So I built my first small form factor PC in a half depth 1U server chassis last night. I was a little concerned about airflow and overheating as the wires from the power supply crisscrossing the internals and USB header and front panel wires added to the rat's nest. This was going to significantly decrease airflow and create a possible heat problem. I originally wanted to go with a 35w or 45w CPU but they were all either on back-order or way over budget so I decided on a low spec 65w CPU to at least get me started. The local Microcenter had the G620 for $60 so I picked one up.
I always do an Intel Burn Test run to make sure my builds are stable even if I don't OC. I like to keep watch during the Burn Test with CoreTemp. I feel like if I can keep the CoreTemp under the tCase for the CPU in a 21'C room and run 10x burn test on High the build passes.
tCase for the G620 is 69'C
During the Burn Test my CoreTemp reading maxed at 63'C and started at 31'C so that was fine but not surprising - what was surprising - at least to me - was that the power draw never went about 26W even when both cores were maxed out. I did not expect a 65W CPU to max out at 26W draw . I am loving it
So my question is twofold:
Are the CoreTemp wattage numbers accurate?
Is it normal for a CPU to max out at 40% of TDP or did I just get an awesome part randomly?
I would not be suprised. Intel overstates their TDP all the time. Its normally a "worst case scenario" deal with them.
Just as an example, the Core 2 QX9650 had a 130w TDP but under max load at stock frequency it would use about 73w, about 56% of its max. The only way it even used 130W was when it was clocked to about 3.8GHz-4GHz, depending on the chip.
So to answer your questions, yes they are probably right and yes its normal for the TDP on Intels CPUs to be well over whats really used.