We've already stated that going into this article we wanted to test GIGABYTE's new thermal solution for the PWMs. Before we could get started, we needed to determine the base temperature.
To start with, we clocked our CPU at 4.0GHz and loaded the processor with various CPU benchmarks. What we found was, the motherboard around the PWM area didn't actually get that hot even after an hour of stress tests. In fact, the only area on this half of the motherboard to see a significant rise was from the PLX chip just under the CPU cooler on the image (M6).
At this point in the test, I'm under the impression that GIGABYTE's new PWM cooler is amazing. With temps like the set we ran we needed to find out what happens when we take the heat sink off.
In the image above, we removed the PWM heat sink and stressed the CPU with performance benchmarks running in a loop. Surprisingly, the warmest IR3550 only reached 45C and a majority of the digital PWM chips hovered right around 40C with am ambient room temperature of 20C.
In the image above you can see a thermal probe going to Measure Point 1.
The temperature probe connects to a data logging device that records three samples per second. In the graph above, we see the temperature ramp up during load of the top left PWM while under load.
In the video embedded above, we see the system with and without the heat sink installed. In both cases, we used a timeline for testing with Cinebench 11.5 with the computer shut down and at room temperature to start with.
Cinebench starts for the first time at 1 minute and 20 seconds. A new Cinebench CPU run starts every 1 minute and 10 seconds after so at 2 min 30 sec, 3 min 40 sec, 4 min 50 sec and finally at the 6 minute mark.
Most of the PWMs reach 40C, but nothing goes over 45C.