It’s a “tock” 12 months: New Nvidia laptop computer GPUs in 25 OEMs’ programs this month


  • Here is a gallery of laptops from April 2020 with the latest Nvidia Max-Q GPUs. First the Razer Blade Advanced with an RTX 2080 Super inside.

  • The Acer Nitro 5, which starts with an RTX 2060 GPU at $ 999.

  • The Predator Triton 500.


  • The ASUS Tuf.

  • The latest HP Omen 15.

  • The Lenovo Legion 7.

  • The Lenovo Legion Y540.


  • The MSI GS66 Stealth.

  • The Clevo PC50.

  • The Gigabyte AERO 15 Studio, OLED edition.

What kind of GPU year can we expect from Nvidia, one of the two largest GPU manufacturers for consumers in the world? The answer is somewhat in the air as Nvidia is in a firm, yet fluid, position. Market worries and announcements-canceled events hover on the one hand, while the company's surprisingly optimistic financial forecast is striking on the other.

In both cases, we reached April by the end of March without the company's usual announcement of new desktop hardware, and we still don't know when entirely new desktop GPUs may be released (more on that later). . Instead, we're starting this month with a different wave of products: a new list of GPUs for laptops, if not as new.

Nvidia has announced a wave of "Max-Q" GPUs that will come from 25 OEMs on laptops by the end of April. Most, but not all, of the company's RTX GPU range. This month's GPU wave consists of three new laptop SKUs (RTX 2080 Super, RTX 2070 Super, GTX 1650 Ti) and minor updates to four existing SKUs (RTX 2070, RTX 2060, GTX 1660 Ti, GTX 1650). Each of these GPUs is based on the manufacturer's Turing 12nm architecture.

  • Nvidia provided this diagram over the April 2020 line of Max-Q GPUs. It does not compare with last year's models.

  • When pressed, Nvidia provided this clearer comparison between some RTX models from 2019 and their "super" upgrades in 2020.

But as we learned from last year's "Super" line of desktop GPUs, these updates mainly consist of mild but welcome jumps on almost identical silicon, which we see in two new Max-Q Super options. Both new Super Max-Q cards benefit the most from an obvious increase in the maximum boost clocks. However, it remains to be seen how these numbers will affect the wild. Last year's cards were launched with different maximum wattages, and this may be the case this time. And they have very slight impairments in tensor core capacity, mainly targeting DirectX-based ray tracing and proprietary Nvidia RTX effects (like DLSS, which we'll discuss later). Otherwise, both still have 8 GB of GDDR6 memory with a maximum memory bandwidth of 448 GB / s and each has received a slight increase in the "CUDA Core" processing units (plus 4.6 percent over the 2080 Super and 11.1 percent over the 2070) Awesome).

A mere press conference prior to the unveiling of the cards showed that Nvidia has found ways to make more profit from these cards, which is why they have horned some new features that are exclusive to this year's models. For the largest program, called "Dynamic Boost", Nvidia works with OEMs to manage the shared thermal workload of CPU and GPU and up to 15W power from a motherboard element based on an application's live frame time data to transfer to another. That means if these sensors activated at the driver level recognize that either the GPU or the CPU has been pushed to the maximum and the other half could save some power, the system redirects its overall performance accordingly.

"All this is free of charge"

Representatives from Nvidia estimate that this system-level optimization will result in a performance increase of around 4 to 8 percent in every game. "But it's all free," said Mark Aevermann, marketing director of Nvidia, in a phone interview with Ars Technica. "[The performance] was locked up there because a controller in the past was not smart enough to do this." It is noteworthy that a full buy-in from an OEM is required for this function before a laptop is shipped with a new Nvidia GPU, as Nvidia has to confirm optimizations at the system level. So don't expect it to be patched into existing laptops with older GPUs. In addition, AMD CPUs do not work with Nvidia Dynamic Boost at the time of printing, and this support will arrive "shortly after", Aevermann said.

Aevermann mentioned additional improvements to achieve even smaller increases in performance, including voltage optimizations for both the GDDR6 memory of the GPUs and for the general voltage regulator of the GPUs. These result in a power recovery of 1 to 2 W, Aevermann said, which was admittedly mild, but insisted that this was crucial: "We want every saving of half a watt, a quarter watt and a watt to be the GPU core drives. "

The last item on the optimization list is a function that sounds great: "Advanced Optimus". This is Nvidia's term for a new connection protocol that is intended to ensure that the integrated graphics solution of a laptop does not impair the direct connection of an Nvidia Max-Q GPU to the display of a laptop. Aevermann said this is simply an optimized way for OEMs to guarantee screen technologies such as variable refresh rates (G-Sync, FreeSync) and frame rates up to 144 fps and beyond, although we've both seen the last two on Nvidia laptops Years. However, this new toolset may also incur licensing costs or other issues, as only one upcoming laptop, the Lenovo Legion 7, has been approved for implementation of Advanced Optimus at the time of printing. "More will follow later." ("Advanced Optimus, like Dynamic Boost – we offer it to OEMs, but it's up to them to implement it or not," Aevermann told Ars.)

  • This whirring blade looked like chaos in the original implementation of DLSS. This year's update takes into account more common visual systems in games, including semi-transparent materials and particle effects.

  • It also does a ridiculously good job of recognizing and "improving" text elements. While these slides were provided by Nvidia, tests in the current version of the game confirm how Ars tested them.

Nvidia seems to hope that consumers are just as excited about these small improvements as they are about the RTX-exclusive Deep Learning Super Sampling (DLSS) image processing system that received a major update at the end of last month. DLSS 2.0 is indeed an impressive offering because its machine-trained anti-aliasing system can upscale 1080p footage to a whopping 4K pixel count with very little perceptible noise or distortion. As long as developers implement a temporal anti-aliasing system (TAA) in their game and forward the results to Nvidia, an automatic model can be generated. If more games were to do so, it would surely be good news for anyone who values ​​their more expensive RTX GPUs, laptops, or otherwise.

Our own anecdotal tests of last week's DLSS 2.0 patch for the beautiful raytraced game control showed that the system works mostly as advertised and the above pictures Dixie don't whistle: They are confirmed by us same game. However, the effect is much better if a 1080p signal with a resolution of 4K is processed. The DLSS 2.0 up-conversion process continues to have problems with lower resolutions, especially its attempts to scale the 960p resolution to 1440p. Therefore, laptop buyers who are aiming for an RTX system with a 1440p monitor should pay attention to this. Aevermann said that his DLSS system will continue to be gradually updated over time. We therefore hope that resolutions with lower sources will be emphasized more in the future, especially for laptop systems that do not always come with 4K monitors.

As we prepared this report, rumors of Nvidia's next series of Ampere generation GPUs picked up speed. A report indicated that we would have to wait until September 2020 to see "RTX 3000" cards. In the meantime, if you want to get involved with interesting new Nvidia hardware, you have to choose this month's Max-Q line. We hope to be able to work with newly equipped laptops by the end of this month and will report on our results.

Listing picture by Razer


Please enter your comment!
Please enter your name here