Are my temps at load safe?

[jzodda]

Maybe you guys can give me some advice

I recently acquired a TEC air cooler (amanda titan)

I have a 6700 running at 500FSBx7 (3500) on my Quad GT

Vcore is 1.50 bios (1.46 actual at load according to Abit Uguru monitoring software)
Mch is 1.60 and DDR2 is 2.35. Using AS5 and contact is good. CPU is not lapped.

orthos blend is stable 12+ hours and running as I type this but my load temps are moving between 64-68c during the test. I realize that at high FSB it runs a little hotter but is this a safe temp for 24/7 operation?

Is this temp normal for the settings I am running at? I am using coretemp .95 in vista to monitor the temps.

I am willing to give up 500 FSB if I am running at an unsafe temp but I honestly don't know whats normal at this FSB and settings tempwise.

 

[Eastcoasthandle]

I will say negative!. It's reasons like this why people decide to water cool their CPU.

 

[ghost6303]

anything under 70c in orthos is completely fine. anything under 80 is supposedly safe for these chips (shutoff is 85), though i like it in the 60s for 24-7 operation. under real world useage you wont see the kind of heat that orthos produces anyway, so 68 with orthos is perfectly fine.

 

[Eastcoasthandle]

anything under 70c in orthos is completely fine. anything under 80 is supposedly safe for these chips (shutoff is 85), though i like it in the 60s for 24-7 operation. under real world useage you wont see the kind of heat that orthos produces anyway, so 68 with orthos is perfectly fine.

I am going to have to disagree here. I had a Intel 660 and with just air cooling, oc to 4.2 it would get a toasty 57C an would start throttling around 60C (if memory server correctly). Ambient temps along with case temps and cpu temps would effect my chipset, ram and video card on a warm day. Sure, you can get away with 65C when all you have to do is open a window on a cold winter day. But when it's warm outside (still to cool for AC) you will find yourself either lowering your OC or not using your PC at all. Besides, are you sure you are not throttling at that temp?
RMClock
CoreTemp

Good post discussing it here
Again, if you want to OC with temps that high it's better to water cool the CPU.

I think your method, unclewebb, is flawed. You cannot concern yourself with just Orthos alone you have to factor in ambient temps inside the case as well.

 

[ghost6303]

core2 duos are completely different do your homework

 

[Eastcoasthandle]

core2 duos are completely different do your homework

Stop being a immature by posting misinformation that it's ok to have temps up to 70C because you have a C2D cpu. And, deal with the fact that I don't agree with you.

 

[unclewebb]

Nice OC jzodda.

Sorry Eastcoast but you're the one posting misinformation.

CoreTemp reports the core temperature of Core2Duo processors and that temperature is not in any way comparable to a temperature measured from a different location on a totally different family of processors.

The most important thing to watch for is stability and if jzodda is able to run Orthos stable for 12 hours then obviously there is nothing wrong with his temperatures of 64C to 68C.

Set CoreTemp to "Show delta to Tjunction temp" and that tells you how close you are to the throttling point. jzodda is nowhere near throttling. The C2D starts to throttle when the delta to Tjunction gets down to about 2 which is equivalent to an absolute core temperature of 83C. Even then, my processor didn't catch on fire or get damaged, it didn't even crash and it continues to run just fine.

As long as you are stable and keeping your temps under the Intel specified maximum then there is absolutely nothing wrong with jzodda's overclock.

Eastcoast: From that link you posted:
"At T_case_max, the processor will begin to throttle.
According to this, T_case_max is 60.1C ."

That is simply not true for the C2D. Throttling is based on the core temperature and not T_case_max.
This is the reason Intel went to a digital on die sensor. So users can reliably run their processors hotter and get more performance out of them.

 

[unclewebb]

This picture confirms that the C2D is running just fine even with Core0 at 82C.

TAT shows "Thermal Monitor Idle" in green and CPUz confirms that no throttling is taking place.

If these processors were being damaged at this level then Intel would have wisely set the throttling and shutdown point to a lower temperature but they didn't.
They even offer a 3 year warranty which shows that Intel stands behind their processors, 100%, even when operating at these extreme temperatures.

Beyond this level if the processor continues to get hotter it is designed to shut itself down which prevents damage and RMA claims.

jzodda: With your memory at 2.35 volts, you'll likely damage that long before you ever hurt your C2D.

