First time overclocking, confused on temps and need help!

genardas · Dec 22, 2013

PC specs:

CPU: Intel Core i5-4670K 3.4GHz Quad-Core Processor
CPU Cooler: Cooler Master Hyper 212 EVO 82.9 CFM Sleeve Bearing CPU Cooler
Motherboard: Asus Z87-A ATX LGA1150 Motherboard
Memory: G.Skill Sniper Series 8GB (2 x 4GB) DDR3-1866 Memory
Storage: Seagate 600 Series 240GB 2.5" Solid State Disk
Storage: Western Digital Caviar Black 1TB 3.5" 7200RPM Internal Hard Drive
Video Card: Asus Radeon R9 280X 3GB Video Card
Case: Cooler Master HAF 922 ATX Mid Tower Case
Power Supply: Corsair RM 750W 80+ Gold Certified Fully-Modular ATX Power Supply

Currently im @ 4.5 GHz w/ 1.25v
i5 4670k w/ Hyper 212 evo with the coolermaster included thermal paste
My questions:

1.) When I run Prime 95 my temps for all my cores don't exceed 80c
Core 1: 40-80C
Core 2: 39-75C
Core 3: 37-74C
Core 4: 35-66C

Now when I run intel burn test for 10 runs on normal settings

Core 1: 40-89C
Core 2: 37-87C
Core 3: 36-82C
Core 4: 35-77C

Now I heard to not go over 1.3v on haswell so I kept it at 1.25 and people say to not go over 80c.

So I stays at 80c and below during prime 95 but exceeds it on IBT.

So are these good temps or not?

2.) I want this pc to last me at the most 5 yrs before upgrading. How bad is it to keep the cpu at this speed and would you recommend a lower speed/voltage for my build. I know overclocking makes pc parts last less time so I dont want to screw anything up since its my first time.

3.) I only ran IBT for 10 runs and Prime 95 for about an hour. No blue screens nothing, so would it be fair to say its stable and good temps or should I run it for a bit long maybe over night.

Thanks for helping a newbie understand!

JackNaylorPE · Dec 22, 2013

First off you shud view the OC tutorials on youtube that JJ did for the Z87 asus boards....

Id drop prime95 from ta test list and use RoG Bench

Pinhedd · Dec 22, 2013

genardas :

You've hit most of the important parts on your own, so bravo on that.

IBT is a good tool to push your CPU to the absolute thermal limit. It's based on Linpack which is a benchmark that tests a CPU's floating point capabilities. Integer, cache and memory subsystems are tested only as a byproduct. If your CPU can handle IBT within acceptable temperatures, it can handle any application within acceptable temperatures.

Prime95 is a good test for longer term stability. It will pass through a variety of modes of operation which test floating point, integer, cache and memory to varying degrees. Your PC will not run as hot as it would with IBT but it will get a more even workout.

My recommendation is 10 passes of IBT on Very High (not maximum) to test for thermal viability and short term stability, followed by 12 hours of Prime95.

genardas · Dec 22, 2013

JackNaylorPE :

Watching them as we speak. Also im guessing RoG Bench is another stress testing program. But why is it so big? Is it cause its for videocards too so it has graphics on there?

genardas · Dec 22, 2013

JackNaylorPE :

Pinhedd :

I just re-ran IBT on very high and got bigger temps

Core 1: 39-92C
Core 2: 37-92C
Core 3: 35-86C
Core 4; 34-80C

I will run prime 95 overnight, but from the looks of it I feel like I shouldnt be overclocking this high since its going over 80C? Or it this alright because its IBT and its not a real life scenario? Also I get way lower temps on prime95 so this is sort of confusing me on which one I should base safe temps off of.

Also when I ran these with the cpu core voltage set to adaptive mode instead of manual, with a core voltage of 1.040v (which I cant change) and a Additional turbo mod CPU core voltage of 1.25 it would crash.

Pinhedd · Dec 22, 2013

genardas :

IBT normally results in temperatures that are about 10-20 degrees above those resulting from Prime95. It's normal; it's even mentioned in the IBT README.

