Sign in with
Sign up | Sign in
Your question
Solved

Overclocking a i5?

Tags:
  • Overclocking
  • Z97
  • OC Genie
  • CPUs
  • Intel i5
Last response: in CPUs
Share
September 10, 2014 12:30:44 PM

So I've checked around Tom's for awhile now and I've stumbled upon a few threads about it, not enough to give me some form of basics tho.

Anyhow, I'm to OC a i5-4690K with a Z97-G55 SLI. What should I look for? Just increase the multiplier? Anything else I should check up on? What's the worst that can happen with a failed OC? Perhaps you could point me in the way of a few good tutorial/guides or give me some quick basics.

Tried using OC Genie, made my PC crash twice.

More about : overclocking

a b K Overclocking
a b à CPUs
September 10, 2014 12:36:53 PM

I'm at work at the moment so I can't provide links but look for some video tutorials online.

The worst that can happen? Destroy your CPU.

Don't overclock until you have a strong understanding of what to do and what not to do.
m
0
l
September 10, 2014 12:40:32 PM

TheConsiderateIlliterate said:
I'm at work at the moment so I can't provide links but look for some video tutorials online.

The worst that can happen? Destroy your CPU.

Don't overclock until you have a strong understanding of what to do and what not to do.


Oh thanks, alright! Thought the worst was loss of data, damn. When'll you get home? Going to sleep in an hour or two. So is it dangerous even if I mildly OC? Such as from 3.5 to 3.8 or 4.0?
m
0
l
Related resources
a b K Overclocking
a c 98 à CPUs
September 10, 2014 12:41:14 PM

Plenty of step by steps out there. If you are looking for an exact match to your configuration that isn't likely. Just look up Haswell overclocking.
m
0
l
September 10, 2014 12:47:04 PM

Eximo said:
Plenty of step by steps out there. If you are looking for an exact match to your configuration that isn't likely. Just look up Haswell overclocking.


Thanks, some night litterature to digest.
m
0
l
a b à CPUs
September 10, 2014 2:32:37 PM

SHATORR said:
Oh thanks, alright! Thought the worst was loss of data, damn. When'll you get home? Going to sleep in an hour or two. So is it dangerous even if I mildly OC? Such as from 3.5 to 3.8 or 4.0?


Everything has risks. Some people have been known to UNDERclock CPU's in order to reduce the thermal stresses. I seem to recall seeing a rule of thumb that for every 10C hotter a CPU runs - it shortens the lifespan by half.

On the other hand, there will be a small number of people who have hot running overclocked CPU's that are still going strong, while a few of the underclocked ones have failed. (rare but it does happen)

There is no way to say the CPU you have will run X years at Y GHz (or Y temp). Every CPU is different, even between identical model/batch/revision.

If you read up on overclocking, some people will say they "won" the OC lottery because they got an i5-2500K that runs 4.8GHz with no voltage increase, while others "lost" because the i5-2500K they have won't go over 4.2GHz no matter how much increased voltage they give it.
m
0
l
September 10, 2014 3:32:01 PM

menetlaus said:
SHATORR said:
Oh thanks, alright! Thought the worst was loss of data, damn. When'll you get home? Going to sleep in an hour or two. So is it dangerous even if I mildly OC? Such as from 3.5 to 3.8 or 4.0?


Everything has risks. Some people have been known to UNDERclock CPU's in order to reduce the thermal stresses. I seem to recall seeing a rule of thumb that for every 10C hotter a CPU runs - it shortens the lifespan by half.

On the other hand, there will be a small number of people who have hot running overclocked CPU's that are still going strong, while a few of the underclocked ones have failed. (rare but it does happen)

There is no way to say the CPU you have will run X years at Y GHz (or Y temp). Every CPU is different, even between identical model/batch/revision.

If you read up on overclocking, some people will say they "won" the OC lottery because they got an i5-2500K that runs 4.8GHz with no voltage increase, while others "lost" because the i5-2500K they have won't go over 4.2GHz no matter how much increased voltage they give it.


Oh alright. I don't care fox Max performance but since I'm to add another Radeon R9 280 later on, I'd guess I'd have to OC the CPU to at least 3.8GHz.

So if I've understood the guides correctly, all I actually have to do is set the voltage to 1.2-1.25V (as high as I can without overheating the CPU) then increase the multiplier slowly, seeing how far up it can go without crashing. Am I missing something? Can the CPU get damaged from clock speeds or just from overheating? Why did my OC Genie crash my PC @3.9GHz?

