Overclocking i5-2500K with Stock HSF

[LOL_Wut_Axel]

I've seen many i7 9xx 45nm CPUs running in the mid 80C with stock cooling and stock speeds. It's still within Intels thermal specifications. The chips still run fine today with zero stability issues. The posting trend I see with LOL_Wut_Axel is that if it's not his way then it's the wrong way.

Alright, go ahead and heat it up to 95C at full load. Let us now how it's handling it in 5 years.

The only two ways to induce CPU degradation is by higher temps and higher voltage. A combination of both is lethal, but higher voltage has a worse effect on degradation than higher temps. Both are still factors to be taken in, and I don't see how anyone could be comfortable running a Core i5 or i7 that's 85-100C at full load for 24/7 use in long periods of time.

 

[arredondo]

I was wondering the same question as the OP.

 

[86waterpumper]

Well my thoughts of o/c any at all on the stock cooler have gone out the window lol. I made sure I mounted the stock heatsink and fan right, on my 2500k but still in the bios I am seeing temps around 60c, which I know the bios puts the cpu partially under a load...In windows, I see temps of around 35 @ idle, but for some reason it's idling at 3300 and not 1600. I'm sure idle temps would be at least down around 30 if it did. Prime 95 ran right up to the low 80s, and I shut it off, so I will be buying a better heatsink fan I will say that I am impressed with how quiet it is, but that's about it. I did use the stock heatsink paste that was already applied, I may try some artic silver, but I doubt it helps alot

 

[notty22]

I was just over-looking Anands original SB launch review and here are his comments.
http://www.anandtech.com/show/4083/...-core-i7-2600k-i5-2500k-core-i3-2100-tested/3

These chips overclock very well. Both my Core i5-2500K and Core i7-2600K hit ~4.4GHz, fully stable, using the stock low-profile cooler.

This is all you need for 4.4GHz​

With a bit more effort and a better cooler, you can get anywhere in the 4.6-5.0GHz range:

 

[86waterpumper]

Hmm yeah different motherboard and cpu maybe who knows. My cores are not far apart from each other, and the system is stable so I don't think i got a bum cpu really. Here in TN, it is 102 degrees today with a heat index of 112. The room my computer is in is at least 85 to 90 degrees so that does not help matters Under general use (installing files, surfing the web etc) I stay right around 50, I will see what it does when I play some games.

 

[LOL_Wut_Axel]

I was just over-looking Anands original SB launch review and here are his comments.
http://www.anandtech.com/show/4083/...-core-i7-2600k-i5-2500k-core-i3-2100-tested/3

Just because you can reach it doesn't mean it's safe for day to day use. That, and there's MANY 2500Ks that can't reach 4.4GHz on stock voltage. You will not want to raise the voltage on the stock heat sink one bit.

 

[arredondo]

According to the article, he didn't merely "reach" that OC speed:

These chips overclock very well. Both my Core i5-2500K and Core i7-2600K hit ~4.4GHz, fully stable, using the stock low-profile cooler.

They titled that page, "Overclocking: Effortless 4.4GHz+ On Air". Looks like it does pretty well with stock cooling.

 

[Markfw]

Just because you can reach it doesn't mean it's safe for day to day use. That, and there's MANY 2500Ks that can't reach 4.4GHz on stock voltage. You will not want to raise the voltage on the stock heat sink one bit.

I have to agree here. I personally would NEVER let one of my boxes (that run 24/7@100%load) run over 70c.......I don't care what anyone says, that's my personal rule.

 

[arredondo]

I must of missed it... the article says the OC is running at 70c with stock parts?

 

[Markfw]

I must of missed it... the article says the OC is running at 70c with stock parts?

There is a trend in this thread, that many says 80-90c is just fine. That is what I am disagreeing with, not that specific article.

I also find it hard to believe that you can get 4.4 at stock vcore.

 

[AdamK47]

There really isn't a hard and fast rule that says what max temp is acceptable for CPUs. Each is different obviously. I wouldn't run my CPU at 90+C for 24/7, but I'm willing to run in the mid to low 80s with it. Take a Pentium 3 from 10 years ago and try to run at these temps and you may end up killing it. Over the years the max TDP has increased and with it the tolerence for higher temps.

 

[RussianSensation]

Oh my word, more temperature paranoia. This is really really starting to bug me. Any temperature is reasonably safe on stock volts.

