New i5-3570K cpu overheating, would love some suggestions??

[cadaveca]

#1. Realtemp is not accurate for IVB at this time. It's not that far off, but it is a bit high at times, that i have seen.

#2. You are on a stock cooler, and the max acceptable temperature on a stock cooler is 105 C. That may seem high, but really, it is not. IF it passes 105 C, it should throttle. If it is not throttling, then the method for measuring temps is definitely not accurate. IF it IS throttling, then you should RMA the chip.

 

[CaptainFailcon]

I think the heat-pipes in the cooler are defective rma the cooler

 

[Tardan]

What are your voltages going up to when running Prime95? I got my hands on a 3770k recently and had heat issues too until I noticed it was shooting up to 1.4v under load. I manually set it to 1.25ish and my temperatures improved drastically and was still stable after 2 hours of running Prime95.

Might want to check your voltage under load and see if its doing the same thing. If so I would manually set it to something lower.

 

[solara2xb]

Thanks man. Its about 23c in my apartment, cant open the window much as im on the ground floor :/ I thought my cooler was a really good one, be-quiet shadow rock pro?? Have tried new thermal past and re mounting and not much change, water cooling that much better?

Water cooling has its advantages, But there is some really good Air Coolers out there that can perform just as good... I've never used the cooler you have so I can't say if its good or not... I have the Havik 140 on one of my system and its a great cooler. I have a feeling something else is not right here... at 23c ambient you should be getting better temps even with IvyB.

Are you using the Pea method I mention to apply your Thermal Paste? Don't spread it yourself. As others have mention, it can create air pockets which will give you bad temps.

Just a small pea size in the middle of the CPU and let the Cooler do the spread for you. Once you install the cooler the compression of the cooler will spread the paste evenly. I've use this method for all my installs and never have I encountered any issues.

Also how is the airflow in your case?? and pics of your set-up?

#1. Realtemp is not accurate for IVB at this time. It's not that far off, but it is a bit high at times, that i have seen.

#2. You are on a stock cooler, and the max acceptable temperature on a stock cooler is 105 C. That may seem high, but really, it is not. IF it passes 105 C, it should throttle. If it is not throttling, then the method for measuring temps is definitely not accurate. IF it IS throttling, then you should RMA the chip.

Yea, I wouldn't use real temp. Its great to have for general purposes, but if you want a more accurate temp readying use Core Temp.

 

[mrmotivator]

What are your voltages going up to when running Prime95? I got my hands on a 3770k recently and had heat issues too until I noticed it was shooting up to 1.4v under load. I manually set it to 1.25ish and my temperatures improved drastically and was still stable after 2 hours of running Prime95.

Might want to check your voltage under load and see if its doing the same thing. If so I would manually set it to something lower.

Vcore voltage seems to be about 1.146v, running prime95 it didnt go above that at all.

I did not use the pea method with the thermal paste that came with the cooler (I was following instructions on a newegg video which said to spread it all over).

I will use that method when the MX-4 arrives and hopefully that will improve the situation!

I think the airflow inside the case is fine, it has 3 big fans inside it - although i can post a pic if you like

Thanks for all the responses guys just hoping i won't have to RMA the cpu!

 

[solara2xb]

Wait! So when you reseated the cooler you didn't reapply new thermal paste??? If that is the case you might have a problem there.. Everytime you remove the cooler its rule of thumb to apply new thermal paste. Clean off the old stuff and apply new TIM. This way you get a even spread when reapplied...

 

[Aquinus]

Wait! So when you reseated the cooler you didn't reapply new thermal paste??? If that is the case you might have a problem there.. Everytime you remove the cooler its rule of thumb to apply new thermal paste. Clean off the old stuff and apply new TIM. This way you get a even spread when reapplied...

He said he didn't use the pea method and that he spread it out himself. I bet you he used too much. Plus, it's the stock cooler, right? It's not going to be amazing anyways.

 

[solara2xb]

He's using a aftermarket cooler.. Not the stock...

 

[Aquinus]

He's using a aftermarket cooler.. Not the stock...

He doesn't have it yet, judging from he said he still is using the stock cooler and the after market one is on its way.

 

[solara2xb]

I have tried re mounting it :/ No its a Be Quiet Dark Rock PRO BK016, thought it was a good one?!

He doesn't have it yet, judging from he said he still is using the stock cooler and the after market one is on its way.

He's using the cooler listed above. he's is waiting on new thermal paste.. I think when he remounted the cooler he didn't use new thermal paste, but rather just stuck the cooler back on the CPU with whatever thermal past that is already on the cooler..

Which I think is still the problem..

That is of course I understand what he did correctly.

 

[MxPhenom 216]

Water cooling has its advantages, But there is some really good Air Coolers out there that can perform just as good... I've never used the cooler you have so I can't say if its good or not... I have the Havik 140 on one of my system and its a great cooler. I have a feeling something else is not right here... at 23c ambient you should be getting better temps even with IvyB.

