CPU Temp Wrong?

[SO•Audiojack]

(2014-06-01 12:20)Pipa Wrote:  

(2014-06-01 09:45)Malibu Wrote:  What I am saying here is just what I was taught by really experienced guys when I started overclocking. You don't have to follow it but I am just trying to help.

Also bsod crashes during overclocking are very rarely related to the temperature, far more likely that you still need to tweak settings to make it stable.

Mine wasn't even overclocked

[Pipa]

(2014-06-01 12:27)Audiojack Wrote:  Mine wasn't even overclocked

I wasn't talking about yours and we weren't talking aboub non overclocked cpus either. We were talking about overclocking the i5 sandy up there.

[SO•Audiojack]

(2014-06-01 12:30)Pipa Wrote:  

(2014-06-01 12:27)Audiojack Wrote:  Mine wasn't even overclocked

I wasn't talking about yours and we weren't talking aboub non overclocked cpus either. We were talking about overclocking the i5 sandy up there.

I was only reinforcing this statement:

(2014-06-01 09:45)Malibu Wrote:  I didn't mean that it will definitely cause your system to fail, I meant that the higher you go over that limit, the more likely you are to get random failures/crashes (BSOD). And again the Tjmax is the limit where it starts to throttle.

[Cola]

Guys please stop the discussion, as it doesn't really answer the question originally stated by OP.

[Pipa]

(2014-06-01 12:30)Pipa Wrote:  I was only reinforcing this statement.

Well simply stating your pc was having stability problems due to heat, may well be true but it has no relevance to the topic we were discussing. I am not saying that he may not be right in general, but from what i have read while overclocking my system which uses a similar processor it is incorrect.

Malibu was stating that heat has a correlation to system stability while overclocking, but if you look at good overclocking guides for the i5 you will notice that while system stability issues are to be expected they are very rarely caused by heat. Even at 100° the cpu starts to throttle itself, but keeps running nevertheless.

Feel free to check the link again i posted:
http://www.overclock.net/t/1198504/compl...ck-edition

(2014-06-01 16:17)Cola Wrote:  Guys please stop the discussion, as it doesn't really answer the question originally stated by OP.

There is hardly any harm in this. Not at the the throwing feces stage yet.

[SO•Audiojack]

Yes, but what I think we're both saying (me and Malibu) is that in a nutshell the stability of any microprocessor is always proportional to its temperature.

This is not based on overclocking but electronics in general, because a semiconductor's resistance changes with the temperature. And in the most basic terms that just makes it perform worse.

Luckily modern semiconductor technology allows for a pretty wide range of temperatures, but you still can't exceed the specifications without hindering performance (and causing errors.) Of course the computer can cope with these errors to an extent so the end user won't notice it until you get the BSOD

[Pipa]

No issue with that and said so before, but such general knowledge may only work in general. So if you actually suggest advice based on a specific processor it would be good if he could look up some guides before doing so.

[Malibu]

Clearly we have different thoughts on this, it's up to the OP who's advice to follow.

[[TC]»Elmo«]

For the record, Tcase and Tjmax are two different sensors.

Tcase is the temp at the external heat spreader. (The metal cap of the cpu)
Tjmax a.k.a. 'core temp' is the internal temperature of the silicon.

The core temp is usually higher (as is its designed maximum). You don't want to ever exceed either of them, although the core temp is more critical.

Some CPUs have 1 sensor per core (numbered 0,1,2,3 etc), 1 somewhere deep in the chip (often just called 'core') and 1 for the heat spreader/cap.

These different sensors are usually the cause of different readings, although it's not as clear cut as that.
Unfortunately, the sensors don't actually give a value in degrees. Each chip tends to use a different scale for the temps, so the software has to convert it into degrees somehow - some programs have more accurate conversions than others.
In theory, the BIOS should be an accurate source, but the CPU is probably running a different mode in BIOS vs windows so comparing the two isn't always that helpful.