NVIDIA to Drop Max-Q and Max-P Differentiators in Mobile GPU Specifications

[AleksandarK]

NVIDIA has recently introduced its 3000 series of Ampere graphics cards designed for mobile/laptop devices. And usually, these GPUs in the past few years have been divided into two configurations: Max-P and Max-Q. The Max-P variant was a maximum performance configuration meant for more power usage and higher temperatures, representing a standard GPU configuration. The Max-Q design was, according to NVIDIA, "a system-wide approach to deliver high performance in thin and light gaming laptops. Every aspect of the laptop, chip, software, PCB design, power delivery, and thermals, are optimized for power and performance." Meaning that the Max-Q variants are more TGP limited compared to the Max-P configuration.


Update 23rd of January 11:35 UTC: NVIDIA spokesperson told Tom's Hardware that: "No, Max-Q branding is not going away. When we originally introduced Max-Q back in 2017, the brand was initially used in GPU naming since Max-Q referred to the GPU TGP only. Today, 3rd Generation Max-Q is broader, and is a holistic set of platform technologies and design approach to building powerful and thin laptops. In addition, to be more transparent about a laptop's exact capabilities, RTX 30 Series laptops now show more information than ever, listing exact TGP, clocks and features supported. You will find this in the control panel which now reports maximum power (TGP+Boost), and support for key features including Dynamic Boost 2, WhisperMode 2, Advanced Optimus, and others, all of which fall under the Max-Q umbrella. We strongly encourage OEMs to list clocks and other technologies a laptop supports, including Advanced Optimus, Dynamic Boost 2, and more. Ultimately, like all laptop features and specs, it is up to the OEM to market what their particular laptop configuration supports.)"


Today, according to Notebookcheck, NVIDIA has decided to drop these naming differentiators in the products, where customers are now unable to know whatever their product uses the high-TGP or low-TGP configuration of a specific GPU SKU. From now on, laptop makers will no longer list the GPU configuration in their specifications and will just list a GPU model, without any explanation whatever it is a Max-P or a Max-Q arrangement. Originally the company did announce both types of SKUs for the 3000 series GPU based laptops. However, this change has happened recently, and now when looking at the laptop specifications, the GPU variant is not listed. Below you can check out the slide posted on Chinese social media spotted by @9550pro on Twitter, showing the differences in Max-P and Max-Q offerings. Laptop OEMs have already excluded the Max-P specification from the specification list.


View at TechPowerUp Main Site

 

[P4-630]

So read reviews first before buying a gaming laptop.

 

[_Flare]

sure, but what if reviewers get a high-TGP sample and enduserproduction gets only low-TGP?

 

[windwhirl]

So read reviews first before buying a gaming laptop.

I get the feeling that won't be enough either, but maybe I'm just being pessimistic

 

[SL2]

As if that's would make much difference anyway. Each model comes in lots of different power configurations, like the mobile RTX 2060 that's either 60, 70, 80, 90.... W. Don't even remember the numbers, don't care.

 

[nguyen]

Just let them GPUs run until they thermal throttle, that's the way to go. This make sense since AMD Cezanne pull much less juice than Intel --> more power available to GPU.

 

[londiste]

Max-P was never a Nvidia term, only Max-Q was used for lowered power limit GPUs. The problem with Max-Q always was the same as mobile GPUs in general - power limit has a range plus it all relies on laptop manufacturer providing good enough cooling.

I do not even know what the good solution to this would be. Max-Q seemed like a good start but it didn't seem to end up useful for actually determining the performance level at glance. There are laptops with Max-Q GPU that runs as fast or faster than a crap laptop "Max-P" laptop. And there are Max-Q laptops that have minimum power limit and are very noticeably slow.

Obviously laptop OEMs request this kind of sliding scale of power limits which meanst the solution should to come from their side. But at the same time they are clearly not interested in clear specifications as we know. Some arm twisting from GPU OEM like Nvidia would be the other way but that isn't exactly a good thing either.

 

[MaMoo]

But I would like to know if my laptop has a PP or a QQ. Very important.

 

[ShurikN]

Max-P was never a Nvidia term, only Max-Q was used for lowered power limit GPUs. The problem with Max-Q always was the same as mobile GPUs in general - power limit has a range plus it all relies on laptop manufacturer providing good enough cooling.

I do not even know what the good solution to this is. Max-Q seemed like a good start but it didn't seem to end up useful for actually determining the performance level at glance. There are laptops with Max-Q GPU that runs as fast or faster than a crap laptop "Max-P" laptop. And there are Max-Q laptops that have minimum power limit and are very noticeably slow.

The only solution that is feasible is laptop manufacturers putting down the available graphics power among the specs.
For example RTX 3070 90W or RTX 3080 80W... something like that

 

[londiste]

The only solution that is feasible is laptop manufacturers putting down the available graphics power among the specs.
For example RTX 3070 90W or RTX 3080 80W... something like that

Absolutely. And many manufacturers are doing that already which is nice. Not all though.

 

[owen10578]

Wow that is pretty shitty...

 

[Verpal]

Best solution might be NVIDIA mandating that OEM have to include TGP, CUDA core count, clock speed and VRAM of mobile GPU in spec sheet, this way interested customer can still can kinda have a grasp on projected performance, whilst OEM can keep the SKU name itself simple for most uninformed customer.

Though, it is in interest of NVIDIA to make sure actual performance is obsfucated, and customer ought to only rely on SKU name to determine relative performance.

 

[Dammeron]

Kinda reminds me of MX150, where some laptops got a shitty version (eg. Thinkpads), but without checking it through GPU-z there was no way to know, which model got which GPU.

