i7-5820K Overclocking and temp weirdness

[zafros]

Hi all - first post here

 

I recently converted from that other religion to the whole windows ordeal... anyway bought myself a spanking new PC, in parts.

 

This has unfortunately yielded some interesting results. My i7-5820K overclocked nicely to 4,4 Ghz on the first day of owning it, and bumped it up to 4,5 Ghz the next day. This remained solidly stable for about two weeks, in which time I also overclocked the graphics card. I was hovering around 60 degrees centigrade during full load at 1,3 volts, which seemed very good to me (what do I know, I haven't had a PC in a decade), consuming about 145 Watts.

 

Well today that all changed, it was unstable as F, and in my mind i went WTH WTF just happened?! I basically had to reset it all and start over, which yielded a result of 4,1 Ghz at 1,25 Volts, consuming 180 Watts with temps never dropping below 75 degrees centigrade, and peaking at 80. So while these temps are still safe in Intel language, I'm loosing my mind as to what spurred this sudden change...

 

oh forgot to mention, my cooler is a CM Nepton 240M

 

So I don't really know where to start, i'm hoping some of you guys will have a clue

 

Have a good one

- zafros

 

 

 

 

Full specs below:

 

MB: Asus X99-a

CPU: i7-5820K

GPU: Asus GTX 970 Strix

RAM: Corsair Vengeance LPX DDR4 2133Mhz 8GB x4

PSU: Corsair RM750

Case: Fractal Design Define R5

[Imakuni]

Perhaps you've changed the fan / pump profile on accident?

 

Tryed cleaning it?

 

Maybe there's something installed that's stressing the CPU (say, a program, vírus, etc), thus causing temps to raise.

 

It's even possible that your ambient temps got higher, is it hotter these last few days?

 

You could.. you know, reapply the cooler, maybe it went lose?

[ChrisCross]

Hi all - first post here

 

I recently converted from that other religion to the whole windows ordeal... anyway bought myself a spanking new PC, in parts.

 

This has unfortunately yielded some interesting results. My i7-5820K overclocked nicely to 4,4 Ghz on the first day of owning it, and bumped it up to 4,5 Ghz the next day. This remained solidly stable for about two weeks, in which time I also overclocked the graphics card. I was hovering around 60 degrees centigrade during full load at 1,3 volts, which seemed very good to me (what do I know, I haven't had a PC in a decade), consuming about 145 Watts.

 

Well today that all changed, it was unstable as F, and in my mind i went WTH WTF just happened?! I basically had to reset it all and start over, which yielded a result of 4,1 Ghz at 1,25 Volts, consuming 180 Watts with temps never dropping below 75 degrees centigrade, and peaking at 80. So while these temps are still safe in Intel language, I'm loosing my mind as to what spurred this sudden change...

 

oh forgot to mention, my cooler is a CM Nepton 240M

 

So I don't really know where to start, i'm hoping some of you guys will have a clue

 

Have a good one

- zafros

 

 

 

 

Full specs below:

 

MB: Asus X99-a

CPU: i7-5820K

GPU: Asus GTX 970 Strix

RAM: Corsair Vengeance LPX DDR4 2133Mhz 8GB x4

PSU: Corsair RM750

Case: Fractal Design Define R5

 

well it looks like something has gone wrong with your cooler. perhaps you applied to much thermal paste and after a couple weeks it dried and startet isolating the cpu from the water block.

 

perhaps you have a broken pump or some other issue with the aio.

 

this is weird

 

where you using prime95?

[zafros]

well it looks like something has gone wrong with your cooler. perhaps you applied to much thermal paste and after a couple weeks it dried and startet isolating the cpu from the water block.

 

perhaps you have a broken pump or some other issue with the aio.

 

this is weird

 

where you using prime95?

 

While that certainly is possible, I'm not getting that feeling just yet. I reapplied thermal paste a couple days after getting it all put together, which did drop idle temps by a couple degrees. Purely unscientifically, it does feel like my cooler is moving a lot more heat after today, it's certainly blowing a lot harder, but the air also feels hotter. The pump is at 2650ish rpm according to AIsuite, which it has always been. the cooler fans topped out today for the first time at 2250rpm.

 

Prime95 has been running without a hiccup for the past 4 hours.

 

One thing I did change recently, was to enable Intel VT-X for VMware to run properly on newer OS virtual machines, but I haven't turned it back on again yet after the reset.