That being said, 92 degrees is a little bit on the hot side, and while 4.5Ghz is impressive it's probably a bit much for a Hyper 212 EVO. I run my 3960x at 4.5Ghz but I have a $500 Swiftech liquid cooling loop keeping it running nice.

If I were you I'd make sure that it does not exceed 85 degrees under IBT conditions. Run IBT at stock (no overclocking), and note down the results. Inch it upwards in increments of 100Mhz until it hits 85, then note down those settings. Repeat the test with those settings a couple of times to obtain a good spectrum of sample values, and keep playing with it until you're comfortable with the results. Then run Prime95 overnight to keep it stable. Stable Prime95 temperatures of around 75 degrees are desirable.

The risk with high temperatures is that while the machine may be stable, high temperatures accelerate the physical degradation of the electrical interconnects on the microprocessor. This cuts into the lifespan of the CPU. At first it will gradually become unstable, requiring more core voltage to meet the same performance; later, eventually it will die completely. This happens to all CPUs, but overclocking and high temperatures cause it to happen faster. Most enthusiasts replace their CPUs long before they fall victim to degradation, but it's a good thing to keep in mind as each CPU is unique. I've only ever seen a few CPUs die early deaths as a result of overclocking, but it does happen.

As for IBT being a real life scenario, that's a yes and no answer. The LINPACK functions used to stress your PC are the same functions used by supercomputers to perform their regular tasks. It's designed to simulate the kind of stresses that designers can place upon your PC. I've recently wondered how many of the BF4 crashes experienced by gamers are the result of poor overclocking rather than a downright buggy game. Just some food for thought.

genardas · Dec 22, 2013

Pinhedd :

genardas :

IBT normally results in temperatures that are about 10-20 degrees above those resulting from Prime95. It's normal; it's even mentioned in the IBT README.

That being said, 92 degrees is a little bit on the hot side, and while 4.5Ghz is impressive it's probably a bit much for a Hyper 212 EVO. I run my 3960x at 4.5Ghz but I have a $500 Swiftech liquid cooling loop keeping it running nice.

If I were you I'd make sure that it does not exceed 85 degrees under IBT conditions. Run IBT at stock (no overclocking), and note down the results. Inch it upwards in increments of 100Mhz until it hits 85, then note down those settings. Repeat the test with those settings a couple of times to obtain a good spectrum of sample values, and keep playing with it until you're comfortable with the results. Then run Prime95 overnight to keep it stable. Stable Prime95 temperatures of around 75 degrees are desirable.

The risk with high temperatures is that while the machine may be stable, high temperatures accelerate the physical degradation of the electrical interconnects on the microprocessor. This cuts into the lifespan of the CPU. At first it will gradually become unstable, requiring more core voltage to meet the same performance; later, eventually it will die completely. This happens to all CPUs, but overclocking and high temperatures cause it to happen faster. Most enthusiasts replace their CPUs long before they fall victim to degradation, but it's a good thing to keep in mind as each CPU is unique. I've only ever seen a few CPUs die early deaths as a result of overclocking, but it does happen.

As for IBT being a real life scenario, that's a yes and no answer. The LINPACK functions used to stress your PC are the same functions used by supercomputers to perform their regular tasks. It's designed to simulate the kind of stresses that designers can place upon your PC. I've recently wondered how many of the BF4 crashes experienced by gamers are the result of poor overclocking rather than a downright buggy game. Just some food for thought.

I guess I will lower the clock speed like you said. On problem though when I ran these tests the first time with the cpu core voltage set to manual instead of adaptive mode. I now set it to adaptive mode, with a core voltage of 1.040v (which I cant change) and a Additional Turbo Mode CPU core voltage of 1.25 it just blue screens when I start the test.

That jj guy in the video said to set it to manual mode when doing test like prime 95 and IBT because if you use adaptive mode then it will actual draw more voltage on those tests.

Pinhedd · Dec 23, 2013

genardas :

Pinhedd :

genardas :

IBT normally results in temperatures that are about 10-20 degrees above those resulting from Prime95. It's normal; it's even mentioned in the IBT README.

