Overclocking i5-2500K with Stock HSF

[blackened23]

edited, duplicate

 

[blackened23]

Guess I have to quote the forum hero again

Say what you will but I had an 860 die after nearly 1.5 years of continuous operation overclocked, temps were in the okay range. Replacing the cpu confirmed that it was the cpu that had in fact died.

Hey if you want to risk it more power to you. I still maintain oc'ing to 4.4 with stock HSF is beyond stupid, especially if you're running at 70c. lmao. When you can buy an H50 which is quiet and cheap, why would you risk coming to a dead computer a few months down the road? Again, beyond stupid.

 

[jvroig]

Say what you will but I had an 860 die after nearly 1.5 years of continuous operation overclocked, temps were in the okay range. Replacing the cpu confirmed that it was the cpu that had in fact died.

Hey if you want to risk it more power to you. I still maintain oc'ing to 4.4 with stock HSF is beyond stupid, especially if you're running at 70c. lmao. When you can buy an H50 which is quiet and cheap, why would you risk coming to a dead computer a few months down the road? Again, beyond stupid.

The problem with your argument is that it is your opinion backed up by a personal anecdote, whereas the argument you are rebutting is backed up by science and industry facts.

No matter how many "lmao" you add to your post, it still won't make it any more valid than it already isn't.

Take that 860 of yours that died. You say temps were in the ok range, and that it is overclocked. If temps were already "ok", then why do you think a better cooler would have made it not die? Would it not be more logical for you to assume that maybe it was the voltage that killed it, OR it was just an unlucky chip (this happens with electronics all the time, it's the nature of the beast, that's why we have RMA policies for everything)? There could be many reasons for a chip to die, but you seem to think it can only be temperature, and thus would be completely averted by a better cooler.

Also, since this is the internet and everybody is just some "random joe" as far as anybody else is concerned, if I were a complete noob and have no clue about the issue yet, I would certainly put more trust in Anand, a trusted and well-respected international reviewer of hardware (and who is under scrutiny by more than just a handful of forum peers), than in some other random internet guy, whether their post count is 1 or 100,000.

Not saying you shouldn't express your opinion of course You have every right to do so, and are very much welcome to. I'm just expressing the facts of the case right there. Your opinion has been read, assessed, and very well noted. I would not find it surprising, though, if it will be rejected in favor of more well-founded opinion, or more credible ones.

 

[blackened23]

The problem with your argument is that it is your opinion backed up by a personal anecdote, whereas the argument you are rebutting is backed up by science and industry facts.

No matter how many "lmao" you add to your post, it still won't make it any more valid than it already isn't.

Take that 860 of yours that died. You say temps were in the ok range, and that it is overclocked. If temps were already "ok", then why do you think a better cooler would have made it not die? Would it not be more logical for you to assume that maybe it was the voltage that killed it, OR it was just an unlucky chip (this happens with electronics all the time, it's the nature of the beast, that's why we have RMA policies for everything)? There could be many reasons for a chip to die, but you seem to think it can only be temperature, and thus would be completely averted by a better cooler.

Also, since this is the internet and everybody is just some "random joe" as far as anybody else is concerned, if I were a complete noob and have no clue about the issue yet, I would certainly put more trust in Anand, a trusted and well-respected international reviewer of hardware (and who is under scrutiny by more than just a handful of forum peers), than in some other random internet guy, whether their post count is 1 or 100,000.

Not saying you shouldn't express your opinion of course You have every right to do so, and are very much welcome to. I'm just expressing the facts of the case right there. Your opinion has been read, assessed, and very well noted. I would not find it surprising, though, if it will be rejected in favor of more well-founded opinion, or more credible ones.

So you put more trust in anand? Well, "lmao", I wasn't aware that anand ever reviewed a recent overclocked CPU with stock HSF, or that he advocated overclocking with stock HSF. In fact every overclocked cpu he has ever tested has not ever had stock HSF recently, to my knowledge. Every review i've seen of sandy bridge processors here has had liquid cooling IIRC.

Go ahead and overclock your cpu with stock hsf, whatever floats your boat. I wasn't aware that taking appropriate safeguards and spending at the most 50-70$ as a preventive safeguard was such a hassle.

 

[john3850]

Tcase != Temperature reported over PECI

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

tcase can and is off in many cases by 6c.many related posts.
inside the intel sb pdf it states that damage can take place when any of the core go above the the tcase temp.
The closer my chip gets to 78-80c @ 4800-5000mhz the quicker the errors happen in prime linX.
The room temp were I live changes now +5c is not a big change but With high oc it matters.
My cpu at 100% load runs good in the low 70-74c now if add the +5c my temp is now over my temp limit.

 

[jvroig]

So you put more trust in anand? Well, "lmao", I wasn't aware that anand ever reviewed a recent overclocked CPU with stock HSF, or that he advocated overclocking with stock HSF. In fact every overclocked cpu he has ever tested has not ever had stock HSF recently, to my knowledge. Every review i've seen of sandy bridge processors here has had liquid cooling IIRC.

If you missed it, it has been mentioned three times in this thread, and the first mention was by myself:

http://forums.anandtech.com/showpost.php?p=32071939&postcount=24

http://forums.anandtech.com/showpost.php?p=32078418&postcount=29

http://forums.anandtech.com/showpost.php?p=32078711&postcount=32

From the article itself:

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

What will you "lmao" now?

 

[blackened23]

lmao indeed (rolleyes). How long did he test with stock HSF? Whatever, it doesn't matter. If you feel like 50$ is too much hassle to protect your investment (and again, CPU degradation doesn't happen overnight) more power to you. I feel like 50$ to protect an investment is an easy choice. With all the reports here of stock HSF running 70c+ at load? Nah, I'll pass on that.

