NVIDIA GeForce RTX 2050 and MX550 Laptop GPUs Tested in 3DMark TimeSpy

The test results were obtained from the Zhihu platform, where preliminary estimates were presented based on production preparation samples. Of the three accelerators announced yesterday, the leaked data covers only two of them – the RTX 2050 and MX550. The RTX 2050 and MX570 are supposed to perform similarly, at least in this 3DMark benchmark.

Note that ComputerBase has confirmed that both GPUs are essentially the same (based on an Ampere GA107 GPU with 2048 CUDA cores), which means they support hardware ray tracing and DLSS (via Tensor cores). However, RT kernels are somewhat limited, which hasn’t been explained in detail yet. In addition, the RTX 2050 has 4Gb, while the MX570 is equipped with 2Gb. In any case, this should not affect the result of the TimeSpy test.

The specs confirm that both MX500 series numbers are limited to the 64-bit memory bus and 12Gbps bandwidth. The clock speeds are 1155 MHz for the MX570 and 1320 MHz for the MX550. The MX550 is a stripped-down GeForce GTX 1650 graphics card based on the Turing TU117 GPU. It retained 1,024 CUDA cores, but the memory bus was split in half to 64-bit. At the same time, the memory technology has been upgraded from the GDDR5 standard to the GDDR6 standard.

The benchmark results are not very impressive. The RTX 3050 laptop GPU with 128-bit memory bus scores about 4800 points in the time-lapse graph. But again, these MX500 results are based on preliminary samples.

They will most likely be paired with next-generation mobile processors (such as Alder Lake-M), which may not deliver the ultimate performance in this state. After all, the RTX2050 and MX500 GPUs are officially coming out in the spring of 2022.


Leave a Comment

Your email address will not be published.

Shopping Cart