Thanks for taking your time!
m
0
l
a b à CPUs
September 10, 2014 3:47:27 PM

Are you using an aftermarket cooler or the stock intel one?
m
0
l
September 10, 2014 4:29:04 PM

@mlga91, I can't seem to find the reply button on my phone..

Anyway, I'm using the CM Seidon 120M, stock fan but getting a 120mm Aerocool DS for the radiator.
m
0
l
September 11, 2014 12:05:54 PM

Did y'all die?
m
0
l
a b à CPUs
September 11, 2014 12:11:40 PM

Not yet, give me a few more years, what have you done already? did you only used OC genie?
m
0
l
a b à CPUs
September 11, 2014 12:44:02 PM

I just made this, dont be impressed by my mad paint skills:



Basically you start stepping up the frequency, then stress test, check temps, if you have problms in the tests step up a little the voltage and try again, always taking care of the temperature.
m
0
l
September 11, 2014 12:50:56 PM

mlga91 said:
I just made this, dont be impressed by my mad paint skills:



Basically you start stepping up the frequency, then stress test, check temps, if you have problms in the tests step up a little the voltage and try again, always taking care of the temperature.


Alright! And yeah, tried OC Genie, after that I just enabled XMP.

So I'm using a 850W 80+ Gold PSU, is it the PSU or CPU that messes up the voltage part?
m
0
l

Best solution

a b à CPUs
September 11, 2014 12:59:21 PM

You change the voltage on the bios, there should be somewhere something that says Vcore, there you change it, always try in little steps, like 0.05, less if you can and want. Also with the frequency use steps of 100 or 50 MHz (0.1 and 0.05 GHz) and again, less if you want and can.
Share
September 11, 2014 1:02:10 PM

mlga91 said:
You change the voltage on the bios, there should be somewhere something that says Vcore, there you change it, always try in little steps, like 0.05, less if you can and want. Also with the frequency use steps of 100 or 50 MHz (0.1 and 0.05 GHz) and again, less if you want and can.


Fawk, clicked the wrong button, oh well.
So should I set the freq and not the multip? And just increase it by 0,05V each time? Shouldn't I jump to smth directly? Like 1.1, 1.2?
m
0
l
a b à CPUs
September 11, 2014 1:04:55 PM

SHATORR said:
mlga91 said:
You change the voltage on the bios, there should be somewhere something that says Vcore, there you change it, always try in little steps, like 0.05, less if you can and want. Also with the frequency use steps of 100 or 50 MHz (0.1 and 0.05 GHz) and again, less if you want and can.


Fawk, clicked the wrong button, oh well.
So should I set the freq and not the multip? And just increase it by 0,05V each time? Shouldn't I jump to smth directly? Like 1.1, 1.2?


The multiplier changes directly the frequency, and with the increases, is a matter of taste, some people step it up without mercy, and other people make it slow. I prefer the slow and more safe way.

PS You can unselect the best ansewer, there should be a button.

m
0
l
a b K Overclocking
a b à CPUs
September 11, 2014 1:08:02 PM

Want me to unselect the solution? However, the guide he made is pretty accurate. If you got time, go by 0.005v steps. Otherwise 0.01 or even 0.02v steps are fine.

You can also fix your voltage to 1.25 (you should be able to cool that) and from then on increase the multiplicator by +1 steps until you are not stable anymore. Go back to the last step you were stable (around 4.5ghz I think) and then decrease the voltage gradually.
m
0
l
September 11, 2014 1:10:26 PM

mlga91 said:
SHATORR said:
mlga91 said:
You change the voltage on the bios, there should be somewhere something that says Vcore, there you change it, always try in little steps, like 0.05, less if you can and want. Also with the frequency use steps of 100 or 50 MHz (0.1 and 0.05 GHz) and again, less if you want and can.


Fawk, clicked the wrong button, oh well.
So should I set the freq and not the multip? And just increase it by 0,05V each time? Shouldn't I jump to smth directly? Like 1.1, 1.2?


The multiplier changes directly the frequency, and with the increases, is a matter of taste, some people step it up without mercy, and other people make it slow. I prefer the slow and more safe way.

PS You can unselect the best ansewer, there should be a button.



Alright, thanks! So increase voltage, check temps, increase freq/multi, check stability, repeat?
m
0
l
September 11, 2014 1:22:13 PM

@DubbleClick No thanks, he deserves it! So I shouldn't go for more than 1,25V? As I said, PSU is beast.

