RTX 3080 trumps 2080ti by a whopping margin

I'm thinking the mainstream laptop high end is going to use wording like "Max Q" which will be operating with a lot less consumption. Even the dual power supply clevo laptops would need bigger bricks; not saying no one will go there but it wont be anything close to mainstream. I've actually enjoyed having so called 2k res and my desktop 1080, it's enough power to drive the screen in everything I've tried so far to decent frame rates at that resolution and to my eye looks great on a 17' screen. It also means I can run in the native res at the desktop, not something this old man could do with 4k. I'd want more with a bigger screen but I did a 19" laptop and that was too much trouble to pack around so 17'' is really my target. I think it could be pretty good with whatever their top mobile Max Q'ish part is, 2k, 17", good enough for my laptop. Or at least that's my hope, the parts will have to prove me right before I go there.

 

Yeah the tdp does seem high.. probably a small gain over the mobile 2080 super as they will have too cut the power draw by two thirds

 

Mobile 3080 will be further crippled vs desktop cards. Mobile 3080 will most likely still be capped at 200W TGP.

 

Yeah at the time of posting I was unaware of tdp values and unless Nvidia works some magic this could be a disappointing year

 

I'll need to see how the rtx 3070/3080 actually performs in medium sized laptops before i jump to conclusions that mobile 3000 cards will only be a hair faster than mobile 2000 series cards. Nvidia did claim a 1.9x perf/watt improvement. Even if thats just marketing BS, gpus do get exponentially more efficient as they are underclocked so we may get 70% of the performance at 40% the power draw.

To be honest, I hope the inefficiency of the 3000 series drives up average gpu tdps in laptops as it will force laptop makers to innovate and create really efficient cooling deigns. That and if it forces nvidia to kill off Max-q gpus, good riddance.

That said, I do plan on waiting for 5nm Nvidia Hopper for an efficiency boost. And in the meantime I'll be investing in a VR setup to tide me over.

 

TGP aside, it uses 320bit memory, so I doubt that we'll see this on a laptop. We'll have to wait and see.

 

Ok guys. I'm on a cell during my vacation and I just watched digital Foundry exclusive look at the 3080...brace yourself...I was wrong it's not a 50 percent increase...it's a whopping 60 to 80 percent faster...un fricken believable

 

At the cost of +95W (+42%) TDP vs 2080 Founders Edition. Mobile 3080 will most likely be castrated with same awful TGP as for Mobile 2080 (S). 200W vs. 320W.

Imagine if 3080 Max-Q come at 80/110W vs the real 3080 at 320W

 

Honestly I'm not very impressed. It looks like around 35% performance per watt improvement at around 150W despite a major process node change. Most of that 35% improvement is due to a bigger core allowing for lower voltages to maintain performance. Clocks are down on the entire 30 series indicating a big voltage dropoff. PPW improvement will likely be only around 20% for the 100W range of mobile parts.

This is interesting for the modding scene though. Power limit mods are going to bring major performance increases, but also insane amounts of power draw. With how fast these cards are though you need 4k 120Hz screens to actually need the extra speed.

 

Nvidia is always able to bring desktop laptop closer correct me if I'm wrong but isn't this then the same jump from Maxwell to pascal