Personally I don't like spreading and prefer ball or line simply because it doesn't look nice to me and I've seen a lot worse. Cant help wonder if it's similar to creating a painting, most people cannot produce amazing results and my own would probably be kindergarten level. Also these days see's a plethora of greases. Not sure if there's a quality control issue with some results too.
Here's an old result from some years back using a specially crafted piece of code to produce constant high core temperatures with Haswell. Using a quad core mobile bare die CPU high thermal latency of the waterblock was kept out of the equation using a duty cycle of 50ms on time to measure temp and 950ms off time. Clocks were increased every 3 seconds or so.
The ZF12 is quoted as 12W/(m⋅K) hence the 12 I suppose. At 4.5GHz power is 200W, not bad for a 47W TDP. Actually at 98C a couple of Watts over and at 77C a couple of Watts under, likely due to increased leakage at higher temps. Extrapolating data for PK-1 gives a BLT of 80 microns @ 10W/(m⋅K) or could be 50 microns at 6W/(m⋅K) etc, I don't have a way to measure actual BLT. The GD900 thermal throttled above 4GHz and came from a tub so was spread on the WB while the other two used the line method on the die IIRC. Note that contact pressure is relatively low due to using bare die and not wanting to crack it. IMO should have used delta temperatures really and better still PECI temperature which produces a rolling average of highest temp to 1/64th of a degree.
@TPCEA I might be behind the times but some pastes that might be of interest to you
Code:
Halnziye
HY410 30g 1.93W/(m⋅K) $1.10
HY510 30ml 1.93W/(m⋅K) $0.30
HY610 25g 3.05W/(m⋅K) $1.30
HY880 1g 5.15W/(m⋅K) $0.50
HY883 1g 6.5W/(m⋅K) $0.50
FNE-210 25g 1.93W/(m⋅K) $1.30
Maxtor
GTG8 4g 12.8W/(m⋅K) $2.60
CTG9 4g 13.5W/(m⋅K) $6.40
Subzero seven
LTP-15 10g 15W/(m⋅K) $4.00 (Liquid Thermal Pad)The LTP is a paste that is applied and becomes a pad with heat IIRC.