[ChrisCross]

While that certainly is possible, I'm not getting that feeling just yet. I reapplied thermal paste a couple days after getting it all put together, which did drop idle temps by a couple degrees. Purely unscientifically, it does feel like my cooler is moving a lot more heat after today, it's certainly blowing a lot harder, but the air also feels hotter. The pump is at 2650ish rpm according to AIsuite, which it has always been. the cooler fans topped out today for the first time at 2250rpm.

Prime95 has been running without a hiccup for the past 4 hours.

One thing I did change recently, was to enable Intel VT-X for VMware to run properly on newer OS virtual machines, but I haven't turned it back on again yet after the reset.

if you used a newer version of prime you probably damaged your cpu.

dont use prime!

[mattebad]

if you used a newer version of prime you probably damaged your cpu.

dont use prime!

Please enlighten me as to how prime can damage your cpu... All it does is generate a butt ton of heat and requires more voltage than the average stress test due to how stupidly hard it is to pass

 

Hi all - first post here

 

I recently converted from that other religion to the whole windows ordeal... anyway bought myself a spanking new PC, in parts.

 

This has unfortunately yielded some interesting results. My i7-5820K overclocked nicely to 4,4 Ghz on the first day of owning it, and bumped it up to 4,5 Ghz the next day. This remained solidly stable for about two weeks, in which time I also overclocked the graphics card. I was hovering around 60 degrees centigrade during full load at 1,3 volts, which seemed very good to me (what do I know, I haven't had a PC in a decade), consuming about 145 Watts.

 

Well today that all changed, it was unstable as F, and in my mind i went WTH WTF just happened?! I basically had to reset it all and start over, which yielded a result of 4,1 Ghz at 1,25 Volts, consuming 180 Watts with temps never dropping below 75 degrees centigrade, and peaking at 80. So while these temps are still safe in Intel language, I'm loosing my mind as to what spurred this sudden change...

 

oh forgot to mention, my cooler is a CM Nepton 240M

 

So I don't really know where to start, i'm hoping some of you guys will have a clue

 

Have a good one

- zafros

 

 

 

 

Full specs below:

 

MB: Asus X99-a

CPU: i7-5820K

GPU: Asus GTX 970 Strix

RAM: Corsair Vengeance LPX DDR4 2133Mhz 8GB x4

PSU: Corsair RM750

Case: Fractal Design Define R5

Your cooler probably wasn't seated properly which is why you had fine temps at first but then they shot up. Reseat your cooler and repaste, make sure it is tightened enough that it can't be moved by your hands. Remember to crossthread as well to avoid air bubbles

[ChrisCross]

Please enlighten me as to how prime can damage your cpu... All it does is generate a butt ton of heat and requires more voltage than the average stress test due to how stupidly hard it is to pass

Your cooler probably wasn't seated properly which is why you had fine temps at first but then they shot up. Reseat your cooler and repaste, make sure it is tightened enough that it can't be moved by your hands. Remember to crossthread as well to avoid air bubbles

well. linus said that in the skylake oc guide which is on vessel. the newer version of prime are powerviruses which mean they generate so much heat and power consumption that its dangerous for your cpu

[zafros]

well, the saga continues.

idle temps are at 26 degrees today, and the bios has apparently reset my OC, so back to the drawing board i go.

[mattebad]

well. linus said that in the skylake oc guide which is on vessel. the newer version of prime are powerviruses which mean they generate so much heat and power consumption that its dangerous for your cpu

But my point was that that doesn't matter if you can cool it. Prime is more of a cooler stress test than anything. It won't damage you cpu unless you are running it for like 24 hours and it is thermal throttling the entire time and you have some insane voltage on it.

 

 

 

 

'Prime95 is not Certified for Haswell/Skylake/Insert Nonsense Here'

This was a myth that was perpetuated by some Youtubers. For Skylake it seems this has died down. But know that Prime95 will not eat your CPU and spit out the remains.

Says it right on the overclockers forum official Skylake overclock guide. I usually trust Linus' word but he is wrong about prime95 being dangerous. Under certain conditions it can be dangerous, but you could say the same for every other stress test if your temps are high enough and you are running them for long enough.

 

Link to guide stating this, it's under the stress testing tab

[ChrisCross]

But my point was that that doesn't matter if you can cool it. Prime is more of a cooler stress test than anything. It won't damage you cpu unless you are running it for like 24 hours and it is thermal throttling the entire time and you have some insane voltage on it.

 

Says it right on the overclockers forum official Skylake overclock guide. I usually trust Linus' word but he is wrong about prime95 being dangerous. Under certain conditions it can be dangerous, but you could say the same for every other stress test if your temps are high enough and you are running them for long enough.

