Purported RTX 2050 Mobile Ampere GPU Benchmarks Revealed
Chinese laptop maker is first out of the door with this mobile GPU, and kindly shares some charts.
Late last year, we reported that Nvidia was preparing a new crop of discrete mobile GPUs, which we were told would debut in Spring 2022. Today, ITHome reported that Chinese laptop brand Mechanical Revolution (MechRevo) launched the Unbounded 16 Pro, which features one of the most intriguing members of Nvidia's new lineup of mobile GPUs, the GeForce RTX 2050. Some GPU performance charts were also kindly shared.
Please take these charts with a heavy dose of salt, as though they look official, and MechRevo has published some previous news pages concerning the Unbounded 16 Pro, we couldn't find the performance slides on the official site.
Above you can see purported scores for the RTX 2050 against incumbent low-end/thin and light mobile GPU stalwarts like the GTX 1650 and MX 450. The performance vs the MX 450 is more than doubled, but the GTX 1650 puts up a better fight with the new contender only gaining a 23% advantage.
GPU performance in benchmarks like 3DMark Fire Strike isn't always a great indicator of gaming performance. This, of course, varies considerably depending on the game.
Seasoned readers will be well aware of some increasingly attractive RTX benefits the newer more powerful RTX 2050 mobile GPU can offer, thanks to RT cores and Tensor cores being present. Nvidia's real-time raytracing possibilities, DLSS, Reflex and Broadcast all play nicely with the RTX 2050, which is Nvidia Optimus technology-friendly.
Nvidia GeForce |
RTX 2060 |
RTX 2050 |
MX 570 |
---|---|---|---|
CUDA Cores |
1,920 |
2,048 |
2048 |
GPU Clocks |
1,185 - 1,560 MHz |
1,155 - 1,477 MHz |
1,087 - 1,155 MHz |
Graphics Subsystem Power (W) |
65 - 115 W |
30 - 45 W |
15 – 25 W |
Memory Speed |
Up to 14 Gbps |
Up to 14 Gbps |
Up to 12 Gbps |
Standard Memory Configuration |
6 GB GDDR6 |
4 GB GDDR6 |
2 GB GDDR6 |
Memory Interface Width |
192-bit |
64-bit |
64-bit |
Memory Bandwidth (GB/sec) |
Up to 336 GB/s |
Up to 112 GB/s | Up to 96 GB/s |
Please note that despite its 20 series moniker, the RTX 2050 uses an Ampere GPU. In fact, it uses the same GA107 as the new MX 570, which is the same silicon from the RTX 3050 and RTX 3050 Ti Mobile. In addition, our December report highlighted that the new RTX 2050's 2048 CUDA cores are higher than the RTX 2060 (Turing) but will be slower with its restricted mobile TDP.
Moving along, we see a slide shared by ITHome from a video editing and processing package called 'Clipping Professional Edition.' This chart shows that the RTX 2050 can help your laptop encode considerably faster than a device relying solely on Intel's Iris Xe 96 EU iGPU. The benefits become more apparent as the video resolution scales up from 1080p, through 1440p, to 2160p (shorter bars mean faster processing).
Stay On the Cutting Edge: Get the Tom's Hardware Newsletter
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
The source report indicates that the MechRevo Unbounded 16 Pro will be available to purchase in early March.
Mark Tyson is a news editor at Tom's Hardware. He enjoys covering the full breadth of PC tech; from business and semiconductor design to products approaching the edge of reason.
-
watzupken Looks like Nvidia has taken a page from AMD’s playbook by introducing 64 bit memory bus for budget GPUs (mobile GPU) in this case. Ray Tracing on 4GB VRAM?Reply -
hotaru.hino
NVIDIA's been doing this for a while on their budget GPUs.watzupken said:Looks like Nvidia has taken a page from AMD’s playbook by introducing 64 bit memory bus for budget GPUs (mobile GPU) in this case. Ray Tracing on 4GB VRAM?
As far as ray tracing goes, well NVIDIA recommends (or "requires") 6GB minimum, so they could disable it for the 2050. But DLSS is still a thing and an RTX feature, so the nomenclature isn't entirely wrong.