HomeReviewsNVIDIA Lovelace AD102 GPU: It will probably reach the GeForce RTX 40 series

NVIDIA Lovelace AD102 GPU: It will probably reach the GeForce RTX 40 series

NVIDIA's next-generation Ada Lovelace GPU architecture has been in the headlines for the past two weeks, as many rumors have been circulating lately.

NVIDIA Ada Lovelace

See also: Nvidia: Introduces ARM-Powered Linux RTX demo

3DCenter.org has compiled a list of NVIDIA AD102 GPU specifications based on rumors from @kopite7kimi and @ greymon55.

In terms of specs, NVIDIA's next-generation AD102 is expected to be built on TSMC's new 5nm hub, with 18.432 CUDA cores (the GA102 on the RTX 3090 has 10.496 CUDA cores to compare). GPU clocks are now expected to be 2,2 GHz or higher, which will result in approximately 81 computational performance TFLOPs (35,5 TFLOPs on the GA102 or RTX 3090).

See also: NVIDIA's new GeForce 471.41 WHQL drivers have been released

The most important mention is that NVIDIA will use the new Micron GDDR6X memory clocked at 24Gbps, from 19,5Gbps to the RTX 3090. The new AD102 GPU will have a 384-bit memory bus, so we'll have to wait 1152 GB / sec. memory bandwidth - 23% increase over 19,5Gbps in GDDR6X memory used on the RTX 3090 (and 936 GB / sec memory bandwidth).

NVIDIA has completed the Ada Lovelace GPU project, which is why we hear so much about it right now - with the next phase in Ada Lovelace being the final design and the so-called "tape-out". We will not receive it in our hands until the end of 2022 - with estimates for Q4 2022, while in terms of efficiency we can expect a 100% + improvement over the GeForce RTX 3080 Ti and GeForce RTX 3090.

This will ensure 4K 120FPS + gaming in next generation games with ray detection and I am sure we will be introduced to the new generation black magik DLSS 3.0 technology with Ada Lovelace, which will only improve things.

A year from now is a long time, so we should expect renewed GPU Ampere in the form of GeForce RTX 30 SUPER graphics cards between now and Q4 2022.

See also: Nvidia can work on the RTX 3080 Super

As for Micron's much faster 6Gbps GDDR24X memory, this is something that was mentioned a year ago, with Micron announcing that it would have 16Gb GDDR6X memory in 2021 that would reach 24Gbps. Micron explained in August 2020: “Its roadmap Micron also points out the possibilities for 16Gb GDDR6X in 2021 with the ability to reach up to 24Gb / s. GDDR6X is powered by a revolutionary new PAM4 configuration technology for Ultra-Bandwidth solutions. PAM4 has the potential to make even more improvements to data rate. "

Source of information: tweaktown.com

Teo Ehchttps://www.secnews.gr
Be the limited edition.