Oh and does anyone know how to reply on a phone?
m
0
l
a b K Overclocking
a b à CPUs
September 11, 2014 2:25:08 PM

Can't reply from a phone by a button, you'd have to manually alter the adress or enter the quotation manually.

1.3v is the safe maximum for haswell, I'd rather stay below 1.25v, though, especially for you who doesn't have further experience in oc'ing.
m
0
l
a b à CPUs
September 11, 2014 2:43:41 PM

DONT TOUCH the voltage first. Raise the multiplier FIRST then once your computer becomes unstable at a clock THEN and ONLY then begin raising your voltage. Multiplier first as that is what actually changes your speeds, then voltage if you get to a clock speed of say 4.2ghz and it's unstable and causing crashes. Voltage makes a overclock stable but to first overclock you must change the multiplier and do it in small increments.
m
0
l
a b K Overclocking
a b à CPUs
September 11, 2014 3:06:43 PM

azzazel_99 said:
DONT TOUCH the voltage first. Raise the multiplier FIRST then once your computer becomes unstable at a clock THEN and ONLY then begin raising your voltage. Multiplier first as that is what actually changes your speeds, then voltage if you get to a clock speed of say 4.2ghz and it's unstable and causing crashes. Voltage makes a overclock stable but to first overclock you must change the multiplier and do it in small increments.


This does not follow any logic at all.
m
0
l
a b à CPUs
September 11, 2014 3:09:59 PM

The multiplier inside the bios is what changes the actual clock speed of the CPU not the voltage. The voltage allows the CPU clock speed to be increased as it changes with the CPU as (for instance turbo boost kicks in) or manually if you are overclocking yourself without using turbo boost. You don't overclock a processor by adjusting voltage you do it by changing the multiplier and then if and when needed adjusting to voltage up to allow for higher clock speeds via the multiplier.
m
0
l
a b à CPUs
September 11, 2014 3:15:03 PM

If I go into my bios right now and raise my voltage the only thing that's going to go up is my temps. Now if I go into my bios and adjust my multiplier up or down my actual clock speeds of the CPU will change. I'm able to take my i5 2500k to 4.2ghz without any voltage change at all over the stock voltage. Now if I want to go to 4.4 or 4.5 I have to raise my voltage up until I find a stable setting. For me at 4.4ghz I run 1.25v I think it is and that's stable for me in gaming. Anything below 1.25v at 4.4ghz and I crash or lock up randomly. So to clock my CPU to 4.2 ghz all I do is change my multiplier from x33 in the bios to x42 save and exit bios and voila overclocked at 4.2ghz. Again anything above that I also have to play with the voltage until I find the lowest stable voltage I can run. Lower voltage = lower temps.
m
0
l
September 11, 2014 3:19:11 PM

If I've understood it correctly, increasing the frequenzy or multiplier leads to a higher clock speed. A higher clock speed means a larger power requirement. Without sufficient power, the CPU will die down. With too much power/voltage tho, the CPU will overheat. Thus, setting voltage first then finding top multiplier would be a safe bet. After that, one could alter both freq and multi to get a speed closer to the power limits.

Have I understood things correctly? Am I missing something?
m
0
l
a b K Overclocking
a b à CPUs
September 11, 2014 3:25:25 PM

you're absolutely correct. However, twaeking frequency as in blck overclocks all other components as well as your cpu. Some systems do not like an altered frequency. To have an easy time, just mess around with voltage and multiplicator.
m
0
l
a b K Overclocking
a b à CPUs
September 11, 2014 3:30:05 PM

As you just want to get a moderate overclock managed, here's what I would do:

1. Disable turbo boost.
2. Set voltage to 1.2v
3. Set multiplicator to 44 (resulting in 4.4ghz).

If stable: lower voltage in 0.01v steps, until not stable anymore. Add 0.01v and you're good.