 

[windwhirl]

The only solution that is feasible is laptop manufacturers putting down the available graphics power among the specs.
For example RTX 3070 90W or RTX 3080 80W... something like that

Best solution might be NVIDIA mandating that OEM have to include TGP, CUDA core count, clock speed and VRAM of mobile GPU in spec sheet, this way interested customer can still can kinda have a grasp on projected performance, whilst OEM can keep the SKU name itself simple for most uninformed customer.

I can agree with that, but it brings the question of whether the majority of consumers could make sense of the specs.

I think I would just offer the one GPU per power target, with no configurable TGP. Say, you want a GPU that never goes beyond 115 W? Have a 3060. 125 W? 3070, and so on and so forth.

But, alas, that's probably not something Nvidia would care or want to do.

 

[londiste]

I think I would just offer the one GPU per power target, with no configurable TGP. Say, you want a GPU that never goes beyond 115 W? Have a 3060. 125 W? 3070, and so on and so forth.

Just think of what that lineup would look like? There are multiple GPUs in generation and power limit is only one of the factors. For example, look at specs of Nvidia's 30-series lineup:

NVIDIA GeForce RTX 40 Series Laptops

The world’s fastest laptops for gamers and creators.

www.nvidia.com

- RTX 3060 Laptop: 60-115W
- RTX 3070 Laptop: 80-125W
- RTX 3080 Laptop: 80-150+W
3080 at 80W is faster than 3070 at 80W and probably matches 3070 at something like 100W.

This is not just an Nvidia thing either. AMD has "up to" values for mobile GPUs and no stated TDP/TGP/powerlimit, but a note that configurations will vary.

 

[rethcirE]

The choice of 115W over 80 or 90W RTX versions was paramount in my laptop choice. Having to dig or hunt for this information seems a disservice to the end user/customer. I somewhat doubt Nvidia would willingly withhold specification details and then turn to OEM’s and require the information they chose to withhold be made easily available to the market. The specifications are being buried because they matter a great deal, and they inevitably sell less lower spec models that few people desire. If you can fudge the numbers and make the cards all appear the same, they will sell more lower spec units because uninformed buyers won’t even know the difference.

 

[londiste]

I somewhat doubt Nvidia would willingly withhold specification details and then turn to OEM’s and require the information they chose to withhold be made easily available to the market.

Mobile GPUs have variable specifications. This is not to fudge things but when building a laptop you as a laptop OEM may have genuine reasons to run a GPU at 80W as opposed to 115W. This can be due to chassis, cooling or something else. Worse, as an OEM you may want to have GPU run at 80W, 90W, 100W, 110W, 115W in your laptop lineup. These seemingly small differences do matter when you are constrained as laptops are.

 

[1d10t]

Drop confusion and start guessing
It's almost like nVidia give full authority for OEM and ODM to toying around with specs, but I say this could be harmful to customer. As an example I had notebook "L" and notebook "A" within same price range and I found out that one of them is performing not quite similar to the others. Fortunately , tiny app GPUz came to rescue and confirmed, both has spec as stated.
With MaxQ or MaxP annotations there's already alot of confussion, imagine when there's is none.

 

[zo0lykas]

WHY are they doing everything so complicated?

same to intel with they cpu naming...

 

[Deleted member 185088]

Pascal mobile GPUs gave hope, then the whole stupid max q rubbish started. Now we are back to the dark ages of m GPUs.

 

[bug]

It matters little what manufacturers will tell you. Most retailers over here sell their laptops simply advertising "Nvidia graphics with 4GB RAM" (sic!). No kidding.

 

[windwhirl]

It matters little what manufacturers will tell you. Most retailers over here sell their laptops simply advertising "Nvidia graphics with 4GB RAM" (sic!). No kidding.

At best you get the GPU name (3060, 3070, etc.)

 

[Sybaris_Caesar]

I saw a guy (or guyette?) criticising Nvidia's max-Q and max-P naming scheme on reddit recently. The wattage budget makes GPUs of the same SKU perform differently in different configs.

But now we've regressed back to the old system. That's why I stated that I consider laptops garbage in a forum discussion few months back. You can't just say oh this lappy has i7 so it must be better than that i5 laptop. You don't know if the laptop has M.2 SATA ssd or NVMe one. RAM sticks are soldered or dual channel or not.

Laptop makers hold too much power over the configuration without disclosing it to the end-user.

 

[windwhirl]

I saw a guy (or guyette?) criticising Nvidia's max-Q and max-P naming scheme on reddit recently. The wattage budget makes GPUs of the same SKU perform differently in different configs.

But now we've regressed back to the old system. That's why I stated that I consider laptops garbage in a forum discussion few months back. You can't just say oh this lappy has i7 so it must be better than that i5 laptop. You don't know if the laptop has M.2 SATA ssd or NVMe one. RAM sticks are soldered or dual channel or not.

Laptop makers hold too much power over the configuration without disclosing it to the end-user.

It's partly due to that obsession with making everything as thin as paper. There's nothing wrong with making things smaller, lighter, etc., but it's getting to be a bit much. Then again, how many people actually bother to do upgrades even if their laptops allow them to do so?

 

[bug]

I saw a guy (or guyette?) criticising Nvidia's max-Q and max-P naming scheme on reddit recently. The wattage budget makes GPUs of the same SKU perform differently in different configs.

But now we've regressed back to the old system. That's why I stated that I consider laptops garbage in a forum discussion few months back. You can't just say oh this lappy has i7 so it must be better than that i5 laptop. You don't know if the laptop has M.2 SATA ssd or NVMe one. RAM sticks are soldered or dual channel or not.

Laptop makers hold too much power over the configuration without disclosing it to the end-user.

Not to mention the (very) limited BIOS that comes with laptops. Often, you can't even disable the IGP, for example.