• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Something wrong with haswell cpus? High temps.....

kaoru.

Junior Member
Hey guys.

I'm having problems as well with haswell. i7-4970k 4.0
Not OC. Not planning.
http://imgur.com/KxRsOto

I'm getting 90's +c if I stress it with encoding/prime.

I reapplied the artic cooler heatsink(which is huge).
Like 3 times already. And STILL SAME EXact TEMPS!!! .
I was going to test it out with original stock cooler
since thats what my other i5 chip is using.

WHile I set in my old i5 cpu into another mobo, I get
GOOD nice low temps. Stock cooler. Max 65c.

Is there something wrong with haswell or MOBO not
reporting TEMPS correctly??????
While MOBO temps are around 30s???? Even
in my MPC PC temps jump into the 40-50s.

Driving me crazy.

Also I was feeling out the heatsink, was WARM AS WELL!
My 780TI gets realllly hot in the 80s+ normal.
100+f here though.

Know whats funny if a CPU is overheating that high? Would
it not blackscreen, shut off due to temps????
Idles at 40s, seems normal cause of heaT WAVE here.

Was reading in other thread.
 
CPUs throttle, nothing bad happens.

However it does sound like its not correctly seated.

What temperatures are you getting with the stock cooler?
 
CPUs throttle, nothing bad happens.

However it does sound like its not correctly seated.

What temperatures are you getting with the stock cooler?

I seated it prety good actually, 4thtime
already. Tried it on other
cpu with same aftermarket cooler, good temps.

Have not installed stock cooler yet. Maybe going to hold
off and buy different thermal pastes see if they work or not.

Was wondering if anyone else has high temps with same
i7-4970k.
 
Wow this is ultra weird. 100C is not normal on non-OC Haswell at all. Especially with a custom cooler. Perhaps there is some kind of a mistake you are doing when mounting the cooler? Perhaps the fan settings are off somehow in bios? For how long did you have this CPU? Could be a flaw of the cpu/motherboard/cooler as well.
 
srly dude .. u gettin paid? (not OP) .. anyway haswell is running hot, haswell's with stock cooling will throttle under AVX loads. What you are seeing is nothing new 🙂
 
Yeah, Haswells throttle under normal condition at stock, with AVX loads. Nothing new there. Hopefully, with 14nm, they might get the heat under control.
 
This is new, I put it together on Saturday. Gigabyte gaming 5 + 8 GB ram+
Cpu. These temps were like that.

Shouldnt my PC like Turn off? Or black/blue screen of death?
if its over 100? Ran prime like for 5 mins.
Havnt had this happen in any of my builds before.
Still running fine. errmmmm. Had nice hopes for this one.....
Only the USB3 on this board made me happy! Considerign I was stuck
with USB 2.

Biggest time waster of a build I've ever had 🙁 > guess ill wait for new thermal pastes.
See how it goes.
 
Last edited:
Shouldnt my PC like Turn off? Or black/blue screen of death?
if its over 100?

No. As already said, it throttles if it hits tjmax. And no, it doesnt crash or anything. You could actually run it without heatsink at all. It would just run very very slow.
 
no multi core turbo going on here if thats what you're getting at.
But even if, with all the reports combined, what is the story?
What does the manual say? You can run this processor on iceland, cause well, its cold, should you so happen to live in africa you should expect it to throttle?
Nice try, come again.
 
no multi core turbo going on here if thats what you're getting at.
But even if, with all the reports combined, what is the story?
What does the manual say? You can run this processor on iceland, cause well, its cold, should you so happen to live in africa you should expect it to throttle?
Nice try, come again.

The question you should ask is, does it throttle below the baseclock? Or is it simply throttle of the turbo bins.
 
Weren't Gigabyte z97 boards raising vcore to ridiculous levels and there was a new bios released to fix it?
 
I was looking through BIOs and Protection.
Cpu vrin pwm thermal protection set to 130.C? isnt that a bit high?
 
AVX loads aside, Haswell seems particularly sensitive to voltage, and automatic settings seem sometimes geared more to stability than manageable temperatures.
 
Yeah, Haswells throttle under normal condition at stock, with AVX loads. Nothing new there. Hopefully, with 14nm, they might get the heat under control.

Shrinks while giving better power efficiency mean higher power densities which makes thermals worse. Probably see a reduction in core TDP to compensate.
 
Haswell's have a "feature" (the new IVR's) that arbitrarily auto-overvolts themselves by +0.1v or so when detecting AVX instructions (like Prime) which can drive up temps by 20-30c or so vs Sandy / Ivy Bridge's. Setting a fixed voltage (instead of offset / adaptive) usually overrides this, and some motherboards are smart enough to apply the fixed only to load (ie, setting fixed won't break Speedstep idle state).

And if you're not planning to OC, maybe try undervolting it instead? You could knock off up to 10-30w / 5-20c or so depending on how much overhead your CPU has. At 4GHz, I can't even get my i5-3570 past 55c on a cheap 212 EVO, yet I've seen 90c Haswell's at stock 3.4GHz before in AVX benchmarks due to the dumb +0.1v self-auto-overvolt thing. It's that, that causes most heat problems when benchmarking.
 
Only see ~0.02V difference in core voltage from idle @3.4 on 4700MQ and running Linpack AVX2 at 170+ GFlops. However this laptop BIOS does not support voltage adjustments except for iGD which is incorrect anyway, so have had to add them myself to use interpolative operation for the limited unlocked multi's.
 
I was looking through BIOs and Protection.
Cpu vrin pwm thermal protection set to 130.C? isnt that a bit high?

No thats normal. The CPU will throttle at 100C and it will shutdown at 130C, should it ever reach that. Its unlikely you ever gonna be able to go over 100C due to throttle.
 
Haswell's have a "feature" (the new IVR's) that arbitrarily auto-overvolts themselves by +0.1v or so when detecting AVX instructions.

The 0.1V increase got nothing to do with the FIVR. When 256bit AVX loads are used, the cache also runs at 256bit. And the voltage is raised 100mv to ensure stability.
 
Thanks guys for the help.

I switched to original stock cooler, I get the same temps 🙁
Only difference is a few degrees at idle 42c compared to 39-41
aftermarket. Later I'll try the new thermal pastes, and compare.
 
The question you should ask is, does it throttle below the baseclock? Or is it simply throttle of the turbo bins.

When I get home Ill disable MCT alltogether and load up 8 threads of prime again.. wanna take a wager on the outcome? Last time i did this test it was snowing outside, about 6 months ago.
 
Sorry but that's not normal at all...not even with AVX Prime and at stock. Yes I do have a waterloop, so get better temperatures, but it shouldn't differ much at stock settings.

I say there is some high overvaulting happening with your UEFI.
 
Back
Top