GPU-Z gpu hotspot and memory identical

[Necrophagist98]

Hello, correct me if I'm wrong, but only the newest GDDR6X graphics cards have a VRAM thermal sensor, right? It shows up in hwinfo for people who have those GPUs. I have a 2080 and hwinfo only tells me my hotspot and my core temp. However, GPU-Z and HWmonitor shows the same identical reading for both my memory and hotspot. My hotspot is 10 degrees hotter when idle and 15 degrees hotter than my core temp under heavy load, which by GPU-Z's logic means that my VRAM is also 15 degrees above my core temp under load. How can it show me my memory temperature when my GPU doesn't have a VRAM sensor? I hope it's a fake reading or something. I've seen many people's readings and every 2xxx series card I've seen has the same readings as me, in that the memory temperature and the hotspot temperature are always identical. However, every 3xxx series card I've seen has different readings for the memory and for the hotspot. Is it possible that GPU-Z is simply copying my hotspot temperature and that it's showing my real memory temp? I've looked online but have come up with nothing. I'm kinda bummed since I'm getting worried that my memory might be throttling without me knowing, that's why I'm hoping it's a fake reading. Thank you!

 

[W1zzard]

might be throttling without me knowing

Memory doesn't throttle afaik. When it overheats, the thermal protection kicks in and reduces GPU clock, which is visible in the GPU clock sensor and the Perfcap Reason

I've seen many people's readings and every 2xxx series card I've seen has the same readings as me, in that the memory temperature and the hotspot temperature are always identical.

Interesting, I guess that could be used to detect support for memory temperature. I'm just showing the values provided by NVIDIA's drivers

 

[Necrophagist98]

Memory doesn't throttle afaik. When it overheats, the thermal protection kicks in and reduces GPU clock, which is visible in the GPU clock sensor and the Perfcap Reason


Interesting, I guess that could be used to detect support for memory temperature. I'm just showing the values provided by NVIDIA's drivers

So the perfcap reason would be THRM if the memory was overheating? Because even when my hotspot and alleged memory temperatures reach 96/97, I still get no perf cap reason except for PWR.

 

[W1zzard]

The memory limit is probably 100°C or higher. Stop your fans with your fingers and see what happens

 

[Necrophagist98]

The memory limit is probably 100°C or higher. Stop your fans with your fingers and see what happens

So the memory hitting its temp limit will show up as perfcap reason THRM, right?

 

[W1zzard]

I think so

 

[Necrophagist98]

I think so

I understand. Thank you very much for the help! Have a wonderful day!