Nvidia Introduces GeForce RTX 2050, MX570 and M550 Laptop GPUs – Update

Spread the love

Nvidia has introduced three new video cards for laptops. The company is coming with a GeForce RTX 2050, in addition to a GeForce MX570 and MX550. Laptops with the new GPUs will be delivered in the spring of 2022.

The Nvidia GeForce RTX 2050 for laptops has 4GB of GDDR6 memory with a bandwidth of 112GB/s as standard. The video card also has 2048 CUDA cores. On paper, that’s more CUDA cores than the RTX 2060 for laptops, which has 1920 CUDA cores. However, Nvidia confirms to Anandtech that the RTX 2050 has a GA107 GPU based on the Ampere architecture. Other RTX 20 video cards are based on the older Turing architecture. Nvidia has changed the internal structure of the CUDA cores to Ampere and therefore these core numbers cannot be compared one-to-one.

Furthermore, the GeForce RTX 2050 laptop GPU has as many cores as the GeForce RTX 3050 for laptops. The RTX 3050 does have more memory bandwidth on top of that. The boost clock of the RTX 2050 is also lower. The GPU therefore has a lower TDP of 30W to 45W. The first laptops with a GeForce RTX 2050 will appear in the spring of 2022.

The GeForce MX570 also gets a GA107 GPU based on the Ampere architecture, while the MX550 gets a TU117 chip based on Turing. However, Nvidia shares few details about these two video cards. The manufacturer does not mention core numbers and no details about the memory arrangement. The company only reports that the MX570 has a ‘higher memory speed’ than previous GPUs in the MX series. These two video cards will also appear in the spring.

Update, 17.10: Initially, the article stated that both the MX570 and the MX550 will get an Ampere GPU. However, the MX550 gets a Turing chip. The article has been adapted accordingly. Thanks to JackAvery.

Nvidia GeForce RTX GPUs for Laptops
Fashion model GeForce RTX 3050 Ti GeForce RTX 3050 GeForce RTX 2050 GeForce RTX 2060
Architecture Ampere Ampere Ampere Turing
GPU GA106 GA107 GA107 TU106
cores CUDA: 2560
Tensor: 80
RT: 30
CUDA: 2048
Tensor: 64
RT: 16
CUDA: 2048
Tensor: 64
RT: 16
CUDA: 1920
Tensor: 240
RT: 30
(different core structure)
Clock speed 1035-1695MHz 1057-1740MHz 1155-1477MHz 1185-1560MHz
Memory 4GB GDDR6 4GB GDDR6 4GB GDDR6 6GB GDDR6
Memory bandwidth 224GB/s 224GB/s 112GB/s 336GB/s
Total graphics power 60-115W 35-80W 30-45W 65-115W
Release date Nov 5, 2021 Nov 5, 2021 Spring 2022 January 2019
You might also like
Exit mobile version