With a downward blowing heatsink fan (such as the stock fan), the heat from the CPU finds its way straight into your case. The temperatures around VRM on the motherboard and the videocard/hard drives will be negatively affected if your CPU is running at 80*C 24/7. All that heat isn't going to be directed towards a top facing or rear exhaust case fan. Enthusiast videocards with reference designs such as the 570/6970/580 can already get somewhat loud. Some people may prefer great case airflow and as low as possible case temperatures to ensure that these cards are cooled at reasonable noise levels (unless you play with headphones ).

While the CPU itself is probably going to be fine, I can't say the same about your motherboard 3-4 years from now. I have disassembled several systems with stock Intel fans and the VRM underneath the boards created visible brown/burned spots. Of course if you rarely stress your CPU at 100%, then it's not really a problem.

Secondly, running your Intel heatsink at those temps practically guarantees that your fan is spinning at high RPMs, resulting in additional noise.

The heatsink market is the best I have ever seen. You can get a Titan Fenrir for $32, Cooler Master Hyper 212+ for $22 (with 10% off ABCDEFG). I have frequently seen Corsair A70 for < $30 on Amazon.com. Thermalright will be re-releasing the True Spirit for just $29.99 soon. I can't imagine trying to save $30 on an aftermarket heatsink when using a $225 CPU for the following reasons:

1) You will likely be able to carry over the heatsink for another 2-3 builds (I have used Tuniq Tower 120 on E6400, Q6600 --> That's basically a $50 investment over 3 years!)

2) You will have the ability to easily increase performance with overclocking by 30-40% (4.6-4.7ghz 2500k in this case). Even if we consider buying expensive aftermarket coolers, you are still better off. The value proposition is clear here: $225 2500k + $80 heatsink such as the NH-D14 or Silver Arrow + 2500k + overclocking is going to beat a 2600k + stock heatsink + mild overclock. But there isn't a necessity for everyone to buy such expensive coolers either.

3) You have a much quieter system. This is especially valuable if you are running a passively cooled videocard.

 

[arredondo]

The quiet angle convinced me. When I upgrade, I'll get a decent heatsink with the CPU since I plan to OC my next rig's processor.

 

[RussianSensation]

I've seen many i7 9xx 45nm CPUs running in the mid 80C with stock cooling and stock speeds. It's still within Intels thermal specifications. The chips still run fine today with zero stability issues. The posting trend I see with LOL_Wut_Axel is that if it's not his way then it's the wrong way.

Voltage beyond spec is far more dangerous than higher temperatures though. Therefore, Alex was correct. Also, I believe his point was running a system loaded for years at a time, not doing a video encoding task for 6 hours once a week.

Alright, go ahead and heat it up to 95C at full load. Let us now how it's handling it in 5 years.

But let's get back to overclocking. Anyone who claims that CPU temperatures don't matter and that's it's fine to run a processor at 95-99*C isn't understanding thermal throttling. Not to mention your power consumption increases with higher temperatures. At 95*C, a Core i5 / i7 CPU is thermal throttling. In fact, Xbitlabs found that figure at just under 93*C for i7 920 (see graph below).

There is a positive correlation between lower CPU temperatures and the CPU's overclocking headroom frequency. Obviously, eventually the CPU is maxed out. But there is no way a stock cooler will allow similar overclocks at higher temperatures.

Generally, you will get lower overclocking potential AND higher temperatures with an inferior cooler:

Source

If you are using the stock heatsink, you are leaving 200-400+ mhz on the table. Plus, you have to put up with the buzzing noise of the Intel heatsink. Whether or not this matters to a user is a personal preference. But a stock heatsink will never allow similar overclocks.

Iv ran both my 2500k and 2600k on stock cooling with closed case and they hit 83-85c linx and ibt.

For how long? 24 hours?

We can't just look at the 80*C figure in isolation. Consider 2 systems: 1 runs 24/7 100&#37; load at 80*C for 2 years vs. 1 that runs 1-2 hours a day at 100% load at 80*C. In both cases, the CPU will 'survive' but is the motherboard going to (blown, overheated VRMs)?

Also, what happens in 2-3 months when your heatsink is covered in dust? Your temps are in the 90s. My aftermarket one with 65*C piles up dust for 2-3x longer and by the time I get to clean it, it's in the high 70s or 80s for example. Lower noise levels, less frequent dust cleaning, higher overclocks, lower component temps (videocard, hard drives, mobo VRMs) are all worth the $30 aftermarket cooler investment imo.

