Thermals and CPU Temperatures
Lately, it has been requested if we could list VRM temperatures. We could hook into sensors and start measuring. Now a good way really is to look at the VRM area with a thermal camera. This way you can detect hot-spots and/or worrying stuff. We run the FPU and CPU torture test in AIDA. The chart plot shows the maximum VRM temperature measured at thermal sensor level. The VRM temperatures you see listed in the chart are not based upon the thermal image, but the max temp reported by the thermal sensors at the VRM stages.
- System setup: normal conditions/default settings / 100% CPU load on all cores
Temperatures
We do not table/chart up temperature results because we'd need to apply identically cooling over and over on all platforms. Also, coolers (RPM) react differently to TDP and variables like BIOS settings on all motherboards, let alone brands. The processor peaks towards 95 Degrees C and then very slowly drops towards 85 degrees C. But during long-term all-core CPU stress, expect 85 Degrees C on the processor package.
Above we measure temperature under load. We hit close to 95 Degrees C on the CPU package sensor in Pl2, then after a while settled at PL1 with ~85 Degrees C
Looking at package wattage, we can see a short state of power at 58 Watts (PL2), then after a few seconds, the SoC returns to a more manageable power state, 35 Watts (PL1).
Performance core Clock ratios then, the highest clock ratio is 54x thus 5400 MHz. Throttling occurs when the PL1 state is going back to it's default 35W. .
Energy core Clock ratios then, the highest clock ratio is 40x thus 4000 MHz. However, at PL1 and more cores active, that's x.22