Overclock.net › Forums › Intel › Intel CPUs › Ivy Bridge - Extreme Core Temperature Differences
New Posts  All Forums:Forum Nav:

Ivy Bridge - Extreme Core Temperature Differences

post #1 of 8
Thread Starter 
Hi.

First I would write important components for my problem from my new rig because I didn't catch time to update my signature (add new rig).

MBO: Gigabyte Z77X-UD5H (BIOS F7)
CPU: i5-3570K (turbo off)
Cooler: Scythe Ninja 3
....


After I got all the components I start testing on my test bench. First test was with everything on default even with the Intel stock cooler. Even at that initial testing I noticed a huge difference between the cores temps.

Idle (room temperature 24C)
C1 = 26-35C
C2 = 25-32C
C3 = 26-36C
C4 = 18-29C

Load
C1 = 53-54C
C2 = 50-51C
C3 = 55-57C
C4 = 43-45C

After that I thought maybe stock cooler did not seat well, and I start to mount the Scythe Ninja 3 cooler and after that the difference became even greater up to 13C between highest temp on core 3 and core 4, after OC-ing on 4.7GHz even 14C.

After OC-ing on 4.7GHz (1.308V)
Idle
C1 = 24-32C
C2 = 19-32C
C3 = 24-33C
C4 = 13-32C


Load
C1 = 58-66C
C2 = 63-67C
C3 = 65-70C
C4 = 53-59C

Even I thought it impossible to have two bad cooler mounting, I decided to reseat cooler again and of course after reseat same thing.

All cores are on same multiplier (turbo off).

One more interesting and significant thing is that 4th core (and core 2 in a small extent) on idle every 10-15 seconds fall to 11-14C (core 2 usually not below 18C) for 1-2 seconds which is impossible because I'm on air cooler and room temperature is 23-25C.

My conclusion is that something is wrong with temperature probes in the CPU. But now I have a fear whether these other probes (not for core 4) measure accurately.

I'll put the images below so you can see how that looks.

495

492
Edited by xPhoto - 6/12/12 at 12:24pm
 
HTPC i3 530@4GHz
(13 items)
 
 
CPUMotherboardGraphicsRAM
AMD Phenom 960T X4  Gigabyte GA-990XA-UD3 HD4350 x3 Corsair DDR3 1600MHz 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Scythe Mugen 2 Windows 7 64-bit SP1 AOC 2236 x6 
KeyboardPowerCaseMouse
Logitech G11 Corsair TX650W CoolerMaster CM690 Logitech G700 
CPUMotherboardGraphicsRAM
Intel i3 530 Gigabyte GA-H55M-D2H Intel® HD Graphics Corsair DDR3 1600MHz 
Hard DriveOptical DriveCoolingOS
WD Green Caviar 1T DVD-RW Scythe SAMURAI-ZZ Windows 7 64-Bit SP1 
MonitorPowerCaseAudio
Philips Cineos Widescreen Flat TV 32PF9541 Corsair CX430 Antec fusion remote silver Harman Kardon AVR 235 7.1 
Other
Speakers Jamo S606-hcs-3 5.0 + Jamo SUB250 
CPUMotherboardGraphicsRAM
AMD Athlon 64 3500+ DFI LANParty NF4 SLI-DR ATI x700 x2 Kingmax DDR 500MHZ 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Thermaltake Sonic Tower Windows XP SP3 Mitshubishi Diamond Pro 2070SB 
MonitorMonitorMonitorKeyboard
LG LCD 17" Samsung LCD 17" Hansol 920D Logitech Multimedia 
PowerCaseMouse
Tagan 480W CoolerMaster CM690 Logitech MX510 
  hide details  
Reply
 
HTPC i3 530@4GHz
(13 items)
 
