Not All RTX 2080s are Created Equal

AlphaAtlas

AlphaAtlas

[H]ard|Gawd
Staff member
Joined
Mar 3, 2018
Messages
1,713
Manufacturers have had some time to stock store shelves and warehouses with Nvidia RTX laptops, but as Techspot pointed out earlier this year, the nomenclature can be very confusing. The laptop "RTX 2080," for example, doesn't have the same performance as the desktop version of RTX 2080, and there are multiple version of the "RTX 2080 Max-Q" with different levels of performance. Hardware Unboxed just tested the performance difference between the various versions.

Check out their video here.

The fact that Nvidia can cram a 545mm^2 GPU into a low-power laptop at all is remarkable, and generally speaking, the RTX chips perform well in their relatively small power envelopes. But as the video points out, be careful if you're in the market for a gaming laptop, as the actual performance level of some RTX GPUs can be difficult to discern.
 
Armenius said:
This is not out of the ordinary for mobile chips.
Click to expand...
Pretty much, they should have kept adding M to the mobile versions. Though, I guess these are indeed close to desktop than older mobile parts which not only had lower clock speed but also fewer cores. AFAIK these all have the same number of cores for equivalent model but in some cases significantly lower gpu/mem clock speed.
 
I actually would be interested in seeing a comparison of fullsize 2080ti showing power, noise, and fps from different release versions with default overclocked and upgraded coolers.
 
The only problem I see here is that there are multiple ways to implement the Q-Max chips and depending on input voltage and cooling they will have wildly different performance but the manufacturer doesn't have to label them differently, so in practice some are running them at 65W and others at 90W but both are labeled as RTX 2080 Q-Max and to know the difference you have to get pretty deep into the specs page.

Researching the product you are about to spend multiple thousands of dollars on easily mitigates this but it will be something to watch for.
 
If you're buying the full-power RTX 2080, you're getting the same chip as a desktop RTX 2080, just clocked a bit lower (which can easily be made equal with a small OC). It's an incredibly powerful chip for a laptop.

It's the low-power RTX 2080 Q-Max that's confusing as there as models that are low power, and models that are ultra low power. I never really understood the q-max line in any case as you're effectively gimping an expensive, high performance chip.
 
harmattan said:
If you're buying the full-power RTX 2080, you're getting the same chip as a desktop RTX 2080, just clocked a bit lower (which can easily be made equal with a small OC). It's an incredibly powerful chip for a laptop.

It's the low-power RTX 2080 Q-Max that's confusing as there as models that are low power, and models that are ultra low power. I never really understood the q-max line in any case as you're effectively gimping an expensive, high performance chip.
Click to expand...
From what I can tell its not a different model for low or high power some manufacturers literally just under-volt and drop the clock speeds to make it fit into the laptops cooling/power envelope.
 
harmattan said:
If you're buying the full-power RTX 2080, you're getting the same chip as a desktop RTX 2080, just clocked a bit lower (which can easily be made equal with a small OC). It's an incredibly powerful chip for a laptop.

It's the low-power RTX 2080 Q-Max that's confusing as there as models that are low power, and models that are ultra low power. I never really understood the q-max line in any case as you're effectively gimping an expensive, high performance chip.
Click to expand...
Perhaps the same physical chip, but the thermal constraints and location of modules etc is undeniable. Most laptop GPU’s are underpowered compared to their similar names desktop full blown dedicated cards.

In some cases a desktop GPU is larger than the entire motherboard including the dgpu in a laptop.

There’s a reason for that. And it shows when you compare performance numbers.
 
In any case, at least nVidia has gotten away from the intentionally confusing mobile chip naming conventions. Remember when a GTX 680m was actually a desktop 670 chip with reduced clocks? When a GTX 880m was a rebadged 780m, which was in fact a desktop GTX 680 chip? Or when a 9800m GT was actually a price-increased 8800m GTX, which was actually close to a 9600 GT desktop chip?

Actually, unlike past mobile cards, the recent Pascal mobile cards, at least at the top end, perform only a hair away from their desktop brethren. My laptop 1080 scores very close to my desktop 1080 did. It doesn't OC as far due to heat and power restrictions (mostly constrained by the PSU), but it's extremely close.
 
bobdabilder said:
I was thinking the issue is that the mobile variants differ.
Click to expand...
Which is nothing new, notebook manufacturers always played around with various nerfed GPUs without labeling them properly. I've had a laptop that I don't know what GPU was in exactly. 3 sources said 3 different things. And not even gpuz or gpushark could give a definite answer.
 
Which doesn't excuse the practice.
M76 said:
Which is nothing new, notebook manufacturers always played around with various nerfed GPUs without labeling them properly. I've had a laptop that I don't know what GPU was in exactly. 3 sources said 3 different things. And not even gpuz or gpushark could give a definite answer.
Click to expand...
 
One of Dell’s Alienware (Area 51M?) uses desktop CPU and GPU, think it’s $5K for 9900K/2080.
 
Armenius said:
This is not out of the ordinary for mobile chips.
Click to expand...

I just remember back in the day when mobile chips were actually distinguishable in that all the mobile GPU variants were all differently numbered and/or had the "m" at the end of the model... for instance the Geforce 950m. Now laptop manufacturers are just calling them the same as their desktop counterparts... of course most of us here know that mobile counterparts to desktop components are not the same (even though they're pretty damn close), but there are too many varying factors. The Max Q naming is obvious, but not even all manufacturers will say "This laptop has a GTX 1060 Max Q variant." Most will just say "GTX 1060." So people should definitely do their research, but I don't see a huge problem here only because there are SO many different laptops with different cooling solutions, different components, different construction materials, varying thicknesses, and different thermal profiles... so the manufacturers I am positive tweak their GPUs to adequately work with their systems which is why reviews are so crucial before buying gaming laptops.
 
Nvidia really needs to stop this shit.

Every model of differing capability needs to have it's own name.

I participate in a Facebook group where the users are generally less informed than us in here. I can't tell you how many people have bought DDR3 versions of the GT1030 and are wondering why the hell they aren't matching the GDDR5 1030 performance figures from reviews.

It's absolutely unforgivable of Nvidia to give vastly different products the same name in order to trick consumers.
 
Zarathustra[H] said:
Nvidia really needs to stop this shit.

Every model of differing capability needs to have it's own name.

I participate in a Facebook group where the users are generally less informed than us in here. I can't tell you how many people have bought DDR3 versions of the GT1030 and are wondering why the hell they aren't matching the GDDR5 1030 performance figures from reviews.

It's absolutely unforgivable of Nvidia to give vastly different products the same name in order to trick consumers.
Click to expand...

Vendors are often more than complicit in pushing the confusion. Dellienware and HP, for example, have been known to sell models with rebranded GPUs right next to each other with the "newer" GPU having a significant price increase. There is no other reason for this than tricking the consumer (by leveraging nV's shady naming conventions) into thinking the rebrand is somehow faster (and worth more).
 
You must log in or register to reply here.
Back
Top