Hmmkay, here is my test system's spec:
Abit BP6 with dual Celeron 400@552MHz at 2.20V
TNT2
Voodoo2
Network card
the usual drives
80mm case fan in front
80mm case fan in rear
80mm fan in power supply
120mm Panaflo blowhole fan on side aiming half into cards, half into CPU area
Alpha 6035 heatsinks with YS Tech 60mm fans, set to pull away from sinks
I booted up and let the system equalize for a minute, then ran Motherboard monitor and had it average 40 samples, one every 5 seconds. Then I stopped MBM, started the Dnet client, let it equalize for a minute, and started MBM and had it average the next 40 samples. Results:
Full idle in Win2kPro:
CPU1 94.41F
CPU2 95.31F
Motherboard sensor 98.60F
Dnet client running:
CPU1 94.91F
CPU2 95.72F
Motherboard sensor 99.28F
Increase in temperature due to Dnet client:
CPU1 +0.50F
CPU2 +0.41F
Motherboard sensor +0.68
On the Abit BP6, the temperature sensor for the CPUs is in contact with the underside of the CPU, which is not quite as good as direct core readings, but still a pretty good relative indicator. Based on these readings, I would have to conclude that there is not a meaningful temperature increase when running the Distributed.net RC5 client on my system.