Overclock.net › Forums › Intel › Intel CPUs › Why do OCers set temp limits at 85 in prime? Prime doesn't equate to real world use.
New Posts  All Forums:Forum Nav:

Why do OCers set temp limits at 85 in prime? Prime doesn't equate to real world use. - Page 2

post #11 of 49
Thread Starter 
Well, prime95 is an ultra-extreme example, and no normal application, even benchmarks, will get a computer as hot as prime does. If I'm not folding, don't have crazy volts (Electromigration) and I'm running benchmarks that still fail to get the CPU above 80, why should I care if my chip gets to 86 degrees if I try to use it to calculate multi-million digit mersenne primes?

This post is of a mostly hypothetical nature, as I've hit a voltage wall on my OC and don't want to risk electromigration, but my agrument stands.

My temps are FINE. That's not the point of my question. My temps are excellent for air cooling and 4.4 Ghz. It only grazes 80 in prime. Considering my volts, my temps are perfectly acceptable.
    
CPUMotherboardGraphicsRAM
i7 950 4.41 GHz 1.36V 24/7 5.00 GHz 1.568V HT ON Asus Sabertooth X58 Sapphire Radeon 5770 4GB G Skill Ripjaws DRR3 1600 @ 1685 
Hard DriveOptical DriveOSMonitor
Western Digital Caviar Blue 320GB Sata III HDD LITE-ON 24X DVD Writer Windows 7 Ultimate AOC 20" 
KeyboardPowerCaseMouse
Logitech MX 3200 SeaSonic 520w NZXT M59 Roccat Kone[+] 
Mouse Pad
Padless! 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i7 950 4.41 GHz 1.36V 24/7 5.00 GHz 1.568V HT ON Asus Sabertooth X58 Sapphire Radeon 5770 4GB G Skill Ripjaws DRR3 1600 @ 1685 
Hard DriveOptical DriveOSMonitor
Western Digital Caviar Blue 320GB Sata III HDD LITE-ON 24X DVD Writer Windows 7 Ultimate AOC 20" 
KeyboardPowerCaseMouse
Logitech MX 3200 SeaSonic 520w NZXT M59 Roccat Kone[+] 
Mouse Pad
Padless! 
  hide details  
Reply
post #12 of 49
different programs stress it differently, even if Task manager says both are at 100%

you might have a lot of loading from cache/RAM instead of doing instruction(cache stall), but the CPU is still running at 100% frequency. I suppose Prime pushes with more instructions, and is a better estimation on whether the chip will BSOD at that frequency. You definitely dont want any program to crash your CPU, so having a high upper bound from Prime is nice
Edited by paperwastage - 2/5/11 at 2:45pm
Upgraded Rig
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel i5-2500k Intel DP67DE mATX XFX 5770 1GB 2x4GB G.SKILL DDR3-1333 
Hard DriveOSMonitorPower
1TB Caviar Black Ubuntu 10.04 x64/Windows 7 x64 2xDell 2209WA, 1xV7 Corsair 400W 
CaseMouse
Apex cheap case Trackball 
  hide details  
Reply
Upgraded Rig
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel i5-2500k Intel DP67DE mATX XFX 5770 1GB 2x4GB G.SKILL DDR3-1333 
Hard DriveOSMonitorPower
1TB Caviar Black Ubuntu 10.04 x64/Windows 7 x64 2xDell 2209WA, 1xV7 Corsair 400W 
CaseMouse
Apex cheap case Trackball 
  hide details  
Reply
post #13 of 49
Thread Starter 
Sorry- I just realized why people were giving odd answers. Title changed appropriately.
    
CPUMotherboardGraphicsRAM
i7 950 4.41 GHz 1.36V 24/7 5.00 GHz 1.568V HT ON Asus Sabertooth X58 Sapphire Radeon 5770 4GB G Skill Ripjaws DRR3 1600 @ 1685 
Hard DriveOptical DriveOSMonitor
Western Digital Caviar Blue 320GB Sata III HDD LITE-ON 24X DVD Writer Windows 7 Ultimate AOC 20" 
KeyboardPowerCaseMouse
Logitech MX 3200 SeaSonic 520w NZXT M59 Roccat Kone[+] 
Mouse Pad
Padless! 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i7 950 4.41 GHz 1.36V 24/7 5.00 GHz 1.568V HT ON Asus Sabertooth X58 Sapphire Radeon 5770 4GB G Skill Ripjaws DRR3 1600 @ 1685 
Hard DriveOptical DriveOSMonitor
Western Digital Caviar Blue 320GB Sata III HDD LITE-ON 24X DVD Writer Windows 7 Ultimate AOC 20" 
KeyboardPowerCaseMouse
Logitech MX 3200 SeaSonic 520w NZXT M59 Roccat Kone[+] 
Mouse Pad
Padless! 
  hide details  
Reply
post #14 of 49
Download "Intelburntest" and then post temps.

