Until a few years ago, before the Pascal era, NVIDIA GeForce GTX mobile graphics cards differed significantly in performance compared to their desktop counterparts. In the middle of 2016 a revolution took place – all NVIDIA GeForce GTX mobile graphics cards 10 x0 offered the same specification of GP chips 106 and GP 104, the only difference was slightly lower core clock. Exactly two years ago, I attended the NVIDIA conference in Las Vegas, during which the Turing architecture was presented for laptops. The situation repeated – we got the same HERE cores 117, HERE 116, HERE 106 and HERE 104. Additionally, some cards supported Ray Tracing and DLSS for the first time. We did, however, feel more maneuvering the memory and core timing. Today, NVIDIA introduces the next generation Ampere for notebooks – GeForce RTX cards 3000. While it will undoubtedly be efficient, it cannot be denied that, in a way, we are returning to the times before Pascal architecture.

NVIDIA officially unveiled the new generation of GeForce RTX graphics cards 3080, RTX 3070 and RTX 3060 for notebook versions. The Ampere architecture is coming to laptops and will bring a big leap in performance. After all, for the first time in years, mobile cards will cease to be their desktop equivalents.

The NVIDIA GeForce RTX series 3000 is by far the most efficient line of graphics cards available ever appeared in laptops. Nevertheless, the Ampere mobile chipsets will not be 1: 1 equivalents to what we got in desktops a few months ago. The lack of specific nomenclature distinctions will only aggravate this inaccuracy. The manufacturer today presented three new graphics chips – NVIDIA GeForce RTX 3080, NVIDIA GeForce RTX 3070 and NVIDIA GeForce RTX 3060. The first two cards are based on the Ampere GA core 102, of which RTX 3080 uses a full GA core 104 with 6144 CUDA cores, while RTX 3070 has GA truncated 104 with 5120 CUDA cores. So while the difference in the RTX specification 3070 is not so noticeable, the RTX 3080 very big changes have already been made by simply using a different core. The mobile GeForce RTX 3080 is therefore something completely different than in the desktop, despite the same naming.

NVIDIA GeForce RTX 3060 is the third Ampere chip, this time based on the Ampere GA core 106 with 3840 CUDA cores and 6 GB of VRAM GDDR6 memory. After all, it is with this card that we expect the greatest performance jump compared to its predecessor (RTX 2060 -> RTX 3060). In this case, gaming performance should be above the mobile GeForce RTX 2080 SUPER. In the case of GeForce RTX 3070 Mobile, we count on performance by 50% higher than GeForce RTX 2070. The most powerful – RTX 3080 – in the mobile version should achieve performance comparable to the desktop NVIDIA GeForce RTX 2080 Ti. Unfortunately, we cannot count on a level comparable to the desktop RTX 3080. We emphasize it again – just like the desktop RTX 3080 uses the GA core 102 and 8704 CUDA Cores, Yes Mobile RTX 3080 is full GA 104 and 6144 CUDA cores. Interestingly, the mobile GeForce RTX 3080 will come with 16 GB of VRAM, while the desktop version has “only” 10 GB.

NVIDIA also presented the third generation Max-Q technology at the conference, additionally driven by artificial intelligence. The foundations of the third generation of Max-Q are also the following techniques: Dynamic Boost 2.0, Whisper Mode 2.0, Resizable BAR and DLSS 2.0. The combination of Dynamic Boost 2.0, Resizable BAR and DLSS should noticeably improve performance in supported games, and translate into longer battery life. The new Ampere cards for laptops are also expected to bring noticeable performance gains in software such as Blender (Denoiser), Adobe Premiere Pro or DaVinci Resolve Studio 17. Laptops with NVIDIA GeForce RTX cards 3000 also belong to the NVIDIA Studio project, where the selected devices are great they are suitable not only for games, but also for professional work, e.g. on creating materials (Content Creators). Laptops with GeForce RTX cards 3000 will be available from 26 January with prices starting from 999 dollars. In the first months, we expect over 70 configuration from all leading manufacturers.

Source: NVIDIA