I have to agree here. I personally would NEVER let one of my boxes (that run 24/7@100%load) run over 70c.......I don't care what anyone says, that's my personal rule.

That's key right there. The load duration at high temps has a strong potential to burn out VRMs and MOSFTETs on motherboards without proper airflow/temperature control.

 

[86waterpumper]

RussianSensation, you seem to know a good bit about heatsinks and fans. I do not want to get another 212, I was not impressed with it at all on my last build. I do like the screw in type retention system rather than clip on, but I'd prefer a smoother base for sure. There is no way you can apply paste in a very neat manner with the rough bottom of the 212. Heck even the stock amd cooler that comes with the phenoms have a nice flat base, and they have heatpipes. Anyway I know that the noctuas are regarded as the best if money is not object, but do you or anyone else know what the best cooler under say 50 bucks would be, that would have the quietest fan and still perform alot better than the stock one?

 

[grkM3]

There is a trend in this thread, that many says 80-90c is just fine. That is what I am disagreeing with, not that specific article.

I also find it hard to believe that you can get 4.4 at stock vcore.

My 2600k did 4.6 on stock vcore but its been threw hell and back and seen well over 1.6 volts for 5.4ghz benching.

From the high vcore it needs more juice now but I just set volts to 1.16 in bios and it loads to 1.25 and its stable at 4.4ghz no problem.

 

[RussianSensation]

I do like the screw in type retention system rather than clip on, but I'd prefer a smoother base for sure .......do you or anyone else know what the best cooler under say 50 bucks would be, that would have the quietest fan and still perform alot better than the stock one?

Assuming you have a decent size case and aren't intimidated by its size, I recommend the Thermalright Macho HR-02 for $39.99.

It has a quiet 140mm fan, a smooth base, and beats Corsair H60, Antec 620 watercoolilng systems!

Here is another review: it's very quiet and its performance is outstanding.

It should be available for sale shortly.

 

[MrTransistorm]

Assuming you have a decent size case and aren't intimidated by its size, I recommend the Thermalright Macho HR-02 for $39.99.

It has a quiet 140mm fan, a smooth base, and beats Corsair H60, Antec 620 watercoolilng systems!

Here is another review: it's very quiet and its performance is outstanding.

It should be available for sale shortly.

What the? Some very minor changes and an added fan for $30 less than the original? Why didn't they do that in the first place? I was robbed! :'(

Oh well, I've moved on to watercooling anyway. My temps are about 15 to 20C lower than they were with my HR-02.

The original HR-02 was a pretty decent cooler for quiet operation. Now that the new model includes a fan for a much lower price, it'll definitely be more competitive.

 

[86waterpumper]

HMmm I have a fractal design mini arc case, and a asrock pro 3m board. I will have to do some measuring to see if it would fit, but it sounds like it will be a great deal of sink for the money!

On a side note, that olive drab color fan is hilarious... Looks like the odd schemes that Noctua uses, not sure that is going to match with any motherboard, however since I don't even
have a windowed case it hardly matters.

 

[RussianSensation]

What the? Some very minor changes and an added fan for $30 less than the original? Why didn't they do that in the first place? I was robbed! :'(

It's unbelievable right. They removed Nickel plating and added the fan and are selling the HR-02 for $39.99. You pretty much get the performance of the Prolimatech Super Mega + fan for $40!

Not trying to turn into a Thermalright fanboy, but it looks like they are on a serious streak to regain market share:

Thermalright True Spirit + fan (originally Thermalright Ultra-120) for $29.99:
http://www.legitreviews.com/article/1657/1/

Not bad.

HMmm I have a fractal design mini arc case, and a asrock pro 3m board. I will have to do some measuring to see if it would fit, but it sounds like it will be a great deal of sink for the money!

http://www.guru3d.com/news/fractal-design-arc-mini-chassis/
* Supports CPU coolers with height of 165mm

You are good.

 

[bryanW1995]

The max you set Turbo to is gonna be for when only one core is in use. Each core you add into the equation, it's one bin lower in clock speed. For example, if you choose 4GHz, it's gonna be 4GHz for one thread, 3.9GHz for two, 3.8GHz for three, and 3.7GHz for four. This is all for the locked CPUs that have Turbo Boost, though. If it's unlocked you just set whatever multiplier you want.