Are you using the Pea method I mention to apply your Thermal Paste? Don't spread it yourself. As others have mention, it can create air pockets which will give you bad temps.

Just a small pea size in the middle of the CPU and let the Cooler do the spread for you. Once you install the cooler the compression of the cooler will spread the paste evenly. I've use this method for all my installs and never have I encountered any issues.

Also how is the airflow in your case?? and pics of your set-up?



Yea, I wouldn't use real temp. Its great to have for general purposes, but if you want a more accurate temp readying use Core Temp.

If Real Temp is only supported for Sandy Bridge at this point, and Core Temp isn't accurate and updated for Sandy Bridge, then how do you expect Core Temp to work well with Ivy Bridge?

 

[solara2xb]

Where did you find that Core Temp isn't accurate for Sandy?? I've used Coretemp on allot of systems and never had any issues. I also use CPUz hardware Monitor and the temps recorded are pretty close...

 

[cadaveca]

Where did you find that Core Temp isn't accurate for Sandy?? I've used Coretemp on allot of systems and never had any issues. I also use CPUz hardware Monitor and the temps recorded are pretty close...

This fact was reported by ASUS(That it is wrong for IVB), and then confirmed by UncleWebb here on TPU, who writes Coretemp. He doesn't have a chip yet(or didn't), and has yet to update it. I'm sure he will when he has a chip.


I noticed myself when testing IVB OC stability that AIDA64, realtemp, and Coretemp all reported different temperatures. According to ASUS, AIDA64 is the one that is right currently(and is also recommended for stability testing).

SB is apparantly working just fine. IVB is different.

 

[solara2xb]

Thanks for the clarification. Didn't know CoreTemp and Ivy had issues. Good to know..

 

[cadaveca]

Thanks for the clarification. Didn't know CoreTemp and Ivy had issues. Good to know..

It's actually not that far off. But it does read higher than it should at times. I have no idea what the issue is, or why, or how...

I think both Realtemp and CoreTemp should be updated soon. Heck, I get those two confused all the time, but both are inaccurate with current versions(unless one came out in the last week).

 

[unclewebb]

RealTemp 3.70
http://www.techpowerup.com/downloads/2089/Real_Temp_3.70.html

The Asus rep did not know what he was talking about.

Core Temp and RealTemp both support Ivy Bridge and both report the core temperature correctly. If there is a problem, it is with the sensors that Intel uses that are not, and never have been, 100% accurate temperature monitoring devices. Intel only designs these sensors for thermal throttling and thermal shutdown control.

The Asus rep was trying to come up with a reason to explain why the Asus monitoring app reads so low. The reason is that the Asus app does not report core temperatures. It reports the CPU temperature from a totally different sensor which is always much cooler than the peak core temperature at full load.

confirmed by UncleWebb here on TPU, who writes Coretemp.

Check out the avatar title.

Hope that clears up the confusion.

 

[solara2xb]

Thanks!

 

[cadaveca]

RealTemp 3.70
http://www.techpowerup.com/downloads/2089/Real_Temp_3.70.html

The Asus rep did not know what he was talking about.

Core Temp and RealTemp both support Ivy Bridge and both report the core temperature correctly. If there is a problem, it is with the sensors that Intel uses that are not, and never have been, 100% accurate temperature monitoring devices. Intel only designs these sensors for thermal throttling and thermal shutdown control.

The Asus rep was trying to come up with a reason to explain why the Asus monitoring app reads so low. The reason is that the Asus app does not report core temperatures. It reports the CPU temperature from a totally different sensor which is always much cooler than the peak core temperature at full load.



Check out the avatar title.

Hope that clears up the confusion.

Yeha, i got you guys confused.


However:

View attachment 47176

Is that Coretemp or Realtemp? I cannot remember. My screenshot, BTW.

ANd I was referring to THIS post of yours:

http://www.techpowerup.com/forums/newreply.php?do=newreply&p=2601104

I have read some forums where all sorts of conclusions are being drawn on how hot Ivy Bridge runs. That's nonsense. These conclusions are based on temperature data that can not be relied on for any of Intel's recent CPUs. Temperature data from Intel CPUs should not be compared to other CPUs in the same CPU family let alone comparisons to CPUs from different families.

RealTemp is a good program and the sensors are useful enough so that if you swap a heatsink you can see if things are better or worse but beyond that, Intel's core temperature sensors are not accurate enough or documented well enough to provide users with 100% accurate core temperatures.

I meant nothing other than what you said, that it was inaccurate.

Your software, adn others, give different temps, so some explanation is needed...I don't really care who is wrong, who is right, or whatever, I expect you guys to sort it out yourselves.

 

[brandonwh64]

I have seen on some heatsinks with the DHPT needing a small layer in the heatsink itself besides the CPU to fill in the cracks between pipes. I would do a thin spread layer on the CPU and very thin later on the heatsink.

