Temperature gradients and boost clock in detail
The cooler honors its name and keeps the card sufficiently cool. After all, 70°C in open construction and a maximum of 75° in the closed housing are now nothing to scare you. The higher Power Target ex works ensures that the clock still plays along quite well. The Founders Edition is much slower and, above all, hotter.
In the stress test, of course, the clock rates break down a little more clearly.
And now the whole thing again in sober numbers in table form:
Initial MSI GeForce RTX 2080 Ti Lightning Z |
Final value MSI GeForce RTX 2080 Ti Lightning Z |
Final value GeForce RTX 2080 Ti Founders Edition |
|
---|---|---|---|
Open Benchtable | |||
GPU Temperatures |
42 °C | 72 °C |
77°C |
GPU clock | 2010 MHz | 1935 to 1950 MHz |
1680 MHz |
Ambient temperature | 22 °C | 22 °C | |
Closed Case | |||
GPU Temperatures |
43 °C | 75 °C |
80°C |
GPU clock | 2010 MHz | 1905 to 1920 MHz |
1665 MHz |
Air temperature in the housing | 25°C | 51°C |
Board Analysis: Infrared Images
The following infrared images show the gaming and torture loops in the open structure and in the closed housing. The differences are very clearly visible, but the cooler still acts quite confidently. As with the Founders Edition, the problem zone is more about memory than GPU and Power Stages. At least not on the left. I had already mentioned in the video that the right side with the 3 phases (6 voltage converters) was treated somewhat stepmotherly.
The very large card with the high waste heat feels visibly uncomfortable in too small housings, which also explains the significant temperature increase. The memory is hot, but not yet in the red danger zone.
The stress test produces a very similar picture, whether it’s a gaming or torture loop.
But even in the Torture loop, the memory remains just as limited in the shot housing.
Since the memory modules are also indirectly cooled via the backplate, they are of course a little cooler when the backplate is mounted. Test measurements with mounted backplate and thermal sensors revealed a delta of 1 to 2 degrees at the bottom of the board. However, how much of this really still arrives at the top of the modules can only be guessed at. However, the built-in heatpipe is rather superfluous, as the two following pictures show.
When everything is thermally heated, it is even clearer that only a very small proportion of the backplate can radiate heat at all. Well-intentioned in theory, but in practice the carbon solution is rather inefficient. What a pity.
Kommentieren