On Mon, 26 Feb 2018 16:56:27 -0500, Flasherly
wrote:


Core Temp lists 176F C_upper threshold

I adjusted, took that down, through the program's user-adjustable
offset setting, by the amount of 46F.

It then reads amount below to reach C_UT is now 130F (176-46)
eg: A 30F reading means temp is 100F

Then, today, I maxed the CPU load out for a neg. Core Temp: -6F
reading.

(Spread-core gain analysis over a directory of 26Gbytes of audio
files;- it's very fast, no more than four, five minutes for returning
appropriate gain reduction tables.)

-6F meaning: 1) either I'm at 136F, or 2) Core Temp's author, by
exceeding AMD's published 158F (18 degrees in addition to derive 176),
gives me an additional 18 degrees lower. I'm not sure I'm misreading
him, where he mentions AMD "won't publish the specs" for an exact
FX-CPU temperature, other than for an allowance of "10-20F"
uncertainty.

That would place me at 120F, if I use the 18F and his mention for
where he's coming up with a C_TH of 176F, when I maxed the CPU cores
out.

I did turn, as well, off the AMD processor Turbo Boost feature prior
to, a day or so ago, loading the CPU down. A BIOS default for TB is
4.2GHz, I had, while running TB, lowered that to no higher than
3.6GHz;- I've never messed with the base clock of 3.2Ghz. I did,
however, notice an immediate 10F lower operational temperature, from
what it had been with the Turbo Boost enabled.

If I'm supposed to not be sure, therefore, maybe it's running "hot" at
120F, which is nice, while not nearer to 140F. Speedfan also runs on
the same temperature sensor bus, and will also derive the "offset"
reading;- what Speedfan does do, that Core Temp doesn't, is as well
provide three additional sensor readings. All temps are linked to the
same Gigabyte sensor chipset -- one being ambient case temp,
potentially, whereas the other two correlate closely until CPU
temperatures rise, for one sensor to follow more closely a higher
reading. A latter and higher reading, still, that I'd offhand say
seldom exceeds 100F (I should have more closely watched for SpeedFan's
upper limit when loading the CPU down today).

Or, possibly, I'm not supposed to be exacting about trying to balance
the two programs into cumulative certitude.

-
Cumulative Distribution Function
(noun) Statistics - a function whose value is the probability that a
corresponding continuous random variable has a value less than or
equal to the argument of the function.