Mau1wurf1977 wrote: So it depends on the board really.
Most definitely. 😀 I only bring it up because my 2.0GHz Willamette has more-often-than-not been a problem with newer 478 boards, and having bought a board or two over the years that end up being no good for it I don't want to see others repeat the same mistake. 😊
I knew there was an Asus (P4P800 if I'm not mistaken) that supports basically everything (iirc it can even take Pentium M with an adapter 🙄), but otherwise I'm not aware of many other "new" boards that play nice with Willamette. That ASRock looks very interesting. 😎
PcBytes wrote:
Well,the temps were those:
Celeron D - 67-70*C
Willamette 1.7GHz - 45-50*
That's too hot for the Celeron D, and I'm honestly surprised to see such an outlandish difference. I'm suspecting there might be a mounting or measurement discrepancy (e.g. one chip is reporting inaccurate information).
Well,it seems the Willamette is cooler than the Celeron D.How does that happen?
It's ~9W TDP lower; that shouldn't produce a 20* C difference though. 😳
If it helps,CPU thermal paste is "Manhattan".(that's the brand,not where I bought it from)
Never heard of it, but that isn't surprising (there's tons of different thermal paste formulations out there).
Also,I tried overclocking the Willamette from 1.7 up to 2.50GHz (so a bit lower than the Celeron D).
While the BIOS was showing the correct speed of 2.50GHz,Windows was detecting it as the normal 1.7GHz speed. I realized that was the CPU name string and not the real speed it was running at 🤣 (I know that from a Foxconn BIOS,where I overclocked once the Celeron D up to 3.0GHz but the CPU name string was still 2.66GHz)
2.50GHz is world-record class for a 1.7GHz Willamette. 😲 Did you verify this with CPU-Z? How hot did it run at 2.5GHz? Also I'd say submit your results to HWBot: http://hwbot.org/benchmark/cpu_frequency/rank … t=0#interval=20
Anyways, to your original question I'd probably upgrade the heatsink to one of the Intel copper-core models and use the Celeron D. Otherwise use the Willamette because at 70* C you're risking damage to the Celeron.
Mau1wurf1977 wrote:Got no Northwood P4 lying around?
A 2.4, 2.6 or 2.8 model should offer a good compromise between performance and power draw.
Northwood is built on a smaller process and runs on a lower voltage, so it should run even cooler 😀
Technically the Celeron D is lower voltage and smaller process (its a 90nm part; Northwood is 130nm), but I'm very surprised by the temperatures being reported. I know when I had one a few years ago it didn't approach that level, even dramatically overclocked, but it was using the copper-core heatsink.
Northwood *is* however a better power/performance compromise, as many of the lower clocked ones run under 70W TDP and perform better than Willamette. 😀