So I do agree with most people here that this most likely seems to be a new AMD GPU that is quite a bit faster than a 2080 Ti, but I also think to not be disappointed if it ends up being something different, no one should feel 100% certain about this. And since I would consider it very unlikely that someone has access to both an AMD engineering sample CPU and an unannounced Nvidia GPU, and is allowed to use both in the same system, I think it is most likely that this is an unannounced AMD GPU. So this benchmark result is very likely done with some unannounced GPU. I guess it might be just about possible to get a '30% higher than 2080 Ti' result by having a 2080 Ti with a world-record liquid nitrogen OC, but I think most likely people using engineering sample CPUs, especially a Ryzen 7 4800H, are not running liquid nitrogen OCs. The majority of 2080 Ti actually perform more around ~80 fps in this benchmark, so the unnamed GPU we talk about here is around 29% faster than a regular 2080 Ti, while being 17% faster than the highest overclocked 2080 Ti. New to Geekbench 6 is a new GPU API abstraction layer and new Machine Learning workloads. Test your GPU's power with support for the OpenCL, Metal, and Vulkan APIs. As usual, there's many GPUs that perform somewhere in the middle, with few being faster and few being slower. L1 cache size 256 KB L2 cache size 2 MB Memory specification. Test your system's potential for gaming, image processing, or video editing with the Compute Benchmark. It hasn't really been mentioned here, but that best performing 2080 Ti in the leaderboard is already a really good performing 2080 Ti. Regarding the GPU, it is definitely some kind of unreleased high end GPU, as being 17% above the best performing 2080 Ti is quite a big step.

The CPU does seem to be a Ryzen 7 4800H and people generally seem to agree about that. To be clear, I do not have access to any more information than you, I also just see the result in the leaderboard, beating all the 2080 Ti. I'm very grateful for every kind of help.Hi, I'm the developer of the benchmark (OpenVR Benchmark), and since I was already asked on the OpenVR Benchmark Discord to comment on the AMD benchmark result everyone seems to be talking about, I guess I will also comment here to let people know what I think. I have really no idea what I'm doing wrong here. Using Mat (by changing UMat to Mat and using img = imread('/home/robin/Pictures/Lenna.png', 1) ) the time is 2628712. High_resolution_clock::time_point t1 = high_resolution_clock::now() img = imread('/home/robin/Pictures/Lenna.png', 1) Imread('/home/robin/Pictures/Lenna.png', 1).copyTo(img) I used the following code to determine the speed: #include 'opencv2/opencv.hpp' Moreover, I tried it without IPP and by playing around with the cmake configuration, which didn't change anything. I was able to run the OpenCL test code from the website and 'clinfo' in terminal works, too. I followed this guide to install OpenCL on my machine. Sudo /bin/bash -c 'echo '/usr/local/lib' > /etc/ld.so.conf.d/nf' I'm using OpenCV 3.2.0 (on Qt 5.9.1) and installed it by:Ĭmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D WITH_TBB=ON -D WITH_V4L=ON -D WITH_IPP=ON -D WITH_OPENCL=ON -D WITH_OPENGL=ON. Doing that, however, is way slower than using Mat on my Laptop. To test if there is a improvement in speed, I tried to use UMat instead of Mat. I want to speed up my programs by using Transparent API.

3320m opencl benchmark