 

[Eastcoasthandle]

What you guys don't understand and don't comprehend is the fact that ambient temps and case temps do play a role in how hot your CPU can get. The truth of the matter is your omitting overall case temps and how higher cpu temps effect your ram, chipset and video card (to name a few) Posting a ss of your CPU at 82C doesn't prove your point, it actually disproves it IMO. The truth of the matter is your omitting overall case temps and how higher cpu temps effect your ram, chipset and video card (to name a few)

Your delta to tjunction temps read 2C/5C away from tcc activation (throttling) using coretemps. From looking at this photo (if core temp is recognizing your C2D correctly) its very possible to start throttling at around 85C. Core Temp should able to read dts (not absolute temps). From what I read throttling is suppose to occur at the point where one of the cores shows exactly the same value as is displayed as tjunction in coretemp. It's the difference in degrees C between the core temperature and the maximum juntion tempearture of the processor. Is this not correct? Do you have any information to suggest otherwise? Do you have any Intel specs that say that C2D don't throttle?
It's incorrect to believe that:
-If your C2D does throttle it suppose to throttle be it 85C or whatever.
-It's normal to have temps that high
-you are suppose to OC your CPU until the heat build up from Orthos shuts your Computer off (do to heat).
-want to operate a CPU at those high temps which also effect other components inside your case
If a C2D does throttle itdefeats the purpose of OC because it slows the performance of the CPU down do to the heat.


If you want to OC at that range with temps that high it's best to water cool. That way you lower your temps dramatically (depending on water cooling setup).

Edit: It appears that the C2D actually do throttle
Section Chapter 13. Section 4

Yup, looks like throttling to me. So I guess C2D also throttle, just like the P4's (found this while browsing).

 

[jzodda]

Thanks for the info guys

BTW this system is in a very cool basement thats cold in winter and even colder in summer with central air on. My system temp is reported at 23C in Uguru and I don't know the room temp is but its never warm down here

As to the ram well OCZ warranties it up to 2.35 (or 2.40 depending on who you ask) I think its 2.35 and so I try to keep it no higher then that.

And yes I have all the "CPU features" disabled. I would never have hit 500 FSB with them on

 

[ghost6303]

umm ok. i wasnt trying to sound like a dick before i was just posting in a hurry, sorry. but it doesnt change the fact that your info is not accurate. just because my processor is at 75 degrees doesnt mean my ram is overheating or ambient temps are too high. the question was:

orthos blend is stable 12+ hours and running as I type this but my load temps are moving between 64-68c during the test. I realize that at high FSB it runs a little hotter but is this a safe temp for 24/7 operation?

and the answer to that question is yes, it is within the limits of being safe. he said/asked nothing about throttling or other hardware temps. the only way your computer would throttle your cpu due to overheating is if you have all the failsafes in bios and hardware monitors running. programs which are notoriously inaccurate. when i put my ram voltage to the defaul 2.2v in bios, it changes to a blood red number as to indicate a high voltage setting. is 2.2v going to kill my ram? no its what it needs to operate. the bios has no idea what is safe and what isnt, which is why 99% of overclockers disable all the auto controls and set everything to settings that are correct. computers have come along way, but they are far from smart.

It is also incorrect to believe that throttling is something any C2D OCer wants. Throttling C2D defeats the purpose of OC because it slows the performance of the CPU down do to the heat.

i agree, which is why everyone who knows what they are doing should disable EIST and C1e and all those settings in bios. then this wont happen.

 

[jzodda]

I am trying to get through a 8+ hour orthos blend test 500x7 with 1.465V in bios and 2.2V DDR2 @ 5-5-5-15 2T

Load temps are now 58C-61C

Its amazing how much higher the temp scales up with just a few notches of Vcore.

 

[ghost6303]

Its amazing how much higher the temp scales up with just a few notches of Vcore.

indeed. this is the reason that when you get to the top of your overclock, getting another 50-100mhz is not worth the 10 degree increase it takes to get there. theres usualy a clear point when it gets to be too hot. thats a pretty nice overclock im surprised that TEC is able to keep up.

 

[jzodda]

I am surprised myself. I had to re-seat it about 10 times before I got good contact. Pain in the ass, but the result is pretty good

 

[unclewebb]

Eastcoast: That 131.69% throttle number is from that RMClock program you recommended. Nice program but the way it calculates percentages is obviously not accurate on an overclocked system. How can you throttle a processor more than 100%?

My guess (in this situation) is that you were throttling well before you reached 83C.

Nice guess but it is totally wrong. In case you missed my second post, at 82C and less the C2D runs at full speed with no throttling. TAT isn't complaining and neither is RMClock. You need to forget about the old P4. The C2D is a different processor and its throttling characteristics are completely different.