4.2GHz on stock volts is not guaranteed, and could definitely cause instability. Also, 4.2GHz will result in another ~2C increase in temps over 3.8GHz. The most I recommend going on these with stock voltage is 4GHz. Since the stock cooler is pretty crappy, if you want to go over 3.8GHz I recommend a custom heat sink. Heat is one of the worst enemies of CPUs.

Huh? Who cares about another 2c temperature delta? That number sounds extremely low to me, but if it's accurate why wouldn't you try it?

Oh my word, more temperature paranoia. This is really really starting to bug me. Any temperature is reasonably safe on stock volts.

Hell, my furnace of a 920 got to sweat it out with the stock cooler for nearly a year and a half at 4.0Ghz with a slight overvolt of 1.3V. It would deviate between around 2 and 10 degrees from TJMax during this time. Linpack would throttle it almost instantly, but I kept it at 4Ghz because it was the only program that would.

I'm not saying that my case is guaranteed representative of every 920 in existence and that overvolting on the stock cooler is fine, I just have a problem with people making random thresholds of temperature that should never be crossed: 80*C in this thread. Let me ask you where you guys got that number from? Do you work at Intel in the "no one listens to me and I think TJMax should be 80*C" department?

The limit for overclocking on stock voltage has and probably always will be the thermal throttling limit. I'll even extend that to only the programs you use. If Linpack throttles your stock voltage overclock, but isn't a program you would otherwise use, it doesn't matter as long as the programs you need the overclock for don't throttle.

I will say that temperature ends up becoming very important once you start upping the voltage however.

I've gone to ~ 90c (so 10c delta to tjmax) many times on my i7 920 with no thermal throttling at all, and that was with vcore of ~ 1.36 or so after vdroop.

hell you can run it at whatever temp you want to and that's fine by me. I am pretty sure MOST people would consider 80C to be pretty damn hot for a cpu though. and me being one of those many people have no intentions of running it that high. if you have a problem accepting that then I do not care one bit.

Why? Because you read that anything with less than 20c delta to tjmax will somehow permanently damage your cpu? If you're on stock voltages and you consistently run your chip at 90c + you won't experience nearly as much degradation as I do on my i7 920 @ 1.36 vcore at ~ 70c. Voltage bumps can and often do cause serious degradation to cpus, but temperature alone rarely if ever hurts them. Well, unless you're planning to give the cpu to your grandkids I guess...

I opened a 2500k box to see the smallest hsf ever made.
I threw the 2500k in the garbage and used the largest hsf from a 65nm c2d which is twice as tall with a copper base just to run off this 2500k before putting it in a case.
Now about this Tcase and SB 72.6*C: thing when your cores reach higher temps of 73c you will start to get errors sooner or later which in turn limit your oc.
Now the the people here with the highest oc have there temps below 74c for 24-7 use.

That statement is just wrong.

There is a trend in this thread, that many says 80-90c is just fine. That is what I am disagreeing with, not that specific article.

I also find it hard to believe that you can get 4.4 at stock vcore.

Finally somebody with experience chimes in. I disagree with your statement, but I can't argue that you've dealt with a lot more cpus than most of us have. Of course, your place must get REALLY hot already, and you are overvolting as well as overclocking. The current discussion is about a safe temperature with no overvolting done at all.

It's unbelievable right. They removed Nickel plating and added the fan and are selling the HR-02 for $39.99. You pretty much get the performance of the Prolimatech Super Mega + fan for $40!

Not trying to turn into a Thermalright fanboy, but it looks like they are on a serious streak to regain market share:

Thermalright True Spirit + fan (originally Thermalright Ultra-120) for $29.99:
http://www.legitreviews.com/article/1657/1/

Not bad.



http://www.guru3d.com/news/fractal-design-arc-mini-chassis/
* Supports CPU coolers with height of 165mm

You are good.

I love my TRUE, though I had to go to push/pull to get 4.2 at reasonable temps on my i7 920. If they're selling them for 1/3 the price with a fan then that's a big win imho. I live being to manually control the fan speed on my cpu coolers, though. I love that on my tuniq's, and I got those 120mm antec tri-cool fans for the TRUE.

