Nvidia Ada RTX 40 Laptop GPU analysis: 175 W RTX 4090 Laptop GPU surpasses a 350 W RTX 3090 in performance and power efficiency

[Redaktion]

Laptops powered by Nvidia's Ada RTX 40 Laptop GPUs have started to become available, presently with flagship GPUs like the RTX 4080 and RTX 4090 Laptop GPU. We have data from three such devices — the MSI Titan GT77 HX 13V and the Schenker XMG Neo 16 both featuring the RTX 4090 and the Razer Blade 16 with the RTX 4080 — and evaluate how Nvidia's latest mobile offerings perform in comparison with other desktop and mobile brethren.

https://www.notebookcheck.net/Nvidia-Ada-RTX-40-Laptop-GPU-analysis-175-W-RTX-4090-Laptop-GPU-surpasses-a-350-W-RTX-3090-in-performance-and-power-efficiency.691313.0.html

[NikoB]

And for me, the main conclusion is again that playing with such wild noise on "gaming" laptops has become even more impossible than before. And their consumption has become simply wild, unthinkable 10 years ago.

And everything is also clear that it is 2 times more profitable to buy a desktop with 4090 than a "gaming" laptop. It will be quieter at times and much faster, at least 1.5 times faster on the video card. And that decides everything.

Why do people (except for shift workers and sailors) buy "gaming" laptops, wasting a lot of money and doing sadomasosism "playing" with wild much more noise and with 1.5 times less fps? I don't understand...

---
The best that the harfware x86 industry could do today is an optical link with a cable length of up to 50m and a bandwidth of 50GByte/s+. To put away these insanely consuming and noisy videocards in the back room or another room, connecting them to the laptop via optical eGPU and forgetting about the wild noise, although the processor requires same cooling...

It's better to put the entire laptop or desktop in the back room. It will leave on the table in living room only a monitor, keyboard, mouse and an optical cable with a bandwidth of more than 50 GByte/s+ plugged into a docking station for peripherals.

[RobertJasiek]

I consider RTX desktop or RTX notebook for use at home in possibly different rooms and the garden. The most relevant aspects are price, noise and sufficient speed (about 3080 10GB desktop, 4080 Laptop or faster). Desktop noise is 37dB+, notebook noise is 39dB+. Both are acceptable. This leaves price as the decisive factor. A notebook may be a bit more expensive but not exaggerated. Desktop dGPUs are overpriced so I wait. Secondary aspects can influence my decision.

Notebooks are reasonably priced at times. Nvidia's greed has taught us patience so I am patient;)

Notebooks are not per se bad but one must avoid impulse purchases.

[RobertJasiek]

"the RTX 4090, RTX 4080, and the RTX 4070 Ti desktop cards offered significant performance and efficiency uplifts compared to their Turing counterparts."

Don't you mean Ampere?

[LL]

For me the relevant aspects are performance, price, screen (bright enough to let me go to the balcony)  and noise.
I don't mind noise up to certain level when it is pushing. But when i am listening to music, seeing a movie, navigating normal web stuff - youtube etc -, office work the computer should be absolute silence. Zero fans. it is intolerable that super processors are unable to work silently in several models for work that requires small performance

[LL]

Note for NBC.
it would be interesting to compare the efficiency of 4090 175 watts models with the 4080 also 175 watts models.

[LL]

Another point on calculating energy efficiency is the time spent to do a job.

It was unfortunate that NBC did put a Blender GPU test or similar.
For example a 50w GPU that takes 8 hours to do a render is significantly less power efficient than a 200w GPU that takes 1 hour.

[Leventebandi]

Was the Neo 16 tested with the Oasis attached btw?
I hope in thestandalone review, you guys will include how does sound, thermals and performance can differ with the liquid cooling attached:)

[Vaidyanathan]

Was the Neo 16 tested with the Oasis attached btw?
I hope in thestandalone review, you guys will include how does sound, thermals and performance can differ with the liquid cooling attached:)

Yes. It was tested with the Oasis attachment.
Will pass on the feedback regarding comparisons with and without the Oasis. Hopefully, whoever is reviewing the device will be able to accommodate the request :)

[Vaidyanathan]

Another point on calculating energy efficiency is the time spent to do a job.

It was unfortunate that NBC did put a Blender GPU test or similar.
For example a 50w GPU that takes 8 hours to do a render is significantly less power efficient than a 200w GPU that takes 1 hour.

That's a great suggestion, LL. Will see if we can incorporate that in our test routines going forward.

[Vaidyanathan]

Note for NBC.
it would be interesting to compare the efficiency of 4090 175 watts models with the 4080 also 175 watts models.

Do you mean including desktop cards limited to 175 W? Or comparing two laptops with the aforementioned GPUs? If it's the latter, it should be already available in the compare. You can also add any configuration to the compare if it's already in our database.

[Vaidyanathan]

"the RTX 4090, RTX 4080, and the RTX 4070 Ti desktop cards offered significant performance and efficiency uplifts compared to their Turing counterparts."

Don't you mean Ampere?

Ah yes. Thanks for spotting it. Corrected :)

[LL]

Thanks you Vaidyanathan for all your replies.

Or comparing two laptops with the aforementioned GPUs?

I mean since the laptop 4080 and 4090 have same maximum TDP 175w it would be interesting to check if 4090 performance came from spending more energy (despite same TDP) or because it is more efficient than 4080 and try to find the reason for that advantage.
To do that would need 2 models with 4090 and 4080 and the 175w TDP for both.
Probably the best way to compare should be with same model line that have both configurations.

[Vaidyanathan]

Thanks you Vaidyanathan for all your replies.

Or comparing two laptops with the aforementioned GPUs?

I mean since the laptop 4080 and 4090 have same maximum TDP 175w it would be interesting to check if 4090 performance came from spending more energy (despite same TDP) or because it is more efficient than 4080 and try to find the reason for that advantage.
To do that would need 2 models with 4090 and 4080 and the 175w TDP for both.
Probably the best way to compare should be with same model line that have both configurations.

Ah that way. At the moment, there aren't too many 4090/4080 samples floating around. Might be possible once we get hands on more devices. :)
Right now, my guess is as good as yours... the 4090 has more hardware than the 4080 and still can work in the same TDP envelope, so it is bound to be more efficient.

I think on the desktop Founders Edition cards, the RTX 4090 consumes 19% higher power at stock settings compared to the 4080 but is also able to offer 34% higher fps in Witcher 3 4K Ultra.

However, the 4090 mobile is actually the 4080 desktop while the 4080 mobile is a cut-down 4070 Ti (duh!). So taking that into account on the desktop side, you get a 6% higher power consumption for a 30% higher fps in Witcher 3 4K.

Of course, the memory subsystem is vastly different on desktop boards which needs to be factored in as well. But I expect this trend to continue on mobile as well, although we need more devices at similar TDPs for a better idea :)

[RobertJasiek]

At the moment, there aren't too many 4090/4080 samples floating around.

Do you know why? I thought there would also be some Aliens and Legions. So far only Titan and Razer 16 plus a few reviewers each got one other different notebook. But some got several presamples, too, which are hardly reviewed yet. I expected hundreds of reviews on Feburary 7 but only a dozen appeared. Do some manufacturers only send tiny numbers of review samples to carefully selected reviewers, which would not reveal the real weaknesses? It was completely different for desktop graphics cards: countless "reviews".