1800X Temps too high with H115i. What's the issue?

[TheNaitsyrk]

Hiho,

 

I benched my CPU back in Sept 2017 with 4.0Ghz oc and 2933Mhz RAM. Fans on full speed as they're quiet and temps were 59C or 60C under fullest load in Aida64.

 

I updated the BIOS of my Crosshair VI HERO motherboard and decided to bench again.

 

This time, the CPU was 77-80C and I have no idea why. It's a 280mm water cooler with fans at full speed (it's also a sandwich 4x fans around the h115i)

 

I thought it's the BIOS so I downgraded to the one I was using previously. Temps just 3-4C lower with the same OC settings, 75C average.

 

HOWEVER,

 

CPU stock with old bios now is cold, just like before only 50C at MAX.

 

With OC (300 Mhz oc literally) 75C. 25C increase? Makes no sense.

 

With new BIOS stock 1800X got 64C hot. STOCK, NO OC, not even memory with H115i with ramped up fans. What the hell?

 

I'm using offset so CPU takes as much as it needs, and it's a tiny offset, nothing big either.

 

Any idea what the problem could be? I was thinking maybe the cooler itself. Thanks in advance.

[knightslugger]

Probably Asus screwing with temp sensors again.

[RadiatingLight]

1800X has a +20c offset. it's actually 60c not 80c (unless you're using a program that corrects for that)

[TheNaitsyrk]

2 minutes ago, RadiatingLight said:

1800X has a +20c offset. it's actually 60c not 80c (unless you're using a program that corrects for that)

I don't think there's an offset any more. Back in Sept it was showing right temps in both Aida and RyzenMaster.

[For Science!]

15 minutes ago, TheNaitsyrk said:

I don't think there's an offset any more. Back in Sept it was showing right temps in both Aida and RyzenMaster.

Are idle temps the same with the 2 Bioses?

[LienusLateTips]

What case?

[RadiatingLight]

21 minutes ago, TheNaitsyrk said:

I don't think there's an offset any more. Back in Sept it was showing right temps in both Aida and RyzenMaster.

Hm.. good point.

[TheNaitsyrk]

8 hours ago, For Science! said:

Are idle temps the same with the 2 Bioses?

I'm not sure. It jumps and goes down then jumps and goes down again. It's never still in RyzenMaster. 

 

8 hours ago, JDE said:

What case?

Phanteks Enthoo Pro M. But it doesn't matter because I already mentioned that temps were fine at one point. 59C max under full load in that exact case.

[For Science!]

8 minutes ago, TheNaitsyrk said:

I'm not sure. It jumps and goes down then jumps and goes down again. It's never still in RyzenMaster. 

Long story short, check the sensemi skew setting in the BIOS. It's an option that artificially lowers the CPU temp by 20 degrees or so for overclockers so that the computer doesnt force shut down due to spikes in temps.

[TheNaitsyrk]

1 minute ago, For Science! said:

Long story short, check the sensemi skew setting in the BIOS. It's an option that artificially lowers the CPU temp by 20 degrees or so for overclockers so that the computer doesnt force shut down due to spikes in temps.

If it's enabled then it means that BIOS was bamboozling me with temps?

[For Science!]

Just now, TheNaitsyrk said:

If it's enabled then it means that BIOS was bamboozling me with temps?

yes. I had a similar "issue" on my Threadripper ASUS board. Basically if you at any point "load 4 GHz profile" in the BIOS (As I did) at any point the senseMI skew is likely to have been enabled. It was very obvious for me because after loading this profile my idle temps (as shown both by the motherboard OLED and HWInfo64) were sub-ambient, as we all know that is not possible unless you are doing some exotic stuff (which I wasn't).

[TheNaitsyrk]

7 hours ago, For Science! said:

yes. I had a similar "issue" on my Threadripper ASUS board. Basically if you at any point "load 4 GHz profile" in the BIOS (As I did) at any point the senseMI skew is likely to have been enabled. It was very obvious for me because after loading this profile my idle temps (as shown both by the motherboard OLED and HWInfo64) were sub-ambient, as we all know that is not possible unless you are doing some exotic stuff (which I wasn't).

Thank you. Once I'm next to my pc I will check it. I'll update once I get to it asap.

 

Alright. I disabled SenseMI in Tweaker's Paradise and same story. Temps are high. Should I update the BIOS and try then?