Sent from my GT-I9000 using Tapatalk
Eysenck Box
(12 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1600 Asus B350M-A XFX AMD Vega RX 64 8GB 2x4GB DDR4 2400mhz 
Hard DriveCoolingOSMonitor
128GB Intel M.2 SSD Stock Windows 10 Pro x64 LG 27UD58 4K 
KeyboardPowerCaseMouse
Razer Blackwidow (Cherry Blue) Corsair TX550 Silverstone TJ08-E Logitech G403 
  hide details  
Reply
Eysenck Box
(12 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1600 Asus B350M-A XFX AMD Vega RX 64 8GB 2x4GB DDR4 2400mhz 
Hard DriveCoolingOSMonitor
128GB Intel M.2 SSD Stock Windows 10 Pro x64 LG 27UD58 4K 
KeyboardPowerCaseMouse
Razer Blackwidow (Cherry Blue) Corsair TX550 Silverstone TJ08-E Logitech G403 
  hide details  
Reply
post #15 of 49
Quote:
Originally Posted by roflolol View Post
Well, prime95 is an ultra-extreme example, and no normal application, even benchmarks, will get a computer as hot as prime does. If I'm not folding, don't have crazy volts (Electromigration) and I'm running benchmarks that still fail to get the CPU above 80, why should I care if my chip gets to 86 degrees if I try to use it to calculate multi-million digit mersenne primes?

This post is of a mostly hypothetical nature, as I've hit a voltage wall on my OC and don't want to risk electromigration, but my agrument stands.

My temps are FINE. That's not the point of my question. My temps are excellent for air cooling and 4.4 Ghz. It only grazes 80 in prime. Considering my volts, my temps are perfectly acceptable.
We set an 85C limit because 85C is really hot!. We know that in a real world scenario it will never get that hot, but that is beside the point. We stop testing at 85C because, even knowing we can oc more, the temps we get testing our oc stability are not worth the cpu degradation risk.
Vengeance 2
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k@4.70GHz(1.350Vcore) Asus P8P67 Pro 2XHD 5770@920c-1300m 4GB Corsair 9-9-9-24@1600(1.64v) 
Hard DriveOptical DriveOSMonitor
Seagate Barracuda 1TB Samsung Multi DVDDL/CD RW Windows 7 Home Premium 64bit HP w2207h@1680x1050 
KeyboardPowerCaseMouse
Microsoft USB Antec TP-750W Antec 300 Illusion Microsoft USB 
Mouse Pad
LOTR-ROTK-Gandalf 
  hide details  
Reply
Vengeance 2
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k@4.70GHz(1.350Vcore) Asus P8P67 Pro 2XHD 5770@920c-1300m 4GB Corsair 9-9-9-24@1600(1.64v) 
Hard DriveOptical DriveOSMonitor
Seagate Barracuda 1TB Samsung Multi DVDDL/CD RW Windows 7 Home Premium 64bit HP w2207h@1680x1050 
KeyboardPowerCaseMouse
Microsoft USB Antec TP-750W Antec 300 Illusion Microsoft USB 
Mouse Pad
LOTR-ROTK-Gandalf 
  hide details  
Reply
post #16 of 49
Quote:
Originally Posted by techheadtrevor View Post
We set an 85C limit because 85C is really hot!. We know that in a real world scenario it will never get that hot, but that is beside the point. We stop testing at 85C because, even knowing we can oc more, the temps we get testing our oc stability are not worth the cpu degradation risk.
It is "really hot" if you are on the beach on some alien planet.

There is practically no degradation risk if the chips are running with spec.

Depending on the model you have, these chips are fine at 80 or even 90. Some don't even throttle untlil they hit 95. Which raises the second point, since it throttles down when it hits that high of a temp, why do people worry at all about 90 or so??
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k Asus Pro Sapphire 6970 Corsair 4x4 
Hard DriveOSPowerMouse Pad
corsair 300 win 7 x64 None My desk 
  hide details  
Reply
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k Asus Pro Sapphire 6970 Corsair 4x4 
Hard DriveOSPowerMouse Pad
corsair 300 win 7 x64 None My desk 
  hide details  
Reply
post #17 of 49
Quote:
Originally Posted by gooddog View Post
It is "really hot" if you are on the beach on some alien planet.

There is practically no degradation risk if the chips are running with spec.

