Temperature behavior
Two practical scenarios are basically simulated to evaluate the temperature behavior. On the one hand, a very high CPU load, as it occurs when rendering videos, for example, and on the other hand, a gaming load, as it occurs when playing current games. Because the hardware and cooling solutions can differ greatly from test to test, I have now banned direct comparisons from my test procedure. However, the graphics card used is often identical, which is why conclusions and comparisons can be made at least partially.
Test system
The following hardware was used for the temperature measurements:
General conditions
- 24.8°C room temperature
- Case fan fixed at 800rpm
- CPU fan fixed at 800rpm
CPU Torture
Cinebench R23 was used for the CPU torture test in the loop. Since the air cooler doesn’t need a long warm-up phase, we could start monitoring after about 10 minutes of warm-up. Although the 5800X3D is a relatively frugal CPU, it has a very high thermal density due to the stacked cache and gets hot accordingly quickly.
Under full load, I measured an average temperature of 83°C, which is a very decent value considering the hot head and the relatively high room temperature. The CPU was slightly undervolted via the Curve Optimizer and was thus able to maintain an average clock of 4325MHz at 1.20V on all 8 cores in the benchmark. The difference between the large custom loop (420mm radiator, large AGB) was surprisingly small in this test (5-6K).
Gaming
Since I will most likely be using the case for quite some time, I was particularly interested in the temperatures of the vertically mounted graphics card. Therefore, I used Borderlands 3 for the gaming test this time. As recent tests have shown, the Unreal Engine in this title demands everything from the graphics card and also reliably reveals hotspot problems. Thus, it is ideal to validate the suitability of this mounting method in conjunction with the graphics card used.
In the game, the 5800X3D ran relatively warm with an average of 65°C, which isn’t surprising since the 6900XT blows its almost 300W TDP directly into the massive CPU cooler. In practice, this isn’t a big deal, because first of all, the allowed temperature for the CPU is 90°C and since I run the graphics card undervolted outside of tests and it’s also seldom continuously loaded to 100°C, the temperatures are not too high usloaded, the temperatures are well below the test values under regular gaming load.
The horizontal arrangement of the heatpipes of the graphics card used in conjunction with the good airflow of the case and the large distance to the side panel is apparently absolutely uncritical. At less than 1500rpm fan speed, the graphics card’s clock speed was continuously maintained at >2400MHz and the temperatures of ~75°C core and ~90°C hotspot were also absolutely harmless.
Interim conclusion
As expected, the temperatures in the Dark Base Pro 901 are right up front. Due to a lack of constant room temperature, different fans etc., a direct comparison with other cases is unfortunately not feasible within the scope of my possibilities. However, I can report that the graphics card runs somewhat cooler in the Dark Base Pro than in the Corsair 7000D Airflow and the fans don’t start up as often when idling – but I must also say that a radiator in the front of the 7000D hinders the airflow somewhat. In general, however, there is a few centimetres more space around the graphics card in the BeQuiet! case, which will benefit the temperatures. The installed Silent Wings 4 fans can be controlled in 400rpm steps via the internal fan controller of the Dark Base Pro 901. At the lowest speed of ~400 rpm, the fans are practically inaudible, even though I keep the front and lid open (and thus not insulated) in favour of airflow.
At 800 rpm, the fans start to become audible, but they move noticeably more air. For a longer gaming session, this is exactly the right choice, because the waste heat from the components is easily removed from the case and the noise level is still below that of the graphics card fans. The settings above this (1200/1600/2000rpm, whereby the Silent Wings are specified with 1900rmp max.) are then only relevant for tests or benchmark runs, because they are simply too loud for sensitive ears and normally not that much air needs to be moved.
85 Antworten
Kommentar
Lade neue Kommentare
Urgestein
Moderator
Urgestein
Urgestein
Urgestein
Urgestein
Urgestein
Veteran
Urgestein
Mitglied
Urgestein
Urgestein
Moderator
Urgestein
Veteran
Mitglied
Veteran
Urgestein
Mitglied
Alle Kommentare lesen unter igor´sLAB Community →