core temp speedfan temp. loc?

[bkolk]

Do core temp and speed fan read temperature readings from different locations?
Coretemp is giving me more volatile readings than speedfan. Coretemp will be idle @ 42-45 and jump up a few degrees anytime a small spike in processor usage. Speedfan has more of a stable readout, and a few degrees cooler than coretemp.
And whats the difference between temp1 and core1?

[carnageX]

Speedfan isn't that accurate. I'd trust CoreTemp more than Speedfan. Check your BIOS temps and compare them to what both programs read, and then see which one is closer to BIOS idle temps.

[arcticap]

Core Temp lets you monitor Intel "Core Duo", "Core Solo" (Yonah), "Core 2 Duo", "Core 2 Extreme", "Core 2 Quad", " Pentium E2000" series, "Celeron 400\500" series (Allendale, Conroe, Merom, Kentsfield, Conroe-L respectively), "Xeon 3000/3200/5100/5300" series (Woodcrest, Clovertown respectively) and all AMD K8 (AMD64) and K10 (Phenom, Opteron) series die temperature.
The temperature readings are very accurate as the data is collected from a Digital Thermal Sensor (or DTS) which is located in each individual processing core, near the hottest part. This sensor is digital, which means it doesn't rely on an external circuit located on the motherboard to report temperature, its value is stored in a special register in the processor so any software can access and read it. This eliminates any inaccuracy that can be caused by external motherboard circuits and sensors and then different types of programs trying to read those sensors.

Core Temp - How it works

I been running CoreTemp and comparing it to Everest which measures the motherboard temperature. Sometimes they're the same and sometimes the MB registers 2C higher.
Coretemp does seem to be more sensitive which I think is an indication of it's accuracy because it spikes according to usage so briefly which reflects that it's responding immediately.
If the CoreTemp log is adjusted to record minute to minute (rather than every second), then the small second-to-second fluctuations don't reflect in the log much if at all.
Maybe Speedfan is trying to minimize reading those spikes too.
In other words, the motherboard (Everest) reading may be high, the Speedfan reading may be lower to compensate, and the CoreTemp reading may be in the middle or just right.
At least that's my gut feeling with my computer running an E8400 Wolfdale.
My core temp. is usally always iding normally at 40C in each core, and spikes 1C or so in one or both cores as I click around, reflecting electrical usage.
I have noticed that a hotter household air temp. will affect the core temp. up to 5C or a little more. That's due to air cooling and the number of case fans and such.
I would trust Coretemp. more, & a couple of degrees C. either way only seems to affect chassis fan RPM's.
It's better that the case fans react to the hotter MB temp. than any lower reading anyway, so the fans can react just before the temperature actually starts to build up in the cores and not after.
Since the CoreTemp program claims that the reading is supplied by an internal CPU sensor in each core, that's why I prefer to put a little more faith in that reading. And even if it was off, at least it provides a solid basis for comparing relative temperature fluctuations.

[bkolk]

coretemp reads my load temp @ stock speed (e8400 3ghz) with zalman 9500at cooler as 57C. Little high for a aftermarket cooler isnt it?
I never had the stock hsf on so i dont know the difference.

[carnageX]

Originally Posted by bkolk

coretemp reads my load temp @ stock speed (e8400 3ghz) with zalman 9500at cooler as 57C. Little high for a aftermarket cooler isnt it?
I never had the stock hsf on so i dont know the difference.

That's because you're using a 45nm CPU (e8400). CoreTemp doesn't read 'em as accurate, nor do many other temp monitor programs. If you do a search around this forum, I think somebody may have found a working temp-logger for 45nm CPU's. Unless, CoreTemp now supports them better of course.

[arcticap]

I have a big Coolermaster Stacker case without some of the extras that you have that add to CPU usage and heat. But I do have more [and noisier!] airflow.
For instance, some of the reviews on Newegg about your MB, while great, mention a hotter northbridge. It sounds like a smaller MB too and since you have a larger video card etc..., you're generating more heat in a smaller space and not venting it enough.

My fans will whine like a vaccuum cleaner to keep the CPU temp. low.

Looking at your specific Antec Sonata II case, I think that it lacks enough airflow, there's not enough vents and it's building up heat [like an oven]. My case has vents on every side including the top and bottom. Not that it makes much difference at idle anyway, but it builds up a lot more after you start using more of the CPU etc...

Sonata II:

Newegg.com - Antec LifeStyle SONATA II Piano Black Steel ATX Mid Tower Computer Case 450W SmartPower 2.0 Power Supply - Computer Cases

GIGABYTE P43:

Newegg.com - GIGABYTE GA-EP43-DS3L LGA 775 Intel P43 ATX Intel Motherboard - Intel Motherboards


Here's my tower case that sits on the floor and has wheels (and my ASRock P35 MB is said to run a little hotter too):

Newegg.com - COOLER MASTER CM Stacker STC-T01-UW Black/ Silver Aluminum Bezel, SECC Chassis ATX Full Tower Computer Case - Computer Cases

Today my CPU is running solid between 40 -42C with my apartment temp. about 10 degrees F. over normal, which raises my CPU temp. by 1 - 2C.

[bkolk]

Originally Posted by carnageX

That's because you're using a 45nm CPU (e8400). CoreTemp doesn't read 'em as accurate, nor do many other temp monitor programs. If you do a search around this forum, I think somebody may have found a working temp-logger for 45nm CPU's. Unless, CoreTemp now supports them better of course.

Does the bios read em correctly? b/c the bios is closer to speedfan temps than coretemp.

[carnageX]

The BIOS should read them right. Like I said, do a search around the this board, and you should find a bunch of topics about the issue, and I know one of 'em has a link to one that works. I just don't trust SPeedfan at all, b/c it's given me wacko readings in the past.

[arcticap]

My bios readings are closer to the MB reading of Everest, but the bios readings for the new 45 nanometer CPU's are what many complain about as not being calibrated properly. I read that many times when searching for a new MB on Newegg recently.
That's why I believe that CoreTemp. has an advantage of being more accurate by by-passing the MB and reading from the core itself.
In truth, there could be a fluke anywhere though, but how else can you check the relative hotness of your case?
You said yourself that you're idling 42-45C which is only a little hotter than some other folks "normal". Any other Coretemp readings should be on a relative scale, say within 5C of normal and based on CPU usage or heat build up from there.
Are you reporting the low CoreTemp temps or the highs?
Only the low temps are displayed on the visual display (monitor).
How many fans & vents does your case have?
According to Everest, my harddrive temp. is 31C and the MB temp. is 39C today.
CoreTemp. temp. is 40C. and hasn't moved.

Could your northbridge be touching your Zalman cooler?

What are your harddrive temps.?

Could your bios need adjusting? Who set the bios and computer up? Could there be some need of a voltage adjustment? Are you running factory spec.'s or overclocking?

Everest is available here:

|MG| EVEREST Free Edition 2.20

[bkolk]

Originally Posted by carnageX

The BIOS should read them right. Like I said, do a search around the this board, and you should find a bunch of topics about the issue, and I know one of 'em has a link to one that works. I just don't trust SPeedfan at all, b/c it's given me wacko readings in the past.

I think the program ur referring to is realtemp, heres a screenshot of that.


Tho it sounds like i need to calibrate it from reading this Real Temp - Documentation

I dunno, my worry about all of this is that i want to overclock but i want to make sure everything is ok b4 i start, that my HSF is working correctly with correct amount of thermal paste applied.
my room is currently ~72F (22C) (cool night windows open)
EDIT:
So the difference between the realtemp/coretemp/speedfan reading is simply tjmax is changed between 95/105/100, explainging the offset readings.