As we said, the idea is to be able to judge the impact of a liquid cooling solution against the original cooling system of our Gainward RTX 4090 Phantom.
The comparative tests will take place in two parts. First with the card at “stock” frequencies and then we’ll push the frequencies with the 666W Galax HOF Bios.
Performance at stock frequencies:
Here’s the table of our results with the stock cooling system and the Alphacool CORE RTX 4090 waterblock. This first table compares the performance of the cooling systems on 3DMark benchmarks where we’ve run a run three times in a row. The idea is to see how the temperatures behave, but also whether the Boost frequency increases if the cooling is better.
During our tests, the temperature in the lab was around 27°C. Well, not surprisingly, temperatures are excellent with the liquid cooling system, even if in terms of performance, there’s not much difference given the performance of the original air cooling.
As we can see, the performance of this Alphacool CORE RTX 4090 is spot on. The gain in GPU temperature is between 7 and 17°C, which isn’t bad at all. We regret, however, that the contact with the chip is not perfect, as it would be possible to have less than 15°C difference between the water and the GPU at 450W, compared with 20°C here.
To push things a little further, we wanted to measure overclocking capabilities with this liquid cooling. We therefore compared performance on 3Dmark with and without overclocking using the Bios 666W Galax HOF for RTX 4090.
No miracle here, the chip doesn’t even manage to reach the power limit here, even at 1.1V. The very low temperature also helps to avoid frequency variation and therefore better performance. For power consumption over 535 W, we stayed below 27°C in relation to the water, which is still acceptable.
Now, bear in mind that switching to liquid cooling comes at a cost. You’ll need to invest not only in the waterblock, but also in a liquid cooling loop.