Methodology
A Ryzen 5 3600 has to be cooled at factory settings, in addition 16GB Crucial Ballistix Sport (2x8GB) DDR4 memory on a MSI B550 Tomahawk (BIOS 7C91vA3). The Asus ROG Strix GTX 1060 6GB is running in zero-fan mode for the CPU stress tests only for the image output, and a BitFenix Whisper M 450W serves as the power source, all housed in the Thermaltake Core P3 as an open benchtable.
All records are logged with HWInfo64 v6.30-4240 – 2000ms logging, the specified temperature is derived from the sensor CPU Die (Average). The load of Prime95 produces an average of 88 watts of CPU package power. Since I don’t have air conditioning, the values in the graphs are the delta between room temperature and CPU temperature. This allows all cooling systems to be compared fairly, even if the ambient temperature is different. The thermal paste supplied with the cooler or pre-applied is used in each case.
Benchmarks
And here comes the most important part, the results of the stress tests:
1 fan |
30% PWM |
60% PWM |
100% PWM |
Fan speed 1 |
1030 RPM |
1510 RPM |
2490 RPM |
Sound level / characteristic |
light hum |
Sync and corrections by n17t01 |
Noise |
Ambient temperature |
25,3 – 24,9°C |
23,9 – 24,1°C |
23,1 – 24,3°C |
Well … well cooled looks different. The graphs are so straight because the CPU is permanently in 95°C thermal throttle. Despite maximum speed and open-air benchtable! At this point I resorted to a granted remedy of all ITX enthusiasts: Undervolting. The following are the results with reduced CPU core voltage:
With 91°C in the peak, the best result is still far behind in comparison, but such temperatures aren’t unusual in really small cases (and also notebooks). So that performance is not secretly lost, I checked each with Cinebench R23:
CB R23 Multi / 10 min |
Points |
CPU Delta |
Surroundings |
Stock |
9010 pts |
61,3°C |
24,3 – 24,7°C |
Undervolt, Offset -0.0500V |
9007 pts |
56,6°C |
24,7 – 24,8°C |
Undervolt, Offset -0.0750V |
9066 pts |
55,4°C |
23,1 – 24,3°C |
Undervolt, Offset -0.1000V |
9050 pts |
53,3°C |
24,0 – 24,6°C |
So the “sweet spot” of my CPU is somewhere between -0.0750 and -0.1000V and due to the slightly better temperatures I can even get some performance out of it.
Summary and conclusion
The correct classification of the test results of such a “niche cooler” is of course a bit difficult, especially since I have no direct comparison to offers of the competition, as far as this form factor is concerned. In and of itself, though, I find it somewhat disappointing that the G200P can’t tame even a mid-range processor under optimal conditions.
Somewhat comforting is the objection that even under the best conditions a Prime95 workload is in most cases far away from the application range of an ITX system, which requires such a small cooler. Priced well below the most similar cooler, the Noctua NH-L9a, Cooler Master offers a cheaper alternative for those who need a really flat cooler. However, depending on the CPU, you still have to be prepared for some fine-tuning in the BIOS.
The test samples are purchased by the customers themselves or, as in this case, provided by the manufacturers without obligation. There was and is no influence on the tests and results. An expense allowance is only paid in exceptional cases, but is shown and also has no influence on the test results.
Cooler Master MasterAir G200P (MAP-G2PN-126PC-R1)
lagernd | 49,90 €*Stand: 30.08.24 09:28 |
Noctua NH-L9a-AM4 chromax.black
Sofort lieferbar, Lieferzeit max. 1-3 Werktage | 54,90 €*Stand: 31.08.24 23:38 | |
Sofort verfügbar, Lieferzeit max. 1-3 Werktage | 54,90 €*Stand: 31.08.24 23:41 | |
siehe Shop | 57,70 €*Stand: 01.09.24 00:38 |
15 Antworten
Kommentar
Lade neue Kommentare
Neuling
Moderator
Urgestein
Urgestein
Moderator
Moderator
Neuling
Urgestein
Moderator
Mitglied
Neuling
Urgestein
Urgestein
Mitglied
Urgestein
Alle Kommentare lesen unter igor´sLAB Community →