 

[john3850]

RE just wrong
What is wrong with not wanting to use that tiny hsf.
When I oc with a new cpu I will not run above 75c
The asrock mb has 775 and 1155 hole mounts.
Because I only use water I only have left over intel hsf.
I wont put mb in a case untill I get a good oc.
So I use the biggest hsf which is 775 hsf.
I read many posts on different forms were people start getting errors at higher temps.
I set cpu for 5000 run linx 30 sec. latter I get errors.
I got errors because my temps are hitting 80c.
If I keep my temps below 74c I dont get errors.

 

[blackened23]

I think its important to note here that running 70/80+c for extended periods is not good for the cpu in the long run. It might be stable today, but how long will it stay that way? Will it be good a year from now? CPU's dont die overnight. High temperatures/voltages take their toll over time, and running at high temps won't make your cpu die immediately but it will in the long run, shorten the lifespan. There's a misconception that a CPU will die immediately if its overheating- that is not how it happens. Even if you're not getting errors at 75c that doesn't mean that the cpu isn't slowly degrading.

I've had overclocked cpu's crap out 6 months down the road because I was stubborn about getting good cooling setups. Never again. My suggestion is to not use the stock HSF no matter WHAT if you're oc'ing, don't learn the hard way like I did years ago. Personally I think running a 4.4 OC is ludicrous with stock HSF, you're just asking for problems to happen down the road.

 

[Ben90]

Guess I have to quote the forum hero again

Anything below TJmax is fine if you don't mind your CPU's expected lifetime to be "on the other side" of the 3yr warranty period. (meaning 3.1yrs or some such)

The QA engineering that goes into the process development itself is all geared towards ensuring this much just from a minimization of extenuating liabilities standpoint to Intel's books.

There is no single temperature threshold above which your cpu's lifetime suddenly diminishes. It is a continuous function that is dependent on an exponential of the operating temperature thanks to the physics that underlie the Arrhenius equation.

As a rule of thumb, for every 10C higher your operating temps the expected lifetime of your CPU is reduced by 50&#37; (think of half-life). Whatever your expected cpu lifetime is if operating at 50C, call it X years, you can expect that lifetime to be cut in half if you operate your cpu at 60C, so X/2 years, and again cut in half once more if you operate it at 70C, so X/4 years, etc.

That may sound dire but understand the lifetime is engineered into the IC from the "top-down" in terms of the thermal specs. Meaning your thermal spec was set for your chip with the desire to minimize the number of in-field fails that would occur under warranty.

So making the assumption that your CPU has an expected lifespan equal to (really we should assume greater than as Intel would not be silly enough to make the mean of the distribution equal to their warranty period and then have to deal with the entire left-hand side of the distribution failing under warranty) the standard warranty period (3yrs) when operating at TJmax is a reasonable assumption. Then for every 10C below TJmax you operate the chip you should double the expected lifespan.

If TJmax is 90C and you operate at 80C then a very reasonable lower-estimate of your CPU's expected lifespan would be 6yrs (2 x 3yrs warranty period). If you operate at 70C then 2x2x3yrs = 12yrs expected lifespan.

What is the basis for my arguing this? At TI we required our process technology to be developed so as to enable the minimum lifetime requirement of 10yrs operating at max spec'ed operating voltage and max spec'ed operating temps in continuous 24/7 operation. It is SOP for the industry.

Now where you can really cook your goose (cpu) is over-volting and running hot. It doesn't take much to be operating your CPU in a voltage/temperature regime that in combination the two factors contribute to lowering the expected lifespan of your CPU to something <1yr.

Not too mention there is always a distribution to the lifespan and your particular chip could have some intrinsic weakness/flaw in it that puts its expected lifespan at a value below the mean of the distribution and by operating at elevated temps and volts it is destined to fail substantially sooner than the warranty period. (I killed my QX6700 in something like 18 months, never operated above TJmax or above Vccmax, but had lapped the IHS so no warranty replacement for me)

 

[Ben90]

I opened a 2500k box to see the smallest hsf ever made.
I threw the 2500k in the garbage and used the largest hsf from a 65nm c2d which is twice as tall with a copper base just to run off this 2500k before putting it in a case.
Now about this Tcase and SB 72.6*C: thing when your cores reach higher temps of 73c you will start to get errors sooner or later which in turn limit your oc.
Now the the people here with the highest oc have there temps below 74c for 24-7 use.

Tcase != Temperature reported over PECI

If your processor errors from 73*C it is defective and you need to RMA it.