The recently announced NVIDIA GeForce RTX 2050 and MX500 graphics processors were tested in 3DMark Time Spy and the benchmarks don’t look great.
NVIDIA GeForce RTX 2050 and MX500 GPUs are entry-level chips for laptops, and their performance so far looks amazing
We now understand that these are entry-level GPUs designed to replace or offer something better than the integrated GPUs that AMD and Intel offer in their laptops, but that’s the thing, not only are they confusing specs, but the parts can offer poor performance. If these standards need to be trusted.
In our post yesterday, we talked about how the NVIDIA GeForce RTX 2050 and MX570 GPUs have a GPU-amp (GA107) with 2048 cores and a 64-bit bus interface, although there is a fundamental difference between the RT and Tensor cores that NVIDIA has yet to make explained. The GeForce RTX 3050 has 4 GB, while the MX570 has 2 GB of GDDR6 memory. As for the MX550, it is based on the much older Turing TU117 architecture and only has 1024 cores, 2 GB and a 64-bit bus interface.
One interesting thing zihu (Via HXL) states that the TGP of new GPUs in laptops not only determines how much the speaker clock will be shifted, but also determines the memory clock. At 20W you get 11Gbps, at 25-35W you get 12Gbps and at 40-45W you get 14Gbps memory speed. The RTX 2050 runs at 30-45W, giving us 14Gbps, so it’s actually faster than the MX570, which has a 15-25W design versus 12Gbps. The same goes for the MX550.
NVIDIA GeForce RTX 2050 3DMark Time Spy Benchmark (Thanks: Zhihu):
NVIDIA GeForce MX550 3DMark Time Spy Benchmark (Thanks: Zhihu):
Now coming to the benchmarks, the RTX 2050 and GeForce MX570 are said to offer roughly similar graphics performance and score roughly 3,369 points. GeForce MX550 scored 2510 points. Now, compared to the GeForce RTX 3050, we’re looking at 40-45% lower performance on the same architecture, but with the RTX 3050 with higher clock speeds and a wider (128-bit) bus. This can be considered the performance of the GTX 1660 series in 2021, which can make good profits on the iGPU. Now that there’s a packet-tracking kernel in both parts, don’t expect to get decent performance out of it unless you’re playing 720p with DLSS set to max power mode.
On the other hand, the NVIDIA GeForce MX550 looks very unsatisfactory, especially considering the leaked AMD Ryzen 6000 ‘Rembrandt’ APU score of 2700 in 3DMark Time Spy. It looks like the MX550 will either be on par or even missing compared to AMD’s RDNA 2 iGPU, which makes this part pretty much a DOA, even if we give it time to release to see if the final chip has any pre-release silicon improvements and whether the drivers Final does add any improvements, but the same can be true for the AMD Rembrandt APU in its final form.
“Analyst. Total tv trailblazer. Bacon fanatic. Internet fanatic. Lifelong beer expert. Web aficionado. Twitter buff.”