 
CPUMotherboardGraphicsRAM
AMD Phenom 960T X4  Gigabyte GA-990XA-UD3 HD4350 x3 Corsair DDR3 1600MHz 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Scythe Mugen 2 Windows 7 64-bit SP1 AOC 2236 x6 
KeyboardPowerCaseMouse
Logitech G11 Corsair TX650W CoolerMaster CM690 Logitech G700 
CPUMotherboardGraphicsRAM
Intel i3 530 Gigabyte GA-H55M-D2H Intel® HD Graphics Corsair DDR3 1600MHz 
Hard DriveOptical DriveCoolingOS
WD Green Caviar 1T DVD-RW Scythe SAMURAI-ZZ Windows 7 64-Bit SP1 
MonitorPowerCaseAudio
Philips Cineos Widescreen Flat TV 32PF9541 Corsair CX430 Antec fusion remote silver Harman Kardon AVR 235 7.1 
Other
Speakers Jamo S606-hcs-3 5.0 + Jamo SUB250 
CPUMotherboardGraphicsRAM
AMD Athlon 64 3500+ DFI LANParty NF4 SLI-DR ATI x700 x2 Kingmax DDR 500MHZ 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Thermaltake Sonic Tower Windows XP SP3 Mitshubishi Diamond Pro 2070SB 
MonitorMonitorMonitorKeyboard
LG LCD 17" Samsung LCD 17" Hansol 920D Logitech Multimedia 
PowerCaseMouse
Tagan 480W CoolerMaster CM690 Logitech MX510 
  hide details  
Reply
post #2 of 8
The first thing that came to my mind was the flatness of your CPU's heat spreader - but the 'lower than room temps' kinda makes this not matter rolleyes.gif

My 2-cents are that: this is strange, could be bad CPU temp probes, could be your mb/bios and in the end does it really matter? thinking.gif Let's think about this. Cores 2,3 will run hotter than 1,4 because they are in the center (my 2,3 run 8 to 10C hotter than 1,4 - my IHS is also concave) and those center probes look likely good enough. your fan speeds will be based on either your hottest core or the package temp i believe so your cooling should be effective... unless the low reading core is the only one fully loaded and it runs much hotter than reported.. I'm not sure how much heat a single core could crank out and how much your cooling solution could dissipate on low fan so i suppose there is the possible risk of damage.

Otherwise i'd say your temps are golden at that clock (assuming they were taken while running IBT or the second test sequence from prime95 or a long running test).

So I'd say you have 3 options.
  1. dont care and hope intels warranty will cover it if it fails.
  2. take intels overclock insurance and know you are covered (google it if you are not aware of it).
  3. send it back now and maybe get a CPU you cant hit 4.7Ghz with.

EDIT:
I looked up your cooler, that thing has a LOT of copper in it - i'm still not sure how much heat one core can crank out but i bet that cooler can nearly passively handle it helpinghand.gif
Edited by Ross K - 6/12/12 at 1:37pm
3770K @4.7Ghz
(14 items)
 
  
CPUMotherboardGraphicsRAM
3770K GIGABYTE G1.Sniper 3 GIGABYTE GV-N670OC-2GD GeForce GTX 670 G.SKILL Trident X DDR3 2400 F3-2400C10D-8GTX 
Hard DriveCoolingOSMonitor
Samsung 830 XSPC RX360 Kit Windows 7 Home BenQ XL2420T Black-Red 24" 120hz 
MonitorKeyboardPowerCase
BenQ FP241VW 24" TFT ROCCAT Isku Illuminated Gaming Keyboard (ROC-12... CORSAIR CMPSU-750TX 750W HAF 932 Advanced  
MouseMouse Pad
Razer Deathadder Black 3500 ROCCAT Alumic Double-Sided Gaming Mousepad (ROC... 
  hide details  
Reply
3770K @4.7Ghz
(14 items)
 
  
CPUMotherboardGraphicsRAM
3770K GIGABYTE G1.Sniper 3 GIGABYTE GV-N670OC-2GD GeForce GTX 670 G.SKILL Trident X DDR3 2400 F3-2400C10D-8GTX 
Hard DriveCoolingOSMonitor
Samsung 830 XSPC RX360 Kit Windows 7 Home BenQ XL2420T Black-Red 24" 120hz 
MonitorKeyboardPowerCase
BenQ FP241VW 24" TFT ROCCAT Isku Illuminated Gaming Keyboard (ROC-12... CORSAIR CMPSU-750TX 750W HAF 932 Advanced  
MouseMouse Pad
Razer Deathadder Black 3500 ROCCAT Alumic Double-Sided Gaming Mousepad (ROC... 
  hide details  
Reply
post #3 of 8
Thread Starter 
Quote:
Originally Posted by Ross K View Post

