CES 2023 – GeForce RTX 40 drops on laptops

alienware m18

Discover when there are no offers

The above links may not work with an ad blocker.

HP Omen 17 (2023)

HP Omen 17 (2023)

Discover when there are no offers

The above links may not work with an ad blocker.

On the eve of the opening of CES, Nvidia held its traditional conference. The firm has made official the GeForce RTX 4070 Ti for desktops, as well as the entire RTX 40 range that will be found in laptops.

GeForce RTX 40 for notebooks is based on the same architecture as its counterparts in the desktop version: Ada Lovelace. Among other things, it introduces new carving finesse, shaders rearrangement (SER) to optimize their processing, and the third iteration of DLSS, which allows images to be embedded and therefore framerates to be increased.

Five shades of GeForce RTX 40

For laptops, Nvidia pulled five changes out of its hat, from the humble RTX 4050 to the monstrous RTX 4090. However, note the similarities between the names: the GeForce RTX 4090 for notebooks actually captures the RTX’s GPU. 4080 for a desktop computer with 9728 computing units (CUDA cores). 16 GB of RAM is available on a 256-bit bus, but not in GDDR6x, as in the desktop version, but in GDDR6. The thermal envelope is processed downwards with a maximum of 150 watts and 25 watts of Dynamic Boost. And mechanically, the frequencies are lower, with a maximum of 2.04 GHz for the portable version and 2.51 GHz for the desktop version.

spec RTX 40 geforce nvidia

It will be necessary to be vigilant against the power envelope applied by the partners.

© Nvidia

Likewise, the RTX 4080 for a laptop is very close to the newly announced RTX 4070 Ti. However, the mobile version has lost 260 CUDA units, along with the slimming treatment needed to integrate the laptop into the PC chassis as we saw above.

Advertisement Your content continues below

As such, the GeForce RTX 4070, 4060 and 4050 see a drop in specs as they go down the range. Thus, the GeForce RTX 4050 ends up with 2560 CUDA units on a 96-bit bus and 6 GB of RAM. For comparison, the GeForce RTX 3060 mobile had 6 GB of RAM on a 192-bit bus, and the GeForce RTX 3050 had 4 GB of RAM on a 128-bit bus.

Large power envelope gap

Nvidia taught us to change the power envelope from simple to double with the GeForce RTX 30. Now we go from single to triple in versions 50 to 70, with a minimum heat envelope of 35 watts and a maximum envelope of 115 watts. So, depending on the integration of notebook manufacturers, GeForce RTX 4070 (60 or 50) can differ greatly in its performance.

A sharp increase in performance

Nvidia also highlighted the performance of the new GeForce RTX 40s, but you have to read between the lines to properly appreciate the order of magnitude. Nvidia likes to announce gains of x3 or even x4, or 300 and 400%, which is simply huge. To announce such numbers, Nvidia compares the RTX 3080 Ti to the DLSS 3.0-enabled RTX 4090, which is incompatible with the RTX 30.

geforce rtx 40 performance

DLSS 3.0 is widely touted to account for the performance increases.

© Digital

Therefore, it is necessary to refer to games that are not compatible with DLSS 3.0 Assassin’s Creed: Valhalla Where Red Dead Redemption 2, the results of which were briefly discussed by Nvidia. So in rasterized games the generation gain will be around 50%, which is still significant. Another tip is software like Blender or Arnold, which have a gain of about 100%.

Advertisement Your content continues below

Improved energy efficiency

Nvidia isn’t just talking about performance gains: the brand is also highlighting the energy efficiency of its new architecture. So, according to the tests, the GeForce RTX 4090 – GeForce RTX 3050 – GeForce RTX 3070 provides performance with a 40-watt envelope. The trick to getting such a result probably comes from enabling DLSS 3.0. , while GeForce RTX 30s do not have this option.

island lovelace nvidia geforce

According to Nvidia, the GeForce RTX 4090 delivers the performance of the GeForce RTX 3070 at only 40 watts.

© Nvidia

In addition to the abundance of these figures, Nvidia has improved Max-Q technology, which includes all options designed to improve the autonomy of laptops equipped with GPUs. In this 5th generation Max-Q, Nvidia integrates new GDDR6 running at lower voltages with added new operating states. Obviously, the architecture itself introduces efficiency improvements. Nvidia now includes DLSS 3.0 as part of its Max-Q technologies, because image generation with artificial intelligence requires less energy than traditional image generation.

nivida partners geforce rtx 40 laptop

All of Nvidia’s partners have already announced their GeForce RTX 40-equipped models.

© Nvidia

Notebooks with GeForce RTX 40 are currently being announced by Nvidia partners. The first models with GeForce RTX 4080 and 4090 will go on sale on February 8 for €2,400. Models incorporating the GeForce RTX 4050, 4060 or 4070 will be available from February 22 with a starting price of €1,200.

Advertisement Your content continues below

Leave a Reply

Your email address will not be published. Required fields are marked *