Back in December of 2021, NVIDIA announced three new mobile GPUs based on its Ampere architecture, the RTX 2050, the MX550, and the MX570. All of these GPUs were given a tentative release date of Spring 2022 with little information officially disclosed by NVIDIA. So, thus far we’ve only had leaks and rumors to rely on to evaluate the performance of these SKUs.
Both the RTX 2050 and MX570 have gotten a major benchmark leak before in which they performed surprisingly similar to each other. This was to be expected however as they are based on the same Ampere GA-107 GPU with the only minute differences. Moreover, despite being from two different series of GPUs, both of these options were aimed to entry-level laptops in hopes of providing a replacement for integrated graphics.
Today, history is repeating itself as we have yet another benchmark, this time for the MX570, which shows the GPU’s performance being almost identical to the RTX 2050. Covered first by Tom’s Hardware, ITHome was the one who initially spotted the Geekbench listing through Benchleaks, and brought it to everyone’s attention. The Geekbench listing detected the laptop as a “HP Zhan 66 14-inch G5” with an Intel Core i5-1235U processor and 16GB of DDR4 memory.
Someone managed to get a hold of this unreleased laptop with an NVIDIA MX570 and ran an OpenCL benchmark, giving us our first new benchmark scores for the MX570 since February. Now I should preface this by saying that by no means is OpenCL is a good aggregator of real-world performance, especially when it comes to gaming, but it’s all we have for now so we’re gonna have to work with it.
As you can see in the screenshot below, the MX570 nets a score of 44,638 points in the OpenCL test. For comparison, the RTX 2050 scores an average of 45,986 points across the entire Geekbench database. However, the highest score for the RTX 2050 does go as high as 48,617 points on a Xiaomi Redmi Book Pro 15 2022 laptop listing. But there’s one very important thing to keep in mind here.
Even though the RTX 2050 and MX570 use the same GPU underneath and have the same amount of CUDA Cores (2048), there are two key differences between the two. First, the RTX 2050 has 4GB of VRAM versus only 2GB on the MX570. More importantly, though, there is a sizable TGP gap between the two. The RTX 2050 has a power limit of 30-45W, whereas the MX570 is limited to only 15-25W.
That power difference alone can have a significant impact on performance as evident by the highest RTX 2050 score being around 4,000 points ahead of the MX570. Though, the fact that the MX570 still remains comparable despite that power limit discrepancy is impressive (and kind of strange) and presents a very good option for thin and light notebooks that prioritize efficiency over performance. This lower TGP could help preserve battery life, too.
Right around the corner
The last time we saw the MX570 pitted against the RTX 2050 was back in February, as mentioned before, where Chinese laptop manufacturer MechRevo officially released the 3DMark Fire Strike scores for the RTX 2050. Before that, we saw our very first benchmark leak for the RTX 2050 and MX570 back in December of last year, literally hours after the GPUs were announced out of the blue.
While those initial benchmarks suggested that the RTX 2050 and MX570 are so similar that they’re essentially interchangeable, I wrote at that time that we should wait for benchmarks closer to the GPUs’ release to evaluate their performance. Well, Spring is here and with this benchmark leak upon us, these GPUs must not be far from release. We could be getting official word from NVIDIA anytime at this point. Till then, let’s hope the leakers keep us updated.