The first thing that came to my mind was the flatness of your CPU's heat spreader - but the 'lower than room temps' kinda makes this not matter rolleyes.gif
My 2-cents are that: this is strange, could be bad CPU temp probes, could be your mb/bios and in the end does it really matter? thinking.gif Let's think about this. Cores 2,3 will run hotter than 1,4 because they are in the center (my 2,3 run 8 to 10C hotter than 1,4 - my IHS is also concave) and those center probes look likely good enough. your fan speeds will be based on either your hottest core or the package temp i believe so your cooling should be effective... unless the low reading core is the only one fully loaded and it runs much hotter than reported.. I'm not sure how much heat a single core could crank out and how much your cooling solution could dissipate on low fan so i suppose there is the possible risk of damage.
Otherwise i'd say your temps are golden at that clock (assuming they were taken while running IBT or the second test sequence from prime95 or a long running test).
So I'd say you have 3 options.
  1. dont care and hope intels warranty will cover it if it fails.
  2. take intels overclock insurance and know you are covered (google it if you are not aware of it).
  3. send it back now and maybe get a CPU you cant hit 4.7Ghz with.
EDIT:
I looked up your cooler, that thing has a LOT of copper in it - i'm still not sure how much heat one core can crank out but i bet that cooler can nearly passively handle it helpinghand.gif

Ross thanks. As a reference value I take hottest core (core 3) or the CPU package when I OCing but still core 4 reading is annoying . But as it looks like this chip is very good sample so I'm no going to return to shop where I bought it, I doubt that thay would take it under warranty just because of temp reading.

Load temp snapshot was taken at the end of wprime 1024M test.

About concave IHS. After removing heatsink thermal paste was evenly spreaded so I think that IHS and heatsink are OK.
 
HTPC i3 530@4GHz
(13 items)
 
 
CPUMotherboardGraphicsRAM
AMD Phenom 960T X4  Gigabyte GA-990XA-UD3 HD4350 x3 Corsair DDR3 1600MHz 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Scythe Mugen 2 Windows 7 64-bit SP1 AOC 2236 x6 
KeyboardPowerCaseMouse
Logitech G11 Corsair TX650W CoolerMaster CM690 Logitech G700 
CPUMotherboardGraphicsRAM
Intel i3 530 Gigabyte GA-H55M-D2H Intel® HD Graphics Corsair DDR3 1600MHz 
Hard DriveOptical DriveCoolingOS
WD Green Caviar 1T DVD-RW Scythe SAMURAI-ZZ Windows 7 64-Bit SP1 
MonitorPowerCaseAudio
Philips Cineos Widescreen Flat TV 32PF9541 Corsair CX430 Antec fusion remote silver Harman Kardon AVR 235 7.1 
Other
Speakers Jamo S606-hcs-3 5.0 + Jamo SUB250 
CPUMotherboardGraphicsRAM
AMD Athlon 64 3500+ DFI LANParty NF4 SLI-DR ATI x700 x2 Kingmax DDR 500MHZ 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Thermaltake Sonic Tower Windows XP SP3 Mitshubishi Diamond Pro 2070SB 
MonitorMonitorMonitorKeyboard
LG LCD 17" Samsung LCD 17" Hansol 920D Logitech Multimedia 
PowerCaseMouse
Tagan 480W CoolerMaster CM690 Logitech MX510 
  hide details  
Reply
 
HTPC i3 530@4GHz
(13 items)
 