The truth of the matter is your omitting overall case temps and how higher cpu temps effect your ram, chipset and video card (to name a few)

Get real. I had to practically turn my cpu fan off to get up over 80C. I certainly don't recommend anyone should run their computer with a core temp over 80C for 24/7 use. My post(s) were to simply debunk some of the myths, like the ones you're spreading Eastcoast.

From looking at this photo its very possible to start throttling at around 60C

That might be how a P4 works but no Core2Duo throttles at a core temperature of 60C. You really need to read up a little on the C2D before spreading misinformation like this around.

Nice overclock jzodda. Checking for stability with your memory at 2.2 volts and trying to lower your core voltage while maintaining stability shows that you know exactly what you're doing. Hopefully my posts helped convince you that you have plenty of head room left.

When you're done playing with it, remember to enter your name in the [H] oc database where it belongs.

 

[theseeker]

Eastcoast: That 131.69% throttle number is from that RMClock program you recommended. Nice program but the way it calculates percentages is obviously not accurate on an overclocked system. How can you throttle a processor more than 100%?


Nice guess but it is totally wrong. In case you missed my second post, at 82C and less the C2D runs at full speed with no throttling. TAT isn't complaining and neither is RMClock. You need to forget about the old P4. The C2D is a different processor and its throttling characteristics are completely different.



Get real. I had to practically turn my cpu fan off to get up over 80C. I certainly don't recommend anyone should run their computer with a core temp over 80C for 24/7 use. My post(s) were to simply debunk some of the myths, like the ones you're spreading Eastcoast.


That might be how a P4 works but no Core2Duo throttles at a core temperature of 60C. You really need to read up a little on the C2D before spreading misinformation like this around.

Nice overclock jzodda. Checking for stability with your memory at 2.2 volts and trying to lower your core voltage while maintaining stability shows that you know exactly what you're doing. Hopefully my posts helped convince you that you have plenty of head room left.

When you're done playing with it, remember to enter your name in the [H] oc database where it belongs.

This is a very interesting thread. I do not wish to highjack this thread but I have a question that I cannot get answered. I am running a QX6700 at 3.3~1.3V, FSB 1333~1.3V and mem at 1066~2.0V and my load temps are 47-49C. The NB never gets over 37C and my GPU's never exceed 52C under load. Sorry for being long-winded, but I would like to know what the max temp is for the QX? I have hit 3.6 and it was fully stable running Orthos and OCCT for 8 hours but my CPU temp hit 62C. Is this acceptable?

 

[ghost6303]

This is a very interesting thread. I do not wish to highjack this thread but I have a question that I cannot get answered. I am running a QX6700 at 3.3~1.3V, FSB 1333~1.3V and mem at 1066~2.0V and my load temps are 47-49C. The NB never gets over 37C and my GPU's never exceed 52C under load. Sorry for being long-winded, but I would like to know what the max temp is for the QX? I have hit 3.6 and it was fully stable running Orthos and OCCT for 8 hours but my CPU temp hit 62C. Is this acceptable?

quad cores have a 15 degree higher max of 100c. anything under 70-80 is supposed to be acceptable. 62 is fine.

 

[Eastcoasthandle]

Eastcoast: That 131.69% throttle number is from that RMClock program you recommended. Nice program but the way it calculates percentages is obviously not accurate on an overclocked system. How can you throttle a processor more than 100%?

Nice try but the truth of the matter is your don't know what you are doing telling people that temps that high are OK.

Nice guess but it is totally wrong. In case you missed my second post, at 82C and less the C2D runs at full speed with no throttling. TAT isn't complaining and neither is RMClock. You need to forget about the old P4. The C2D is a different processor and its throttling characteristics are completely different.

I guess that 83C is also wrong. Did you post that SS to discredit yourself? Or do you actually game at those temps?

Get real. I had to practically turn my cpu fan off to get up over 80C. I certainly don't recommend anyone should run their computer with a core temp over 80C for 24/7 use. My post(s) were to simply debunk some of the myths, like the ones you're spreading Eastcoast.

All that for a SS? So, you let your CPU run for a few seconds just to take a screen shot and hurry up and turn your fan back on. How noble of you to do so. Your previous post read differently but now you say this. Looks like the you need to get real!

That might be how a P4 works but no Core2Duo throttles at a core temperature of 60C. You really need to read up a little on the C2D before spreading misinformation like this around.