 

Link to guide stating this, it's under the stress testing tab

http://www.overclock.net/t/1510388/haswell-e-overclock-leaderboard-owners-club/2390#post_22900116

 

i would say thats not a bullshit source

[mattebad]

http://www.overclock.net/t/1510388/haswell-e-overclock-leaderboard-owners-club/2390#post_22900116

 

i would say thats not a bullshit source

I'm not saying that you are bullshitting. All I asked for was a source.

 

Raja also said, "It can be lethal for a CPU to see that level of current for prolonged periods."

 

Now even I don't recommend Prime 95 since I think its a retarded test. But like I said before, if you run it long enough, (I exaggeratted a bit before but wasn't too far off) like overnight for around 8-9 hours. It will damage it and maybe kill it. It's a crazy test that shouldn't be used at all.

 

And like I said, "Under certain conditions it can be dangerous, but you could say the same for every other stress test if your temps are high enough and you are running them for long enough."

 

I just didn't think it would kill a cpu, maybe just damage it. Good to know that it can kill one. 1 Point for LMG Media, 0 For Retarded Horse

[ChrisCross]

I'm not saying that you are bullshitting. All I asked for was a source.

 

Raja also said, "It can be lethal for a CPU to see that level of current for prolonged periods."

 

Now even I don't recommend Prime 95 since I think its a retarded test. But like I said before, if you run it long enough, (I exaggeratted a bit before but wasn't too far off) like overnight for around 8-9 hours. It will damage it and maybe kill it. It's a crazy test that shouldn't be used at all.

 

And like I said, "Under certain conditions it can be dangerous, but you could say the same for every other stress test if your temps are high enough and you are running them for long enough."

 

I just didn't think it would kill a cpu, maybe just damage it. Good to know that it can kill one. 1 Point for LMG Media, 0 For Retarded Horse

 

well the first time someone told me that i was thinking he was bullshitting. i was just assuming

[Sintezza]

things i would do.

 

- reseat cpu

- reseat cpu cooler.

- reset bios to stock factory defaults.

- check temps at stock.

- start to overclock again.

- check temps again.

- stable yes / no ?

- Bios updates?

 

You wont damage a cpu that easaly realy.

I´m more likely to think that it might be something according to the mobo or bios.

Or a bios setting.

[Warchylde]

If you figure it out or have any better results be sure to post.

 

I didn't have this same problem but I can't OC my 5820k very high. Running a GTX H100i Cooler. Reseated it 3 times when I first did my build trying to get lower temps.

I must have lost the silicon lottery, cant get above 4.3ghz @ 1.25v w/o getting too hot - 80c @ full load.

[reptileexperts]

I believe the rumors behind the prime95 virus (and likewise the furmark virus) is in issue when using any voltage regulation other than manual. If you use adaptive voltage or auto voltage in prime, it is known to draw too much voltage, which could push you into dangerous territory, or above intels recommended specs. Because haswell has a lower voltage max recommended value over previous generations, you could potentially damage the CPU with a "good' cooler that manages the heat from higher voltages. But these higher voltages again could do permanent damage.

 

Again, my interpretation. I'll stuck with XTU Benchmark stressing, and AIDA64 stability tests FWIW

[mattebad]

I believe the rumors behind the prime95 virus (and likewise the furmark virus) is in issue when using any voltage regulation other than manual. If you use adaptive voltage or auto voltage in prime, it is known to draw too much voltage, which could push you into dangerous territory, or above intels recommended specs. Because haswell has a lower voltage max recommended value over previous generations, you could potentially damage the CPU with a "good' cooler that manages the heat from higher voltages. But these higher voltages again could do permanent damage.

 

Again, my interpretation. I'll stuck with XTU Benchmark stressing, and AIDA64 stability tests FWIW

Aida 64 is great for stress testing, but I'm not so sure about XTU stress test. I usually use a mix of Aida64, Realbench, and OCCT

[reptileexperts]

Aida 64 is great for stress testing, but I'm not so sure about XTU stress test. I usually use a mix of Aida64, Realbench, and OCCT

 

What I've seen / learned with using stress methods is that if I can pass xtu benchmark 10 times in a row, the OC is generally rock stable. It's a quick and dirty check, with Aida64 being used for longer runs.

[bomerr]

If you figure it out or have any better results be sure to post.

 

I didn't have this same problem but I can't OC my 5820k very high. Running a GTX H100i Cooler. Reseated it 3 times when I first did my build trying to get lower temps.

I must have lost the silicon lottery, cant get above 4.3ghz @ 1.25v w/o getting too hot - 80c @ full load.

That's slightly bad but perfectly fine.

 

I hit 44 @ 1.25v and 45@ 1.35v 

 

@ OP your AIO is prob broken