Back in December 2021, NVIDIA unexpectedly announced three new mobile GPUs aimed at the low-end/entry-level market. One of these three new GPUs was the RTX 2050. Despite being a RTX 20-series card, it was actually based on a GA-107 Ampere GPU with 2048 CUDA Cores carrying 6GB of GDDR6 VRAM. These GPUs were meant more so as a replacement for integrated graphics rather than actual discrete GPUs offering better performance.
Interestingly, right after these were announced, the benchmarks for the RTX 2050 leaked hours later and showed extremely underwhelming results. It was similar to MX 570 in terms of performance because both are based on the same GA-107 die. Back then, in my coverage I mentioned that we should wait for benchmarks close to the GPU’s Spring 2022 launch before judging them. That’s exactly what we’re looking at today.
MechRevo, a Chinese laptop manufacturer has just shared benchmarks of the RTX 2050 laptop GPU, comparing it against other GPUs in this segment. The benchmark in question is 3DMark Fire Strike and, as you can see below, the RTX 2050 is roughly 23% faster than its predecessor, the GTX 1650. Moreover, RTX 2050 also pulls ahead of the MX 470 graphics by about 134%.
Apart from this, MechRevo were also kind enough to provide video encoding efficiency results for a bunch of GPUs competing against the RTX 2050. The GPU topped the list, being second to only Intel’s top-spec integrated graphics. Intel Iris Xe with 96 Execution Units was able to beat every single other GPU on the list despite being an iGPU.
This puts the RTX 2050 in an interesting position where it’s significantly faster than the generation it succeeds but doesn’t exactly perform like a chart-topper against similarly-spec’d graphics configurations. NVIDIA hasn’t disclosed the starting price for laptops equipped with the RTX 2050 so it’s hard to say whether the GPU will be worth it or not at the moment. But, with the way things stand right now, the RTX 2050 discrete graphics chip is about as performant as a high-end iGPU.