Depending on the model you have, these chips are fine at 80 or even 90. Some don't even throttle untlil they hit 95. Which raises the second point, since it throttles down when it hits that high of a temp, why do people worry at all about 90 or so??
Yes , yes, of course.. how could I have missed it! It is called common sense. You know that thing in the back of you mind telling you that something your about to do is a bad idea. AS I have stated before...85C or even 90C might not kill your cpu in 5sec...but why put your cpu through that for a minor overlclocking gain? I think it is funny that you guys use intels 95C throttle temp as a safe max.
Using the chips own self preservation mechanism as a temp limit is asinine. That mechanism is meant to save the chip from immediate death, and you say 5c away from that is fine?

The rest of us actually care about our systems.
Edited by techheadtrevor - 2/5/11 at 3:44pm
Vengeance 2
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k@4.70GHz(1.350Vcore) Asus P8P67 Pro 2XHD 5770@920c-1300m 4GB Corsair 9-9-9-24@1600(1.64v) 
Hard DriveOptical DriveOSMonitor
Seagate Barracuda 1TB Samsung Multi DVDDL/CD RW Windows 7 Home Premium 64bit HP w2207h@1680x1050 
KeyboardPowerCaseMouse
Microsoft USB Antec TP-750W Antec 300 Illusion Microsoft USB 
Mouse Pad
LOTR-ROTK-Gandalf 
  hide details  
Reply
Vengeance 2
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k@4.70GHz(1.350Vcore) Asus P8P67 Pro 2XHD 5770@920c-1300m 4GB Corsair 9-9-9-24@1600(1.64v) 
Hard DriveOptical DriveOSMonitor
Seagate Barracuda 1TB Samsung Multi DVDDL/CD RW Windows 7 Home Premium 64bit HP w2207h@1680x1050 
KeyboardPowerCaseMouse
Microsoft USB Antec TP-750W Antec 300 Illusion Microsoft USB 
Mouse Pad
LOTR-ROTK-Gandalf 
  hide details  
Reply
post #18 of 49
Quote:
Originally Posted by techheadtrevor View Post
Yes , yes, of course.. how could I have missed it! It is called common sense. You know that thing in the back of you mind telling you that something your about to do is a bad idea. AS I have stated before...85C or even 90C might not kill your cpu in 5sec...but why put your cpu through that for a minor overlclocking gain? I think it is funny that you guys use intels 95C throttle temp as a safe max.
Using the chips own self preservation mechanism as a temp limit is asinine. That mechanism is meant to save the chip from immediate death, and you say 5c away from that is fine?

The rest of us actually care about our systems.
This. +rep for common sense and a great explanation.
Eysenck Box
(12 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1600 Asus B350M-A XFX AMD Vega RX 64 8GB 2x4GB DDR4 2400mhz 
Hard DriveCoolingOSMonitor
128GB Intel M.2 SSD Stock Windows 10 Pro x64 LG 27UD58 4K 
KeyboardPowerCaseMouse
Razer Blackwidow (Cherry Blue) Corsair TX550 Silverstone TJ08-E Logitech G403 
  hide details  
Reply
Eysenck Box
(12 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1600 Asus B350M-A XFX AMD Vega RX 64 8GB 2x4GB DDR4 2400mhz 
Hard DriveCoolingOSMonitor
128GB Intel M.2 SSD Stock Windows 10 Pro x64 LG 27UD58 4K 
KeyboardPowerCaseMouse
Razer Blackwidow (Cherry Blue) Corsair TX550 Silverstone TJ08-E Logitech G403 
  hide details  
Reply
post #19 of 49
It's to check for stability. Prime puts so much stresses on your CPU, tons more than your average game.


It's to be safe, you could run a game just fine, but it may not be 100% stable.
post #20 of 49
We set a safe temperature max to make sure you don't kill your chip.
Prime95 (or equivalent) is the only thing that will test your CPU for stability, generate the most heat, and use the most resources.
Without prime how would you test if your OC is stable?
You will never reach those temps in real world use, but you will in prime. And prime is the only way to test OC stability.
So you don't really have a choice. And so a safe max temp was created.

Example scenario:
Your temps are 85c at 4.0ghz.
You are aiming for 4.2ghz and your temps exceed 90.
You know that your real world temps will never exceed ~75, but you can't maintain <85 in prime.
You'll never get a stable 4.2.
We have no choice.
Edited by kiwiasian - 2/5/11 at 4:19pm
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
  • Why do OCers set temp limits at 85 in prime? Prime doesn't equate to real world use.
Overclock.net › Forums › Intel › Intel CPUs › Why do OCers set temp limits at 85 in prime? Prime doesn't equate to real world use.