It is you who is posting misinformation (something I said before). Do you have any Intel specs that say that C2D don't throttle? In any case, as I said before, if your temps are that hot (at OP) you might want to consider water cooling.
Besides, I never said nor imply that C2D throttle at 60C. LOL

 

[ghost6303]

first of all- calm down.
secondly-

It is you who is posting misinformation (something I said before). You provide no information to back up what you say and you offer no source.

i was under the asumption it was common knowledge.
but here is some sources. most specificly this section...

The temp scale shown below illustrates the normal ~ 25c Delta between Idle and TAT @ 100% Load, and the typical ~ 15c Delta between Tcase and Tjunction on an example system overclocked with 1.4 Vcore. 50c Tcase and 65c Tjunction are safe and sustainable temperatures.

Tcase/Tjunction

--60--/--75--75-- Hot
--55--/--70--70-- Warm
--50--/--65--65-- N
--45--/--60--60-- O
--40--/--55--55-- R
--35--/--50--50-- M
--30--/--45--45-- A
--25--/--40--40-- L
--20--/--35--35-- Cool

(Tcase = ambient case temp, Tjunction = core temp 1-2)
as you can see, above 70 is getting warm, upper 70s and low 80s is hot, above 85 is dangerous, just because the program your using to throttle it is set at the wrong temps doesnt mean those temps are what it should be. RMclock could be programed to start throttling at 35c that doesnt make above 35 dangerous does it?

 

[jzodda]

This is an interesting discussion



I am back for some more screens. (500x7) 8 hours Orthos stable @1.465Vcore. Now I will work a little on timings and if I can go tighter that will be nice but if not I think I have found my sweet spot for now.

 

[Eastcoasthandle]

first of all- calm down.
secondly-

i was under the asumption it was common knowledge.
but here is some sources. most specificly this section...


(Tcase = ambient case temp, Tjunction = core temp 1-2)
as you can see, above 70 is getting warm, upper 70s and low 80s is hot, above 85 is dangerous, just because the program your using to throttle it is set at the wrong temps doesnt mean those temps are what it should be. RMclock could be programed to start throttling at 35c that doesnt make above 35 dangerous does it?

As I said before:

It's incorrect to believe that:
-If your C2D does throttle it suppose to throttle
-It's normal to have temps that high
-you are suppose to OC your CPU until the heat build up from Orthos shuts your Computer off (do to heat).
-want to operate a CPU at those high temps which also effect other components inside your case
If your temps are that high you might want to consider water cooling.
This is the point I am making here.

 

[ghost6303]

"It's incorrect to believe that:
-If your C2D does throttle it suppose to throttle"

you brought up throttling in the first place, i never use it as dont most people.

"It's incorrect to believe that:
-It's normal to have temps that high"

IT IS NORMAL TO HAVE TEMPS THAT HIGH

"-you are suppose to OC your CPU until the heat build up from Orthos shuts your Computer off (do to heat). "

you are the one who said that in the first place, this is definately not the way to overclock. your supposed to monitor temps as you overclock, once you get to computer shutting down you risk damage.

"-throttling is something any C2D OCer wants."

again, you brought up throttling

"-want to operate a CPU at those high temps which also effect other components inside your case"
refer back to answer #2 they are not high temps, also the CPU temp has very little to do with other component temp. granted- when you increase cpu voltage you generaly have to increase ram and northbridge voltage to match, thus increasing temps of the overall system. CPU temp has a minimal effect on surrounding hardware. about 20% of heat from the CPU is dissapated thru the motherboard, and unless you have zero airflow this isnt a problem as if your running your CPU at 80c (5c below the red line) 20% of 80c is only 16 degrees leaked into surrounding components. and sence most people use heatinks and fans, there is enough airflow to eleminate this temp increase.

edit
heres a screen of my second computer, e6300 at a mild 2.1ghz with the crappy shuttle heatsink/fan, which has been on for 4 days strait, fans on ultra low, 66c loaded with orthos
completely normal operating temperature

 

[unclewebb]

Eastcoast Throttling Myth Debunked:

Here is a cool Core2Duo running SuperPi:

Here is the same processor running SuperPi at 76C:

Can you see a big difference? Is the C2D running at part throttle at this temperature? Of course not.

The difference in performance between these two runs is the normal variation between two runs using SuperPi.

Sorry I couldn't get up into the 80C range to run this comparison but without TAT running in the background I couldn't get my C2D that hot.

The fact is that the Core2Duo does not start to throttle until it gets to within a couple of degrees of Tjunction or TjMax as Intel calls it.

