Might be kinda tough with the way [currentyear] AMD is interpreting their VRAM temps
I don't see too much relevance with some of the comparisons to RDNA2 in here though, since the huge jump in VRAM temps happened in RDNA3 and seems to be continuing into RDNA4
Code3 wanted to shows us how cool he runs his card on page 12, in 2 min Furmark test
. I pointed him towards Superposition 4k optimized 20 min test cinematic loop.
The only relevance is how older cards run their VRAM temps compared to new ones. Is not much but, we have to start at some point. Whoever is willing to share their tests I'm thankful.
Erryone should be using a high performance chassis when using enthusiast grade hardware
As an enthusiast, I hate to see hot hardware, unless its doing something cool.
My case is far from extreme hardware, glass panel is 3cm away from the GPU heatsink exhaust(hot air recirculation). During tests can get over 40 C or gaming 1 hr long sessions.
Grills are horrendous, that's why I cut the rear grill fan off.
Yeah right, Lian Li OD11 Dynamic XL ROG Nullified "
Extreme cooling" with with suffocating GPU exhausts by the side panel.
Still, with thermal putty instead of the old leaky pads and fan on the back plate it does well in 25 min 8K optimized cinematic loop Superposition.
View attachment 393187View attachment 393191
Challenge accepted
But I don't think so given that in 2 minutes the temperature stabilizes. For a 20 minute stress test Superposition Benchmark requires the paid version, a well ventilated PC case and a good heatsink are everything, in any case as requested here is the fresh screen just out of the oven....rofl Furmark pushes the vram more and it stabilizes at about 66°-67° while cinebench uses it all but no 3d so I don't understand the point furmark is more than enough since in a real scenario there is no such thing and we all know that once a temperature is stabilized the rest doesn't matter, if the new 9700 xt work at higher temperatures I don't see the problem here we say, you wanted the bicycle so pedal.
So we have: cinebench 20min, furmark 20min, superposition 2 bench in a row
Thanks.
You can do Superposition Free test in Cinematic loop as I did, see test bellow. My 3 fans run at 1400 RPM during the test but,
are 2.5 cm thick fans(deshrouded card) compared to
1cm stock fans and is not something you should ignore. Maybe you should run yours at 1800 RPM.
You can do 8K optimized and I promise you is more demanding than Furmark over your VRAM usage and temps. As you can see Furmark uses 700 mb Superposition will use around 7000-8000 MB in 8k optimized.
This how you do Cinematic loop in Superposition 8K optimized as you have 12GB VRAM, so is ok:
The most important value you see already encircled with red in HW info image
Average(media) temps and
time running. Than of course GPU temps , Hot spot and VRAM values should be visible.
Note: Hwinfo has to be open seconds before starting the test or after 3 min warm up otherwise Average(media) temps column will not show accurate temps.
When you decide the test is over, let's say VRAM to hot or 20-25 min passed(I did 25 min in my test) take a screen shot of HWinfo and superposition in 1 click and save it, make sure HW info shows the time.
I'm on other machine atm that is why you see R9 380 Nitro my main PC card is 1080 TI 11 GB.