If not stable: increase voltage - if you need more than 1.25v (highly unlikely), drop multiplicator to 43 on 1.25v and start dropping voltage as described.
m
0
l
a b à CPUs
September 11, 2014 3:34:30 PM

The only problem with adjusting your voltage first is you don't know what your CPU is capable of. Some cpu's require voltage increases right off the bad where some don't require voltage increases at all until you get into much higher clocks. Like my 2500k for instance I got lucky with being able to hit 4.2ghz without having to do anything to my voltage. Others have had to up voltage at 3.7ghz. The reason so many people adjust only the multiplier first on K series processors is because you might be able to hit the clock speed you want without changing voltage at all. If you jump right into voltage you are taking a guess at what voltage to set at and the. Going back and changing your multiplier. If you do just the multiplier first you find the limits of your CPU without adding voltage. You have no limits technically with your voltage as you can crank it really high however that may damage your CPU but you don't know at what voltage to stop at. With the multiplier you find the limit without damaging anything by overheating from too much voltage and the. Once you have found the highest multiplier setting you can achieve with no voltage adjustments you can start adding voltage from there. I have just never heard of anyone doing voltage first. That's a first even being here on tomshardware I have never seen that recommended.
m
0
l
September 11, 2014 3:41:09 PM

You're basically preaching two different versions.. However, I've read about Click's way before, even if I understand Azzazel's point.

Aaanyway, was reading this guide but it feels overly complicated, am I wrong? http://www.overclock.net/t/1411077/haswell-overclocking...
m
0
l
a b à CPUs
September 11, 2014 4:23:34 PM

If you want to increase the frequency raise the multipler, you only raise the voltage for stability. Some people go straight to 1.25 because is below the maxium safe voltage, then they reach a certain frequency and then they lower the voltage until stability permits it, everyone have their methods and most of them works, but since you've never overclocked before start with the slow and safe method. Also raising the voltage is not much problem for the PSU, the problem is with the cpu, it doesn't like more than 1.3v.
m
0
l
September 12, 2014 1:29:48 AM

Alright, thanks!
m
0
l
September 12, 2014 1:35:27 AM

Last few questions, should I OC now or get a Noctua NH-D15 first? How big's the chance of my CM Seidon 120V breaking? Would warranty cover other hardware possibly affected by the malfunction?

Thanks in advance!
m
0
l
a b à CPUs
September 12, 2014 3:53:28 AM

When you overclock you void any and all warranties on hardware. That being said them being able to prove you were overclocked is a different story but I'm sure they as manufacturers have a way to find out. Anywho anytime you do custom overclocking on PC hardware the manufacturer normally states it voids their warranty. Only company I have ever seen warranty overclockes is evga on their gpu's. Even that wasn't till recently though.
m
0
l
September 12, 2014 4:25:47 AM

azzazel_99 said:
When you overclock you void any and all warranties on hardware. That being said them being able to prove you were overclocked is a different story but I'm sure they as manufacturers have a way to find out. Anywho anytime you do custom overclocking on PC hardware the manufacturer normally states it voids their warranty. Only company I have ever seen warranty overclockes is evga on their gpu's. Even that wasn't till recently though.


I know, was thinking about CPU liquid coolers, if they were to break and take something with it, hah.
m
0
l
a b K Overclocking
a b à CPUs
September 12, 2014 4:29:57 AM

No guarantee in that case, I guess. The chances of that happening should be small, though. Just don't move your pc.
m
0
l
September 12, 2014 4:43:24 AM

DubbleClick said:
No guarantee in that case, I guess. The chances of that happening should be small, though. Just don't move your pc.


Won't. I plugged it to a CPU fan slot tho, should I move it to a case fan? Read somewhere it needs a constant speed, thus a CPU slot would be wasted.
m
0
l
a b K Overclocking
a b à CPUs
September 12, 2014 4:46:25 AM

I'm not experienced with liquid cooling. Staying away from clc's as far as I can, as they offer extremely poor value/noise, which is important to me and didn't yet want to spend the money and thought assembling a custom loop.
m
0
l
September 12, 2014 4:50:42 AM

DubbleClick said:
I'm not experienced with liquid cooling. Staying away from clc's as far as I can, as they offer extremely poor value/noise, which is important to me and didn't yet want to spend the money and thought assembling a custom loop.


Yeah, I've noticed that! That's why I'm getting the Noctua NH-D15.. A shame it's almost half the price of another GPU..
m
0
l
a b à CPUs
September 12, 2014 5:00:28 AM

I know corsair covers your PC hardware if one of their clc's fail and take out a CPU or mobo
m
0
l
September 12, 2014 6:07:53 AM

azzazel_99 said:
I know corsair covers your PC hardware if one of their clc's fail and take out a CPU or mobo


Thanks! Made a thread specifically about this, perhaps someone can understand the terms on their website, haha
m
0
l
!