TAT will clearly show you when thermal throttling is taking place because the part in green that normally says "Thermal Monitor Idle" will turn to red and change to "Thermal Monitor Active."

 

[unclewebb]

Anyone interested in learning a little more about the digital temperature sensors built into the Core2Duo can read Chapter 13 from the:

Intel® 64 and IA-32 Architectures Software Developer’s Manual
http://www.intel.com/design/processor/manuals/253668.pdf

 

[Slammin]

Anyone interested in learning a little more about the digital temperature sensors built into the Core2Duo can read Chapter 13 from the:

Intel® 64 and IA-32 Architectures Software Developer’s Manual
http://www.intel.com/design/processor/manuals/253668.pdf

Agreed, and they should also avoid this thread because Eastcoast is simply polutting it with too much disinformation.

 

[Eastcoasthandle]

"It's incorrect to believe that:
-If your C2D does throttle it suppose to throttle"

you brought up throttling in the first place, i never use it as dont most people.

"It's incorrect to believe that:
-It's normal to have temps that high"

IT IS NORMAL TO HAVE TEMPS THAT HIGH

"-you are suppose to OC your CPU until the heat build up from Orthos shuts your Computer off (do to heat). "

you are the one who said that in the first place, this is definately not the way to overclock. your supposed to monitor temps as you overclock, once you get to computer shutting down you risk damage.

"-throttling is something any C2D OCer wants."

again, you brought up throttling

"-want to operate a CPU at those high temps which also effect other components inside your case"
refer back to answer #2 they are not high temps, also the CPU temp has very little to do with other component temp. granted- when you increase cpu voltage you generaly have to increase ram and northbridge voltage to match, thus increasing temps of the overall system. CPU temp has a minimal effect on surrounding hardware. about 20% of heat from the CPU is dissapated thru the motherboard, and unless you have zero airflow this isnt a problem as if your running your CPU at 80c (5c below the red line) 20% of 80c is only 16 degrees leaked into surrounding components. and sence most people use heatinks and fans, there is enough airflow to eleminate this temp increase.

edit
heres a screen of my second computer, e6300 at a mild 2.1ghz with the crappy shuttle heatsink/fan, which has been on for 4 days strait, fans on ultra low, 66c loaded with orthos
completely normal operating temperature

Yes, I brought up throttling but I did so in a question. And, it's not relevant in the context in which you post about it. The point of this thread is to discuss his temps. IMO, I suggested that he water cool if he wants his CPU at this overclock. Therefore, your posts are getting way off subject as you only want to discuss throttling which is not the point of my original posts (or posts thereafter).
As I said before:

It's incorrect to believe that:
-If your C2D does throttle it suppose to throttle
-It's normal to have temps that high
-you are suppose to OC your CPU until the heat build up from Orthos shuts your Computer off (do to heat).
-want to operate a CPU at those high temps which also effect other components inside your case
If your temps are that high you might want to consider water cooling.
This is the point I am making here. IMO it is NOT normal to run temps that high.

 

[Eastcoasthandle]

Here is a cool Core2Duo running SuperPi:

Here is the same processor running SuperPi at 76C:

Can you see a big difference? Is the C2D running at part throttle at this temperature? Of course not.

The difference in performance between these two runs is the normal variation between two runs using SuperPi.

Sorry I couldn't get up into the 80C range to run this comparison but without TAT running in the background I couldn't get my C2D that hot.

What throttling myth? This is getting way off topic. My position is to use water cooling if your temps are that high (at least if your gaming temps are).
Oh so you don't normal run your CPU at 70C+/83C now? So all you do is take ss at 83C to try and make a false point of view. Let stay on topic here as the OP asked if his temps where safe. Also, when you ran Super PI what was your CPU clocked at?

As I said before:

It's incorrect to believe that:
-If your C2D does throttle it suppose to throttle
-It's normal to have temps that high
-you are suppose to OC your CPU until the heat build up from Orthos shuts your Computer off (do to heat).
-want to operate a CPU at those high temps which also effect other components inside your case
If your temps are that high you might want to consider water cooling.
This is the point I am making here so the only thing debunk is your attempt to re-invent what I said, deal with it. Oh, and those scores are not remotely similar. It is clear that your score decreased with higher temps. Regardless of the reason, the higher temps have effected your score. Also, you can't did do a run at 80C either. The absence of that speaks volumes to me. I guess it was good enough to take a quick ss at 83C and hurry up to turn your fan back on LOL. You are only debunking yourself.

Below are 2 runs of Super PI. I included TAT to monitor temps and to time stamp the runs. So that it is clear I simply didn't copy and paste twice.


