Shekel pinching has never been more obvious!
I’ve gone on plenty regarding NVIDIA’s mobile segment for their GeForce RTX 30-series such as the abysmal configurations of the GeForce RTX 3080 and 3060 more specifically and how NVIDIA has outright abandoned the naming scheme of their Max-P and Max-Q SKU’s which separate the TGP gimped trash from the other, however the case that falls upon Ampere mobile GPUs is even worse.
With Max-Q and Max-P GPU’s simply getting the axe NVIDIA put as much emphasis onto OEM partners to be as transparent as possible regarding power and frequency limitations of their “Laptop GPUs” as they have selling desktop componentry for MSRP.
Review embargos for RTX 30 mobile devices has been lifted, meaning that now you can finally view a plethora of reviews for laptops and notebooks that contain the new exciting RTX GPUs, however unsurprisingly the OEM partners have simply ignored NVIDIA’s suggestion to advertise and describe the respective TGP values and frequency of RTX 30-series “Laptop GPUs” on their websites.
So much so that ComputerBase has reported that across the new RTX 3000 mobile devices that there are a staggering 28 different variations of Laptop GPUs inside them, from the three GPUs made available all of them offering different TGP limitations and therefore performance figures.
I shit you not, across the RTX 3080, 3070 and 3060’s mobile variants, laptop manufacturers have in total 28 different configurations in total with differing TGP limitations and therefore overall performance.
Thus making NVIDIA’s own specifications and or performance figures listed on their website completely redundant which it merely depicts a fucking range of TGP and frequency anyway.
Using a compiled sheet from the data collected by the ComputerBase article, kindly provided by Videocardz we’re able to see the differences in effective performance and varying TGP values across NVIDIA’s mobile lineup.
Of course the standard desktop RTX 3070 and RTX 3080’s are both complete write-offs, despite having fewer SM’s compared with the mobile RTX 3080 (40 vs 48) the desktop RTX 3070 is bound to pull ahead every single time. That’s just the way how it is when you’re buying mobility grade componentry.
However for the mobile segment at least, OEM manufacturers have provided us with 11 different configurations of the RTX 3080 “Laptop GPU”, three of which are the unofficial Max-Q variants which are easily distinguished via their seriously low base and boost clock speeds and most importantly the much slower 12 Gbps GDDR6 memory.
There are six different variants of the RTX 3070 mobility graphics card two, with an equal amount of them also being the much inferior Max-Q variety while the RTX 3060 Laptop GPU offers twelve different flavors in total with just THREE of which are Max-Q GPUs.
That’s just fucking appalling, and of course there’s no real indication from their respective manufacturers regarding these TGP values, frequencies and rated memory speeds either. They all fall under the banner of being an RTX 3000 Laptop GPU.
Among the list however you’d be able to spot many cases in which an RTX 3070 mobile devices are superior to that of the Max-Q variants of the RTX 3080 mobility units in theoretical FP32 TFLOPs, with three Max-Q variants of the RTX 3080 Laptop GPU offering different TGP limitations ranging from 80W to 90W it’s actually also appalling that the slower 80W RTX 3080 also ends up being slower than standard RTX 3060 mobile GPUs that range from 100-watts and above.
That’s right. Your undisclosed RTX 3080 mobile graphics card could end up being slower than an RTX 3060 laptop.
This shit is like a fucking lottery and you’re bound to lose more often than not because even the RTX 3070 Max-Q graphics cards perform even worse. Ironically the one with the 90W TGP has pretty much identical FP32 throughput as the RTX 3080 Max-Q with the 80W TGP.
Plethora of RTX 3060 Laptop GPUs are faster than the 3070 Max-Q variants, with the exception of the 80W and 85W limits on standard “Max-P” RTX 3060 graphics cards which perform like absolute dogshit second to none above the Max-Q variants with slower 12 Gbps memory which is sure to have an impact on actual gaming performance as well.
These figures aren’t entirely accurate as there’s still a variance of anywhere up to a maximum of 20W that these mobility GPUs can potentially siphon from the CPU in mobile devices thanks to its PPAB range.
The RTX 3080 Max-Q at 80W has a throughput of 9.6 TFLOPs, the RTX 3070 Max-Q at 90W offers 9.5 TFLOPs while a standard RTX 3060 at 95W can offer 9.3 TFLOPs.
That is fucking depressing.
Because NVIDIA has scrapped the Max-Q and Max-P denunciations in a clearly anti-consumerist act the mere suggestion for mobile manufacturers to list and describe the specifics of the Laptop GPU they’re actually using is downright hysterical. They’re not going to properly describe anything because otherwise you wouldn’t buy it now would you?
You spend top dollar on buying the latest and greatest portable devices and you’re given shit like this, where your RTX 3060’s and 3070’s can potentially be faster than the supposed “flagship” RTX 3080 but of course the cheaper variants will be offering low TGP trash that offers performance literally miles off its desktop comparison and more in lines with shit from yesteryear.
And the goyim are none the wiser to this scheme. Just another reason why gaming laptops are fucking garbage products.