra "DLSS On" performance to it's 20xx series (which is not supported on the 1050) from a list of games it cherry picked.

For starters, using a low end card like a 1050 as a base index is simply misleading. The 1050 launched at about $100, 1/4rth the price of the 2060.

Second, thanks to Nvidia's card naming, we have no idea which parts are mobile and which are not. The fact that mobile and desktop chips share the same name leaves things like this open to manipulation. I'm not saying this is happening, only that it's extremely confusing and disingenuous.

Third, Nvidia decided to show RTX cards in the best possible light by enabling DLSS on newer cards and only testing on the few DLSS supported games. It's not representative of the performance you will see at large nor is it apples to apples as the 1050 does not support DLSS.
Yes I understood that they are relative to the 1050, but what are they? I'm guessing they are all laptop - ie MAX-Q cards which are very different to true 2070 and 2060s?
 
Yes I understood that they are relative to the 1050, but what are they? I'm guessing they are all laptop - ie MAX-Q cards which are very different to true 2070 and 2060s?

We know for sure that the two new SUPER cards are mobile but given that Nvidia no longer denotes mobile in model names, it's impossible to tell if the other cards are their mobile variants or not.

A mobile RTX 2070 is called the exact same thing as a desktop RTX 2070.

Nvidia has done a very poor job with model names in general recently and I have no idea why they made the change.
 
We know for sure that the two new SUPER cards are mobile but given that Nvidia no longer denotes mobile in model names, it's impossible to tell if the other cards are their mobile variants or not.

A mobile RTX 2070 is called the exact same thing as a desktop RTX 2070.

Nvidia has done a very poor job with model names in general recently and I have no idea why they made the change.
It's worse than that - as there are the Max-Q parts too which also are intended for laptops etc. Both the mobile and max-q parts are unsurprisingly considerably slower than the proper desktop parts.

A really good explanation with benchmark comparisons:

https://www.notebookcheck.net/GeFor...e-RTX-2080-Max-Q_9551_9611_9621.247598.0.html
 
A lot of people I know come to me asking about laptop choices and I always steer them towards the 2060 equipped HP Omen or Alienware.

I don't like buying old hardware because it's usually less efficient even if it claims to be more powerful, and as you've seen, the newer product is more likely to be supported longer.

There's virtually no game you can't play comfortably on a Core i7 laptop with a 1060 nowadays. A Core i7 with a 1660 will cost you around $799 -$999 while a 2060 will run about $999-$1400.

The 2070 and 2080, I honestly don't see spending more than $2000 for. The 2080 models right now are extremely high in price and about to be outdated by Ampere GPU.

 
Back