 
CPUMotherboardGraphicsRAM
AMD Phenom 960T X4  Gigabyte GA-990XA-UD3 HD4350 x3 Corsair DDR3 1600MHz 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Scythe Mugen 2 Windows 7 64-bit SP1 AOC 2236 x6 
KeyboardPowerCaseMouse
Logitech G11 Corsair TX650W CoolerMaster CM690 Logitech G700 
CPUMotherboardGraphicsRAM
Intel i3 530 Gigabyte GA-H55M-D2H Intel® HD Graphics Corsair DDR3 1600MHz 
Hard DriveOptical DriveCoolingOS
WD Green Caviar 1T DVD-RW Scythe SAMURAI-ZZ Windows 7 64-Bit SP1 
MonitorPowerCaseAudio
Philips Cineos Widescreen Flat TV 32PF9541 Corsair CX430 Antec fusion remote silver Harman Kardon AVR 235 7.1 
Other
Speakers Jamo S606-hcs-3 5.0 + Jamo SUB250 
CPUMotherboardGraphicsRAM
AMD Athlon 64 3500+ DFI LANParty NF4 SLI-DR ATI x700 x2 Kingmax DDR 500MHZ 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Thermaltake Sonic Tower Windows XP SP3 Mitshubishi Diamond Pro 2070SB 
MonitorMonitorMonitorKeyboard
LG LCD 17" Samsung LCD 17" Hansol 920D Logitech Multimedia 
PowerCaseMouse
Tagan 480W CoolerMaster CM690 Logitech MX510 
  hide details  
Reply
post #4 of 8
Every Ivy Bridge I have seen is like this.
post #5 of 8
Thread Starter 
Quote:
Originally Posted by Darth Oscar View Post

Every Ivy Bridge I have seen is like this.

I notice by screenshots from Ivy owners on this forum that difference between hottest and coldest cores are bigger then on my old i7 920 but all in range 5-6C nothing near 12-14C, and even more strange in my case is droping 10C below room tepmerature on idle with air cooler.
 
HTPC i3 530@4GHz
(13 items)
 
 
CPUMotherboardGraphicsRAM
AMD Phenom 960T X4  Gigabyte GA-990XA-UD3 HD4350 x3 Corsair DDR3 1600MHz 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Scythe Mugen 2 Windows 7 64-bit SP1 AOC 2236 x6 
KeyboardPowerCaseMouse
Logitech G11 Corsair TX650W CoolerMaster CM690 Logitech G700 
CPUMotherboardGraphicsRAM
Intel i3 530 Gigabyte GA-H55M-D2H Intel® HD Graphics Corsair DDR3 1600MHz 
Hard DriveOptical DriveCoolingOS
WD Green Caviar 1T DVD-RW Scythe SAMURAI-ZZ Windows 7 64-Bit SP1 
MonitorPowerCaseAudio
Philips Cineos Widescreen Flat TV 32PF9541 Corsair CX430 Antec fusion remote silver Harman Kardon AVR 235 7.1 
Other
Speakers Jamo S606-hcs-3 5.0 + Jamo SUB250 
CPUMotherboardGraphicsRAM
AMD Athlon 64 3500+ DFI LANParty NF4 SLI-DR ATI x700 x2 Kingmax DDR 500MHZ 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Thermaltake Sonic Tower Windows XP SP3 Mitshubishi Diamond Pro 2070SB 
MonitorMonitorMonitorKeyboard
LG LCD 17" Samsung LCD 17" Hansol 920D Logitech Multimedia 
PowerCaseMouse
Tagan 480W CoolerMaster CM690 Logitech MX510 
  hide details  
Reply
 
HTPC i3 530@4GHz
(13 items)
 
