The entry-level Ampere GPU might get a new life with a newer GPU die, but don't expect any big impacts on the performance due to a cut-down GPU.
GPU-Z Version 2.67.0 Software Receives Support for RTX 3050 "A", Boasting the AD106-A Die Instead of GA106/107, Hinting at a Refreshed RTX 3050 GPU for Both Laptop and Desktop
It's not the first time we've seen NVIDIA deploying a different GPU chip in an existing card, but it's rare to see GPUs receiving newer GPU dies after years. The GeForce RTX 3050, which was released in 2021 for laptops and in 2022 for desktops, is reportedly getting a new iteration but with the newer "Ada Lovelace" architecture. The original GPU brings the Ampere-based GA106 die for the desktop and the GA107 die for the laptop variant, but the newer one will feature the "AD106".
TechPowerUp released its latest GPU-Z software, version 2.67.0, which mentions the new GeForce RTX 3050 in the release notes. It's not just the RTX 3050, but it's an RTX 3050 'A', which will indicate the difference between the Ampere and Ada-based variants. NVIDIA did confirm the new RTX 3050 A last year, but it was only for the mobile platform. The specifications include a newer AD106 die but a severely cut-down version (to match the performance of Ampere-based RTX 3050), which brings only 1792 CUDA cores compared to 2048 on the Ampere-based RTX 3050 laptop GPU.
There are downgrades in various areas, but the memory configuration has been retained. With TechPowerUp's latest GPU-Z version, the RTX 3050 A is mentioned without "mobile", which indicates it could be a desktop graphics card. We haven't seen any previous GPU-Z versions mentioning the support for either the mobile or desktop RTX 3050 A variant, and the new release notes could be mentioning just the mobile variant. However, NVIDIA would likely want to use its leftover stack of Ada silicon for this entry-level GPU.
The new RTX 3050 A desktop variant will also use the AD106 if NVIDIA is really going for a new iteration and should bring cut-down cores similar to the laptop variant. We should see 8 GB GDDR6 memory with a 128-bit memory bus and a TDP ranging from 100- 150W. NVIDIA will be trying to match the performance of the Ampere-based RTX 3050 and will cut down the specs accordingly. It is also possible that we may see slightly higher performance, similar to the uplifts we saw in synthetic benchmarks of the RTX 3050 A mobile. Nonetheless, the good thing is that users will be able to identify between the Ampere-based and Ada-based RTX 3050 cards based on the naming convention.
News Source: Techpowerup
Follow Wccftech on Google to get more of our news coverage in your feeds.