That being said, 92 degrees is a little bit on the hot side, and while 4.5Ghz is impressive it's probably a bit much for a Hyper 212 EVO. I run my 3960x at 4.5Ghz but I have a $500 Swiftech liquid cooling loop keeping it running nice.

If I were you I'd make sure that it does not exceed 85 degrees under IBT conditions. Run IBT at stock (no overclocking), and note down the results. Inch it upwards in increments of 100Mhz until it hits 85, then note down those settings. Repeat the test with those settings a couple of times to obtain a good spectrum of sample values, and keep playing with it until you're comfortable with the results. Then run Prime95 overnight to keep it stable. Stable Prime95 temperatures of around 75 degrees are desirable.

The risk with high temperatures is that while the machine may be stable, high temperatures accelerate the physical degradation of the electrical interconnects on the microprocessor. This cuts into the lifespan of the CPU. At first it will gradually become unstable, requiring more core voltage to meet the same performance; later, eventually it will die completely. This happens to all CPUs, but overclocking and high temperatures cause it to happen faster. Most enthusiasts replace their CPUs long before they fall victim to degradation, but it's a good thing to keep in mind as each CPU is unique. I've only ever seen a few CPUs die early deaths as a result of overclocking, but it does happen.

As for IBT being a real life scenario, that's a yes and no answer. The LINPACK functions used to stress your PC are the same functions used by supercomputers to perform their regular tasks. It's designed to simulate the kind of stresses that designers can place upon your PC. I've recently wondered how many of the BF4 crashes experienced by gamers are the result of poor overclocking rather than a downright buggy game. Just some food for thought.

I guess I will lower the clock speed like you said. On problem though when I ran these tests the first time with the cpu core voltage set to manual instead of adaptive mode. I now set it to adaptive mode, with a core voltage of 1.040v (which I cant change) and a Additional Turbo Mode CPU core voltage of 1.25 it just blue screens when I start the test.

That jj guy in the video said to set it to manual mode when doing test like prime 95 and IBT because if you use adaptive mode then it will actual draw more voltage on those tests.

Any sort of automatic or dynamic overclocking should be avoided like the plague for exactly the reason that you just described, they're horrible. Set it to manual and leave it there.

genardas · Dec 23, 2013

Pinhedd :

genardas :

Pinhedd :

genardas :

IBT normally results in temperatures that are about 10-20 degrees above those resulting from Prime95. It's normal; it's even mentioned in the IBT README.

That being said, 92 degrees is a little bit on the hot side, and while 4.5Ghz is impressive it's probably a bit much for a Hyper 212 EVO. I run my 3960x at 4.5Ghz but I have a $500 Swiftech liquid cooling loop keeping it running nice.

If I were you I'd make sure that it does not exceed 85 degrees under IBT conditions. Run IBT at stock (no overclocking), and note down the results. Inch it upwards in increments of 100Mhz until it hits 85, then note down those settings. Repeat the test with those settings a couple of times to obtain a good spectrum of sample values, and keep playing with it until you're comfortable with the results. Then run Prime95 overnight to keep it stable. Stable Prime95 temperatures of around 75 degrees are desirable.

The risk with high temperatures is that while the machine may be stable, high temperatures accelerate the physical degradation of the electrical interconnects on the microprocessor. This cuts into the lifespan of the CPU. At first it will gradually become unstable, requiring more core voltage to meet the same performance; later, eventually it will die completely. This happens to all CPUs, but overclocking and high temperatures cause it to happen faster. Most enthusiasts replace their CPUs long before they fall victim to degradation, but it's a good thing to keep in mind as each CPU is unique. I've only ever seen a few CPUs die early deaths as a result of overclocking, but it does happen.

As for IBT being a real life scenario, that's a yes and no answer. The LINPACK functions used to stress your PC are the same functions used by supercomputers to perform their regular tasks. It's designed to simulate the kind of stresses that designers can place upon your PC. I've recently wondered how many of the BF4 crashes experienced by gamers are the result of poor overclocking rather than a downright buggy game. Just some food for thought.