[Salacious B Crum]

I may have missed it but when you put the overclock back to stock did you reset the cmos? try that

[TheNaitsyrk]

1 hour ago, mrbilky said:

I may have missed it but when you put the overclock back to stock did you reset the cmos? try that

I updated the BIOS again. I'm just getting proof now. I will reset the CMOS just after I'm done with this. I will post results of RyzenMaster, CPU-Z and Aida64. Stock and OCed.

 

8 hours ago, For Science! said:

yes. I had a similar "issue" on my Threadripper ASUS board. Basically if you at any point "load 4 GHz profile" in the BIOS (As I did) at any point the senseMI skew is likely to have been enabled. It was very obvious for me because after loading this profile my idle temps (as shown both by the motherboard OLED and HWInfo64) were sub-ambient, as we all know that is not possible unless you are doing some exotic stuff (which I wasn't).

Here's the Stock and OC states. OC literally crashed (not literally, just stopped the test, but it's definitely stable at 4.0Ghz because it worked with previous bioses etc). I think I forgot to give the RAM the right voltage at 3200Mhz so that must be the reason for hardware failure.

I used Pstate OC just like with the old BIOS. I can now do normal OC but results are the same.

 

SenseMI Skew changed nothing.

 

I'm out of options. I don't know what's causing it. Should I take off my cooler and reapply thermal paste or something?

[AngryBeaver]

My guess is your voltages are too high. You probably need to dial them in. I would double check what your voltage is under load.

[TheNaitsyrk]

2 hours ago, AngryBeaver said:

My guess is your voltages are too high. You probably need to dial them in. I would double check what your voltage is under load.

No because previously volages weren't to high, and nothing changed.

 

I did and it's actually less than before. 1.376V under load.

[RyzenUp]

Thermal paste check your Thermal paste

[partymarty]

3 hours ago, ForChickenNuggets said:

Thermal paste check your Thermal paste

That wouldn't have made such an enormous difference is such short amount of time. Thermal paste won't go bad in 4-5 months. If the application was good at the start of the build, than it's still good

[TheNaitsyrk]

1 hour ago, partymarty said:

That wouldn't have made such an enormous difference is such short amount of time. Thermal paste won't go bad in 4-5 months. If the application was good at the start of the build, than it's still good

Correct. I'm also not using some random cancer. I'm using Thermal Grizzly.

[TheNaitsyrk]

1 hour ago, partymarty said:

That wouldn't have made such an enormous difference is such short amount of time. Thermal paste won't go bad in 4-5 months. If the application was good at the start of the build, than it's still good

I also flashed the newest BIOS released yesterday 6001 for Crosshair VI HERO and temps are even worse. 1.267V at stock, RAM at 2133 and CPU got to 67C with fans blasting at 100%... 4 of them sandwiched around the radiator. Imagine what would happen if I left them at like 40%...

 

I contacted Corsair about this since no one knows what is happening, and they said that they'll RMA my H155i.

[partymarty]

1 hour ago, TheNaitsyrk said:

I also flashed the newest BIOS released yesterday 6001 for Crosshair VI HERO and temps are even worse. 1.267V at stock, RAM at 2133 and CPU got to 67C with fans blasting at 100%... 4 of them sandwiched around the radiator. Imagine what would happen if I left them at like 40%...

 

I contacted Corsair about this since no one knows what is happening, and they said that they'll RMA my H155i.

You could try to make sure if it's the cooler by taking some random aircooler and holding it onto the chip and bootup the pc and then check the temperatures.

If they are lower it's probably the cooler. If they are higher it's most likely something else (mobo or cpu). 

 

But to do this you'll need some random aircooler laying around.

[TheNaitsyrk]

6 hours ago, partymarty said:

You could try to make sure if it's the cooler by taking some random aircooler and holding it onto the chip and bootup the pc and then check the temperatures.

If they are lower it's probably the cooler. If they are higher it's most likely something else (mobo or cpu). 

 

But to do this you'll need some random aircooler laying around.

This is AM4, I have no random "AM4" cooler lying around.

[knightslugger]

Just now, TheNaitsyrk said:

This is AM4, I have no random "AM4" cooler lying around.

just hold something down, it doesn't need a crapton of force.

[TheNaitsyrk]

Just now, knightslugger said:

just hold something down, it doesn't need a crapton of force.

If I showed you a pic of my PC you'd be like "never mind". Plus, it's either this or the motherboard. If this persists after H115i swap then I'll RMA the mobo.

[knightslugger]

Just now, TheNaitsyrk said:

If I showed you a pic of my PC you'd be like "never mind".

i personally don't think that's the issue though. I think this is just the BIOS manipulating the sensor data.