CPU Fan Profiles

[watercooled]

They should probably drop some more cores in there then, not that that's a valid reason anyway...

It's just a reason I've heard, something to do with thermal cycling of soldered dies being more likely to cause cracking with small dies. I'm not sure how true that is.

But either way, they could really do a much better job of attaching the IHS, even if it really is impractical.

[Mr_Jon]

Yes, controlling the fan using SpeedFan overrides it.

A most useful program. The "warning" threshold temp for relevant sensor can be bumped to whatever for kicking the fan to 100% anyhow. I became well versed at that by not realising that I'd set it on one sensor, and then spent a good half hour staring blankly at it wondering why it was happening.

[watercooled]

I think to start with I'd set it at something around 85C - not because that temperature is worrying in itself, but because of the noise I'd be aware of it and could adjust the ramp based on it.

[Mr_Jon]

Yeah, I worked backwards from an initial high point. I picked 60 (62 is the throttle point on my cpu, and it's rare to go above 55 even when stress testing) and ramped down from there. The graph resembled more of a death-slide than a curve though, as when the fans rarely need to get above 75%.

[Xlucine]

It's just a reason I've heard, something to do with thermal cycling of soldered dies being more likely to cause cracking with small dies. I'm not sure how true that is.

But either way, they could really do a much better job of attaching the IHS, even if it really is impractical.

I'm doubtful of that. A small die ought to do a better job at keeping the whole thing at a similar temperature, whereas a big die would have a bigger temperature differential so more stresses due to the different strains

[watercooled]

Ughhhhhh: https://www.overclock3d.net/news/cpu..._be_soldered/1

And they can't be using the 'it's too small to solder' excuse with this one, surely.