Temperature Results and Analysis

We installed the new Ballistix modules in a Gigabyte GA-H55N-USB3 motherboard, and set the XMP timings in the BIOS. Next we installed the MOD software in the Windows 7 Home Premium OS and proceeded to test it under Prime 95 and in Crysis for 15 minutes each, while the system was left idle in between to settle the temperature back to a level value.

We also recorded the temperature during the test in a CSV file which we've also plotted below - mostly for our own curiosity.

We compared the recorded temperatures against a third party temperature sensor we manually placed between the PCB and heatsink (above the SPD chip) to see what the comparative difference was.


Both the temperature sensor and the thermal probe reacted almost in sync to temperature changes, so we can theorise the on-chip probe is reading a little hotter because of its direct contact with the PCB underneath.

The obvious downside to the sensor on the Crucial DIMM is it's only reading the PCB temp though. It's not closer to the actual memory chip, which is far more difficult to engineer. Unfortunately it certainly can't be on the memory chip, because they simply aren't made with on-die thermal sensors.

No matter how close we tried to put our manual probe to the memory chips, the temperature was always a degree or two lower than the memory itself claimed: could this be an indication that Crucial slightly over-compensated on the reading because of its placement? It's difficult to say.

Either way, the bottom line is that the sensor on the Crucial DIMM really works and it's a rare new feature for memory; one that might attract a few who love to know the ins-and-outs of their PC. If DDR3 temperature reading takes off in other system monitoring software already covering the popular parts such as CPU, GPU and case temps, the benefits grow very obviously, but unfortunately it's both waiting for DDR3 manufacturers to act.