 

[unclewebb]

I noticed myself when testing IVB OC stability that AIDA64, realtemp, and Coretemp all reported different temperatures.

Another question is I noticed that real usually records lower temps then core temp. Any reason for that?

All monitoring programs that report core temperatures for Intel Core i CPUs are reading the same temperature data from the same register within the CPU. They also should be reading the exact same TJ Max value from another register in the CPU and they are all using the following formula to determine the core temperature.

Core Temperature = TJ Max Value - Digital Thermal Sensor Reading

Intel standardized where these two registers are located in the CPU and how to get core temperature data from them back in November 2008 when the first Core i CPUs were released. Nothing has changed since then. Ivy Bridge, Sandy Bridge and all the rest of them are identical when it comes to reading core temperature data.

When lightly loaded, core temperatures are constantly changing. It is impossible for 2 different programs to sample the temperature sensors at the exact same moment in time so slight differences should be expected. When the CPU is fully loaded and the core temperature has had a chance to stabilize, the amount of difference between various programs should be minimal.

The advantage RealTemp has is that it works at a higher Windows priority level than the competition does. When a CPU is fully loaded running LinX, RealTemp is at the front of the line when it comes to asking the CPU for some data from the core temperature sensors. This allows it to get temperature data more frequently than some other monitoring programs. That's the reason why you might see it reporting higher core temperatures than the competition.

If you want to use a utility that can sample the CPU more frequently as well as at a higher priority level than any other program, check out ThrottleStop.

ThrottleStop 5.00 beta 1
http://www.mediafire.com/?wqo7qmmimh31w92

According to ASUS, AIDA64 is the one that is right currently

The Asus rep doesn't even understand the difference between core temperatures and CPU temperatures. I wouldn't put too much faith in his choice of monitoring software.

Intel's core temperature sensors are not accurate enough or documented well enough to provide users with 100% accurate core temperatures.

That statement is still 100% true. These sensors are still not fully documented and the amount of error is still unknown. Intel saves money by not using sensors that are 100% accurate from idle to full load. RealTemp is quite capable of reading Ivy Bridge temperature sensors correctly. If the reported core temperatures are not 100% accurate, you need to ask Intel to use some higher quality sensors.

 

[cadaveca]

When lightly loaded, core temperatures are constantly changing. It is impossible for 2 different programs to sample the temperature sensors at the exact same moment in time so slight differences should be expected. When the CPU is fully loaded and the core temperature has had a chance to stabilize, the amount of difference between various programs should be minimal.

The advantage RealTemp has is that it works at a higher Windows priority level than the competition does. When a CPU is fully loaded running LinX, RealTemp is at the front of the line when it comes to asking the CPU for some data from the core temperature sensors. This allows it to get temperature data more frequently than some other monitoring programs. That's the reason why you might see it reporting higher core temperatures than the competition.

Great explanation.

THat screenshot i posted is from na ASUS Z77 product, which obviously shows the much lower CPU temp. I've seen other boards that use the highest "CPU CORE" temp as the reported temp there...so while Intel may have standardized teh data, OEMs aren't always using said standard method.

Frankly, myself, I don't care what temps are unless the CPU throttles.

 

[D007]

I just the paste it came with, was hard to spear over the CPU as was so thick but did the best i could?! I changed the settings in the bios to every option and no improvement, should be running all the time.

Oh no u din't!!
Jebuss..lol.. Smear you say?...
Apply a small beebee sized drop to the cpu and place the cooler squarely on top of it.
Do not slip or slide it, place it firmly and straighlt down to not create any bubbles.. Firmly tighten all of the coolers corners, going from cross end to cross end. Do not go from corner to corner in a clockwise/counterclockwise fashion. It's like putting on a tire.. Tighten the screws diagonal across from eachother, then work to the others in the same manner. You never want to tighten one more than the others to much, let them all tighten at the same pace.

Get Arctic silver or something for paste, the stuff that comes with it is garbage..
I like ac5 still.
Anything new come out that's better?

 

[solara2xb]

Oh no u din't!!
Jebuss..lol.. Smear you say?...
Apply a small beebee sized drop to the cpu and place the cooler squarely on top of it.
Do not slip or slide it, place it firmly and straighlt down to not create any bubbles.. Firmly tighten all of the coolers corners, going from cross end to cross end. Do not go from corner to corner in a clockwise/counterclockwise fashion. It's like putting on a tire.. Tighten the screws diagonal across from eachother, then work to the others in the same manner. You never want to tighten one more than the others to much, let them all tighten at the same pace.

Get Arctic silver or something for paste, the stuff that comes with it is garbage..
I like ac5 still.
Anything new come out that's better?

Try IC Diamond, Been using this stuff and its been great. I use to use AS5 and switched from that to IC Diamond. Best stuff I have used in years.