Elapsed Time: 16.875
Temp: 31/29
Time: 12:36:08pm


Elapsed Time: 16.875
Temp: 32/28
Time: 12:38:25

As you can see I get the same results twice. It's clear that milliseconds do play a role. That's why we use SuperPI Mod 1.5 XS elapsed time runs.

Side note:
Below you will find Super PI (regular). As you can see it does not include milliseconds. Therefore, when using SuperPI Mod 1.5 XS it is clear that milliseconds do play a role in elapsed time runs.

Because of the difference between SuperPI and Super PI mode 1.5 XS it is clear that incremental, elapsed time runs do play a role with this program. However, it can be defined according to why the program was used to begin with. In this case it was run to show no tangible difference between 46c/42c and 76C/71C. In this case that turn out to be FALSE because there is a difference between your elapsed time runs. However for me, they are exactly the same (time runs) using similar temps. I cannot say why for certain but it's clear that your time runs are not the same but they should be (in order to prove your point that high temps play no tangible role in performance). In all, it's clear that temps do affect your performance and thus the reason why I suggest you water cool. Again, this is the point I am making here.

 

[unclewebb]

What throttling myth?

Eastcoast, weren't you the guy on the previous page that posted this nonsense?

...its very possible to start throttling at around 60C. My guess (in this situation) is that you were throttling well before you reached 83C.

You come to this forum and deliberately mislead people with nonsense. The Core2Duo does not throttle itself at any temperature near 60C,

I come to this forum and put my processor on the line for the benefit of all [H]ard users and prove that the Core2Duo does not throttle until its core temperature exceeds 82C and then you reply that you don't know what I'm talking about.

You can go back and edit your previous comments all you want but you were the one that brought the big, bad, throttling bogeyman up for discussion.

jzodda originally asked if his core temps of 64C to 68C while running Orthos is safe and your first reply is to encourage him to get a water cooling system. You back that horrible advice up with stories about the Pentium 4 era which is completely irrelevant.

My position is to use water cooling if your temps are that high (at least if your gaming temps are).

More great advice. jzodda quoted his temps while running Orthos which is working both cores to the max. More than 99% of the games on the market today are single threaded and only utilize one core at a time. Gaming core temps are typically 10C to 15C less than Orthos temps.

SuperPI:
Anyone who has ever run this program more than once or twice has quickly learned that a small difference in time from one run to the next is NORMAL.

SuperPI mod 1,5 may use 3 digits after the decimal point but that does not prove that it is accurate to within 0.001 seconds. It is not and your results prove that eastcoast.

The two runs you posted show an initial time, 19 intermediate times and a final time. Your runs that "prove" how consistent your cool processor runs from one run of SuperPi to the next is about as misleading as it gets. Of the 21 times posted for your two runs of SuperPi there was a difference of:
0.032 secs 6 times
0.031 secs 2 times
0.016 secs 10 times
0.000 secs 3 times

It's nice your final scores were dead even at 16.875 seconds but all through your run there were differences and the average difference was greater than 0.016 seconds. Your data proves that a difference of 0.016 seconds from one SuperPi run to the next is completely NORMAL which is exactly the difference I encountered when running at 46C or 76C.

If you ever complete an introductory course in Statistics at the university level you will learn about what "statistically insignificant" means.

Another thing they will teach you in school is that if you look at the data you collected from SuperPi you will see that 16 of the 21 data points have a difference which is a multiple of 0.016. This makes it far more likely that the resolution of the timer that SuperPi uses is accurate to only 0.016 seconds and not 0.001 seconds.

Oh, and those scores are not remotely similar. It is clear that your score decreased with higher temps. Regardless of the reason, the higher temps have effected your score.

You previously confirmed your lack of knowledge about the Core2Duo and what is normal when it's overclocked but statements like that only help confirm your lack of knowledge about mathematics as well.

Also, you can't did do a run at 80C either. The absence of that speaks volumes to me.

I thought most normal users would be happy that I did a SuperPi run at 76C without any problems or statistically significant change in performance. SuperPi runs fine at over 80C but because it is only a single threaded app it is very difficult to run the program at an exact temperature and end up with any statistical meaning in the final results.

I apologize to any user, besides Eastcoast, that feels I have badly misled them in any of my posts and for Eastcoast I'll try to get my hands on a space heater to get my C2D up to a nice consistent 80C for a SuperPi run to prove, ummm, prove something I guess!

