Unlock the Potential of AI and Immersive Web Applications with WebGPU.

NVIDIA Vulkan 1.3.271 developer driver 537.96.NVIDIA GeForce driver 546.29 Game Ready.Now if a tool performs some kind of low level sensor access, maybe this access is invisible for the driver and that could explain the spikes… Maybe, actually, the graphics driver properly handles multiple sensor readings but only if these readings are done with official driver API functions. Now a question: why these multiple accesses are not handled by the graphics driver? Then, I will update shortly all my tools (FurMark, TessMark, FluidMark, ZoomGPU, GPU Shark, GPU Caps View) with a spike filter that will use some complex math routines ( if gpu_temp > temp_max then spike! ) 😉

So from time to time, there is certainly a conflict between GPU-Z and ZoomGPU because both tools want to access to the same sensors at the same time… And this is not limited to GPU-Z or ZoomGPU but to all hardware monitoring utilities. But at the same time, FurMark uses ZoomGPU to read GPU temperature and display the monitoring graph. If you look at the screenshot at the beginning of the post, you’ll notice that GPU-Z is running and FurMark displays GPU-Z data.

Now it is possible to enable special math data filtering algorithm independently for each hardware monitoring graph to reject misreading spikes caused by sensor access conflicts when running multiple hardware monitoring tools at the same time Yesterday, a new beta version of MSI Afterburner 2.0 has been published and in the release notes, we can read:Īdded optional data filtering mode for all hardware monitoring graphs. I guess I found the explanation of the spikes in temperature monitoring graphs that some users have noticed with FurMark.

Instaling Geeks3D FurMark 1.37.2