 
CPUMotherboardGraphicsRAM
AMD Phenom 960T X4  Gigabyte GA-990XA-UD3 HD4350 x3 Corsair DDR3 1600MHz 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Scythe Mugen 2 Windows 7 64-bit SP1 AOC 2236 x6 
KeyboardPowerCaseMouse
Logitech G11 Corsair TX650W CoolerMaster CM690 Logitech G700 
CPUMotherboardGraphicsRAM
Intel i3 530 Gigabyte GA-H55M-D2H Intel® HD Graphics Corsair DDR3 1600MHz 
Hard DriveOptical DriveCoolingOS
WD Green Caviar 1T DVD-RW Scythe SAMURAI-ZZ Windows 7 64-Bit SP1 
MonitorPowerCaseAudio
Philips Cineos Widescreen Flat TV 32PF9541 Corsair CX430 Antec fusion remote silver Harman Kardon AVR 235 7.1 
Other
Speakers Jamo S606-hcs-3 5.0 + Jamo SUB250 
CPUMotherboardGraphicsRAM
AMD Athlon 64 3500+ DFI LANParty NF4 SLI-DR ATI x700 x2 Kingmax DDR 500MHZ 
Hard DriveCoolingOSMonitor
Samsung HD103SJ Thermaltake Sonic Tower Windows XP SP3 Mitshubishi Diamond Pro 2070SB 
MonitorMonitorMonitorKeyboard
LG LCD 17" Samsung LCD 17" Hansol 920D Logitech Multimedia 
PowerCaseMouse
Tagan 480W CoolerMaster CM690 Logitech MX510 
  hide details  
Reply
post #6 of 8
It seems to be a design fault, my 3770K has 13*C difference between cores mad.gif
At 1.3Vcore, 4.7GHz (100x47) and 70 minutes IBT Maximum the temps are: 69/79/82/75
I'm using a Zalman CNPS10X Extreme,t he IHS is lapped and the TC is changed to Zalman STG-2.
post #7 of 8
I ran a 2500k for a year overclocked to 5Ghz 1.48v with 12c difference between hottest and coldest core even after lapping the CPU. I still have it and works fine.

My 3930k has a 16c difference between hottest and coldest core but that's because it's VERY concave so this is an example on the more extreme side.

It's common, nothing to worry about.

As far as temps being below ambient, I agree it's impossible however I suspect it is the monitoring software at fault. It's always recommended to use the latest RealTemp for Intel CPU's.

If RealTemp still reports sub ambient temps for the coolest core, it's def a bad sensor. I wouldn't RMA for that personally as I always go by the hottest or second hottest core and it would cause downtime, however it will likely hurt the resell value.
Edited by Scorpion667 - 6/25/12 at 3:39pm
    
CPUMotherboardGraphicsRAM
3930k 4.7 1.368v RIVE BIOS 2105 680L 1330/7011 16GB Samsung 2133Mhz  
Hard DriveCoolingOSMonitor
Crucial M4 240GB  H110 after 4x H220 RMA Win 8.1 64bit XL2420T Lightboost  
KeyboardPowerCaseMouse
SS 6Gv2  EVGA 1300G2  Arc Midi R2 Zowie AM, many opticals 
Mouse PadAudioAudioAudio
SS QCK+ Xonar Essence STX AK100 MKII Shure SE846  
AudioAudioOther
Ultrasone Pro 900 Sennheiser HD700 Intel EXPI9301CT NIC 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
3930k 4.7 1.368v RIVE BIOS 2105 680L 1330/7011 16GB Samsung 2133Mhz  
Hard DriveCoolingOSMonitor
Crucial M4 240GB  H110 after 4x H220 RMA Win 8.1 64bit XL2420T Lightboost  
KeyboardPowerCaseMouse
SS 6Gv2  EVGA 1300G2  Arc Midi R2 Zowie AM, many opticals 
Mouse PadAudioAudioAudio
SS QCK+ Xonar Essence STX AK100 MKII Shure SE846  
AudioAudioOther
Ultrasone Pro 900 Sennheiser HD700 Intel EXPI9301CT NIC 
  hide details  
Reply
post #8 of 8
Apologies for getting this one going again.

I have the same problem with my i5-3570K where core #4 is up to 10 degree's different but it appears to have got worse since installing the CPU from new. I will redo TIM this week and see if it solves the problem.

My chip is not brilliant, only manages 4.4Ghz on air before a significant increase in Vcore is required. mad.gif
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
Overclock.net › Forums › Intel › Intel CPUs › Ivy Bridge - Extreme Core Temperature Differences