 

[jbmx4life]

Your temps are fine. Unless you got some serious multi-threaded stuff going 24/7 that has the cpu fully loaded constantly, then there won't be anything else that puts the load of Orthos on the cpu. As unclewebb stated, games are mostly single threaded so you won't get near that temp while gaming.

Eastcoast you need to stop comparing apples to oranges here. Everyone has told you that P4 is totally different than a C2D and you need to realize that already. Unclewebb's testing is truly amazing and fully legitimate and all you can do is try to knock when you ain't got shit to back up what your spewing except for piping about your P4 based CPU. No comparison and you are wrong.

 

[Arcygenical]

Eastcoasthandle.

At 65w TDP, the maximum case temperature (Tc) for a C2D processor is 60.1c for a 24/7 operation.

The THERMTRIP signal is sent if the temperature rises to 25c higher than this value, a temperature of 85c. The THERMTRIP is the total shutdown of the CPU.

Throttling occurs when the DTS report rises to 20c over the Tc value (60.1c) by means of voltage and frequency modulation.

This is all in their documents.

As for the variance within unclewebb's SuperPi shots... Well, Coming from someone who has done over 3000 runs in superPi, I can tell you that 100ms is easily lost from accidently touching your mouse a few times... Or by IntelTAT reading an extra temperature or two from the DTS. The variance is WELL within the margin of error, and I've experienced this myself while running nowhere near this DOCUMENTED 85c number.

In all, it's clear that temps do affect your performance

Uhh, no... I'm sorry. It's not.

 

[Eastcoasthandle]

Eastcoast, weren't you the guy on the previous page that posted this nonsense?

My suggestion has always been to water cool.

You come to this forum and deliberately mislead people with nonsense. The Core2Duo does not throttle itself at any temperature near 60C,

You are still talking about something other then the point of view I made suggesting to water cool.

I come to this forum and put my processor on the line for the benefit of all [H]ard users and prove that the Core2Duo does not throttle until its core temperature exceeds 82C and then you reply that you don't know what I'm talking about.

Again, not the point of my posts.

You can go back and edit your previous comments all you want but you were the one that brought the big, bad, throttling bogeyman up for discussion.

No, I only asked the question about it. You have taken this way out of context. Again, my suggestion is about water cooling the CPU.

jzodda originally asked if his core temps of 64C to 68C while running Orthos is safe and your first reply is to encourage him to get a water cooling system. You back that horrible advice up with stories about the Pentium 4 era which is completely irrelevant.

No, that's incorrect again, I suggest to water cool. I only related the heat of the P4 in relations to other components in the case. Which you don't believe exist.

More great advice. jzodda quoted his temps while running Orthos which is working both cores to the max. More than 99% of the games on the market today are single threaded and only utilize one core at a time. Gaming core temps are typically 10C to 15C less than Orthos temps.

My suggestion to water cool was intended for when ambient is higher then normal. I didn't limit my reasoning to Orthos only.

SuperPI:
Anyone who has ever run this program more than once or twice has quickly learned that a small difference in time from one run to the next is NORMAL.

Well that doesn't explain why my scores are not different. What is normal is the difference between each time run (as SuperPI is running). That's normal.

SuperPI mod 1,5 may use 3 digits after the decimal point but that does not prove that it is accurate to within 0.001 seconds. It is not and your results prove that eastcoast.

That is not the point of why I posted that. You suggest that having your CPU that hot doesn't matter. I showed otherwise. This is not an issue about the accuracy of timed run itself. But to show if temps affect performance.

The two runs you posted show an initial time, 19 intermediate times and a final time. Your runs that "prove" how consistent your cool processor runs from one run of SuperPi to the next is about as misleading as it gets. Of the 21 times posted for your two runs of SuperPi there was a difference of:
0.032 secs 6 times
0.031 secs 2 times
0.016 secs 10 times
0.000 secs 3 times

It's nice your final scores were dead even at 16.875 seconds but all through your run there were differences and the average difference was greater than 0.016 seconds. Your data proves that a difference of 0.016 seconds from one SuperPi run to the next is completely NORMAL which is exactly the difference I encountered when running at 46C or 76C.

The facts speak for itself. I am not hear to argue but to make it clear that those pics do show a difference between runs when temps are not the same.

I thought most normal users would be happy that I did a SuperPi run at 76C without any problems or statistically significant change in performance. SuperPi runs fine at over 80C but because it is only a single threaded app it is very difficult to run the program at an exact temperature and end up with any statistical meaning in the final results.