I guess I will lower the clock speed like you said. On problem though when I ran these tests the first time with the cpu core voltage set to manual instead of adaptive mode. I now set it to adaptive mode, with a core voltage of 1.040v (which I cant change) and a Additional Turbo Mode CPU core voltage of 1.25 it just blue screens when I start the test.

That jj guy in the video said to set it to manual mode when doing test like prime 95 and IBT because if you use adaptive mode then it will actual draw more voltage on those tests.

Any sort of automatic or dynamic overclocking should be avoided like the plague for exactly the reason that you just described, they're horrible. Set it to manual and leave it there.

What I was referring to was the voltage clock. If it is set to manual then it always runs @ 1.25ghz but if set to adaptive mode @ 1.25ghz it will only got to 1.25ghz when needed and go down in voltage when needed also. At least that's how it was explained jin the video.

Pinhedd · Dec 23, 2013

genardas :

Pinhedd :

genardas :

Pinhedd :

genardas :

IBT normally results in temperatures that are about 10-20 degrees above those resulting from Prime95. It's normal; it's even mentioned in the IBT README.

That being said, 92 degrees is a little bit on the hot side, and while 4.5Ghz is impressive it's probably a bit much for a Hyper 212 EVO. I run my 3960x at 4.5Ghz but I have a $500 Swiftech liquid cooling loop keeping it running nice.

If I were you I'd make sure that it does not exceed 85 degrees under IBT conditions. Run IBT at stock (no overclocking), and note down the results. Inch it upwards in increments of 100Mhz until it hits 85, then note down those settings. Repeat the test with those settings a couple of times to obtain a good spectrum of sample values, and keep playing with it until you're comfortable with the results. Then run Prime95 overnight to keep it stable. Stable Prime95 temperatures of around 75 degrees are desirable.

The risk with high temperatures is that while the machine may be stable, high temperatures accelerate the physical degradation of the electrical interconnects on the microprocessor. This cuts into the lifespan of the CPU. At first it will gradually become unstable, requiring more core voltage to meet the same performance; later, eventually it will die completely. This happens to all CPUs, but overclocking and high temperatures cause it to happen faster. Most enthusiasts replace their CPUs long before they fall victim to degradation, but it's a good thing to keep in mind as each CPU is unique. I've only ever seen a few CPUs die early deaths as a result of overclocking, but it does happen.

As for IBT being a real life scenario, that's a yes and no answer. The LINPACK functions used to stress your PC are the same functions used by supercomputers to perform their regular tasks. It's designed to simulate the kind of stresses that designers can place upon your PC. I've recently wondered how many of the BF4 crashes experienced by gamers are the result of poor overclocking rather than a downright buggy game. Just some food for thought.

I guess I will lower the clock speed like you said. On problem though when I ran these tests the first time with the cpu core voltage set to manual instead of adaptive mode. I now set it to adaptive mode, with a core voltage of 1.040v (which I cant change) and a Additional Turbo Mode CPU core voltage of 1.25 it just blue screens when I start the test.

That jj guy in the video said to set it to manual mode when doing test like prime 95 and IBT because if you use adaptive mode then it will actual draw more voltage on those tests.

Any sort of automatic or dynamic overclocking should be avoided like the plague for exactly the reason that you just described, they're horrible. Set it to manual and leave it there.

What I was referring to was the voltage clock. If it is set to manual then it always runs @ 1.25ghz but if set to adaptive mode @ 1.25ghz it will only got to 1.25ghz when needed and go down in voltage when needed also. At least that's how it was explained jin the video.

On my X79 I simply set the VCore to 1.325 volts and the maximum turbo multiplier to 45. Under low-power conditions it will step down to about 0.9 volts and 1.2Ghz before jumping up to 1.325 volts and 4.5Ghz.

Search

First time overclocking, confused on temps and need help!

genardas

Honorable

Pinhedd

JackNaylorPE

Titan

Pinhedd

Champion

genardas

Honorable

genardas

Honorable

Pinhedd

Champion

genardas

Honorable

Pinhedd

Champion

genardas

Honorable

Pinhedd

Champion

TRENDING THREADS

Latest posts

Moderators online

Share this page