Measurement procedure and results
I used Red Dead Redemption 2 for the measurements. In 1440p with ultra details, the RX 6900XT works up quite a sweat and pulls in around 280W (GPU only, read with GPU-Z) almost continuously despite the partial CPU limit of the lame Ryzen APU, which makes for a considerable waste heat.
With a room temperature of 20.8°C measured, the temperatures in the case leveled out after about 15 minutes of gameplay and I could start the test run, which also lasted about 15 minutes.
The following measurements were performed:
- Open Case – 2 x 140mm front IN, side panel open
- Closed Case – 2 x 140mm front IN, side closed
- +1 fan – 2 x 140mm front IND, 1 x 120mm rear OUT, side panel closed
- +2 fans – 2x 140mm front IN, 1 x 120mm rear OUT, 1 x 140mm top out, side horn closed
- +3 fans – 2x 140mm front IN, 1 x 120mm rear OUT, 2 x 140mm top out, side horn closed
All case fans and the fan on the CPU cooler were fixed at 800 rpm for comparability!
Closed Case
With the case closed and only the two standard front intakes, some heat builds up in the case, but it does not yet lead to reduced clock rates or thermal throttling. With 95°C GPU hotspot temperature, this was the hottest test.
+1 fan
Here, in addition to the standard front intakes, a 120mm fan was mounted as an exhaust in the rear.
Both the average GPU temperature and the hotspot temperature were 2 Kelvin lower than the previous run despite a lower GPU fan speed. In return, the CPU got a bit warmer, contrary to expectations. I assume that the exhaus fan pulls the hot air upwards and thus heats the CPU cooler additionally. However, with a difference of just 0.3 K, it could simply be a measurement inaccuracy.
+2 fans
In addition to the last scenario, another 140mm fan was mounted in the lid as an exhaust.
With the additional fan, the temperatures have improved further. GPU, GPU hotspot and also the CPU have remained over 1 K cooler on average, the GPU fan speed has further reduced.
+3 fans
In this scenario, another 140mm fan was mounted in the lid as an exhaust. Thus, we have achieved the full configuration of all fan slots.
Due to the full configuration, all temperatures could be reduced by 1 K again, while the GPU fan speed dropped once more. However, a normal mortal will not hear any difference between the determined 1515 revolutions of this test and the 1560 revolutions of the previous test.
Open Case
For the sake of completeness, however, I then also performed a run with the side panel open. Compared to the full configuration, all temperatures dropped again by 1 K, the GPU fan speed by another gradation and the GPU even boosted by 30 MHz on average. The fact that the temperatures only changed so slightly when the side panel was open definitely speaks for the fact that the airflow in the North works very well.
29 Antworten
Kommentar
Lade neue Kommentare
Urgestein
Urgestein
Veteran
Veteran
Moderator
Veteran
Veteran
Mitglied
Mitglied
Mitglied
Veteran
Veteran
Mitglied
Moderator
Mitglied
Veteran
Urgestein
Urgestein
Alle Kommentare lesen unter igor´sLAB Community →