In earnest the issue is your deviation from the OP. I have said more then once my intentions. Be that as it may your runs did show a change in performance. When you are corrected you throw a fit about it.

I apologize to any user, besides Eastcoast, that feels I have badly misled them in any of my posts and for Eastcoast I'll try to get my hands on a space heater to get my C2D up to a nice consistent 80C for a SuperPi run to prove, ummm, prove something I guess!

Grabing for straws now? There is no need to be this upset over what's nothing. The truth of the matter is simple, water cool your CPU.

 

[jbmx4life]

There is no need for the OP to go to water. His temps are fine and well within the max temp documented by Intel.

 

[burningrave101]

Grabing for straws now? There is no need to be this upset over what's nothing. The truth of the matter is simple, water cool your CPU.

Watercooling is more expensive, a pain to setup if you don't have the perfect case for it, and it has to be maintained. You have to change out the water every so often and always keep an eye out for leaks. His temps aren't that high. I have about the best heatsink on the market, the Thermalright Ultra 120 Extreme and my E6600 @3.6Ghz+ is sitting around 59C at the moment under full load of Orthos Small FFTs. My temps will probably drop some more once I get the top case fans installed in my TJ09 but I have the side panel off right now so that helps too. The Core 2 Duo processors seem to run hotter than Pentium 4 processors because the thermal diode isn't in the same place any more. The thermal readings are coming from a much hotter point of the processor that's actually closer to the real temp of the hottest portion of the core. The real temps of your Pentium 4 were even higher than what you perceived them to be because of the location of the temp readings.

 

[ghost6303]

im not sure whats being argued here. all these points are facts not theories.
this is the question in the first post...

orthos blend is stable 12+ hours and running as I type this but my load temps are moving between 64-68c during the test. I realize that at high FSB it runs a little hotter but is this a safe temp for 24/7 operation?

the answer is not an opinion, its factual information based on specs from intel. these temps are within the recomended operating temperature for this specific CPU.

the first mention of throttleing was by you in post #4

I had a Intel 660 and with just air cooling, oc to 4.2 it would get a toasty 57C an would start throttling around 60C (if memory server correctly).

Prescott and conroe cores are as different as apples and walnuts. they are both grow on trees, but thats about it with similarities. the temperature limits on each family are completely different, as is archetecture, egineering processes, voltage requirements, speed, memory, socket type, etc.

No, I only asked the question about it. [throttling]

where was the question in there?

I only related the heat of the P4 in relations to other components in the case. Which you don't believe exist.

no you compared his CPU temps to yours, and i just explained why you cant do that.

Oh so you don't normal run your CPU at 70C+/83C now? So all you do is take ss at 83C to try and make a false point of view. Let stay on topic here as the OP asked if his temps where safe.

just because he has a better cooling soloution and doesnt normaly run as hot as the OP doesnt mean its not valid to increase the temp to run tests. and yes lets stay on the topic of the origional question, to which the answer is yes, his temps are fine according to Intel.

 

[unclewebb]

As for the variance within unclewebb's SuperPi shots... Well, Coming from someone who has done over 3000 runs in superPi, I can tell you that 100ms is easily lost from accidently touching your mouse a few times....

OH sh*t, busted. I touched the mouse!

Trying to get my processor up into the mid 70C range while simultaneously running SuperPi at full speed was not easy. I turned the fan down to a crawl, warmed the cores with TAT, shut that down and then ran Orthos to maintain the temperature.

If you go from this kind of heat and load to no load you'll cool off by 20C in the blink of an eye. The only thing I could do was to start SuperPi and immediately kill Orthos. The moving mouse killed my hot run time by 0.016 seconds!

Back to the original question. There is absolutely nothing wrong with a Core2Duo running Orthos in the mid to high 60C range and you certainly don't have to run out and get a water cooling kit.

Stability is the key. If you are Orthos stable for several hours and your temps are within the Intel spec and your processor is not throttling as displayed by TAT, then there's no need to worry about your temps because some guy on a forum claims that it's a big number.

The Core 2 Duo is a fantastic processor with plenty of safe guards built into it. Don't be afraid to use it to its fullest potential, the way Intel intended.

 

[SuperKeijo]

I turned off case fan and cpu fan and let the cpu burn tat @ 1.5v 3200mhz and temps went up to 98 according to tat and all was fine up to that point, then at some point it just rebooted LOL, this was with a e4300...

Besides C2D's (except quad cores and ex6800s..) don't cost shit anymore and will propably be almost free second hand in a year or two so I guess the lifespan won't be much an issue on these things anyway.