 

[toyota]

even 20 bucks will get you a massively better cooler that will allow you to run 20 C lower temps over stock cooler.

 

[jvroig]

lmao indeed (rolleyes). How long did he test with stock HSF? Whatever, it doesn't matter. If you feel like 50$ is too much hassle to protect your investment (and again, CPU degradation doesn't happen overnight) more power to you. I feel like 50$ to protect an investment is an easy choice. With all the reports here of stock HSF running 70c+ at load? Nah, I'll pass on that.

FWIW, we have an emoticon for that, you don't actually have to write "rolleyes"

The OP asked how high can he go using stock HSF. Citing Anand, the answer is "you may be able to reach 4.4Ghz". Of course, overclocking is never guaranteed, so for all we know all he can reach is 3.7 and that still wouldn't make his chip defective.

It's as simple as that. It's not about whether $50 is too expensive or not. The question is just what the OP asked - how high using stock HSF is safe? There are many answers in this thread, but nothing really as impressive as what Anand and his team published: effortless 4.4GHz, using the stock cooler. (Again, no OC is guaranteed, so this is more a guideline than a rule)

I'm sure there are people who were impressed with your answers. I'm sure they went "Whoa, this dude blackened23 is so awesome, he's just 'lmao'-ing the competition, clearly proving he is more credible than Anand or science or industry insiders! Oh, joy!"

But until then, I would stick with Anand and science.

You can keep repeating your answer over and over, but unless you actually can prove you are more credible than Anand, or that we should believe you more than what the fab and CMOS process engineers say about industry methodology for determining warranty, chip longevity and operating temps, you won't really become more credible than you already aren't.

In other words, science > lmao.

even 20 bucks will get you a massively better cooler that will allow you to run 20 C lower temps over stock cooler.

Yep, my hyper 212+ is more or less that amount when I got it. I like it a lot. This is good advice to give to the OP. It's just that his original question was about the stock HSF. I gave him the answer. It's still true that a better HSF would be good advice, but that doesn't change the validity of the answer to his original question.

 

[86waterpumper]

My question is though, where did Anand get his cpu? If intel sent it to him to test then that is a whole other kettle of fish isn't it It could have very well been cherry picked at least to some extent. Obviously every chip is different, some will run hotter, some will o/c better it's all luck of the draw.

In the end though, with all things considered these things are pretty fast at stock speeds without overclocking at all. In fact, I don't even feel compelled to mess with it for awhile like I usually do right off the bat but alot of that is probably due to the ssd...

 

[RussianSensation]

My question is though, where did Anand get his cpu? If intel sent it to him to test then that is a whole other kettle of fish isn't it It could have very well been cherry picked at least to some extent. Obviously every chip is different, some will run hotter, some will o/c better it's all luck of the draw.

I remember reading this exact review, thinking I should be able to do around 3.4-3.5ghz on my i7 860 on the stock heatsink as did Anand. When I purchased my i7, I put AS5 on the heatsink and got this at stock speeds:

Took the heatsink off, reseated it the 2nd time, this time with Tuniq TX-2 thermal paste. No difference.

Went out, got the Prolimatech Megahalems. Got this:

Then proceeded to overclock the i7 to 4.0ghz and still got lower temperatures vs. with the stock Intel heatsink at stock speeds:

Conclusion: Intel stock heatsink sucks for overclocking. Does it work? Sure. But you are leaving a ton of overclocking headroom on the table for what, a savings of $30-40 when you just spent $500+ on new parts?

 

[monkeh624]

I've got a 2500K and Asus P8P67Pro Mb. I did a 'lazy overclock' using the mobos built in tweaker.
This set my CPU to 4357Mhz, using the stock HSF. Its been at that speed for around 2 months now. No problems so far. Makes me wonder how far I could push it with a decent HSF.
I'm quite lazy when it comes to these things though, atm I'm quite happy with its current speed

 

[toyota]

I've got a 2500K and Asus P8P67Pro Mb. I did a 'lazy overclock' using the mobos built in tweaker.
This set my CPU to 4357Mhz, using the stock HSF. Its been at that speed for around 2 months now. No problems so far. Makes me wonder how far I could push it with a decent HSF.
I'm quite lazy when it comes to these things though, atm I'm quite happy with its current speed

just because you have "No problems so far" does not mean you are not potentially doing harm. the fact that you did not even mention temps make me wonder if you have even checked them.

 

[Crap Daddy]

I've got a 2500K and Asus P8P67Pro Mb. I did a 'lazy overclock' using the mobos built in tweaker.
This set my CPU to 4357Mhz, using the stock HSF. Its been at that speed for around 2 months now. No problems so far. Makes me wonder how far I could push it with a decent HSF.
I'm quite lazy when it comes to these things though, atm I'm quite happy with its current speed

You probably have a Vcore just under 1.4V and your temps while gaming are in the 70s. If you ever do a handbrake encode - which is an application and not a torture test- your CPU will get over 90C. If you feel confortable with this then it's your choice.

 

[monkeh624]

just because you have "No problems so far" does not mean you are not potentially doing harm. the fact that you did not even mention temps make me wonder if you have even checked them.

You're right, I had neglected to check the temps. Did a quick test just now, fired up crysis2 and played for a bit, then checked the temps. One core was sitting at 71C, which is too hot for my liking.
Hardware monitor reports the vcore at 1.24V max. I have no idea if this is normal, I didnt touch the volts, but maybe the asus tool did.
So it looks like a new HSF would be a good idea.

 

[john3850]

Why would anybody give a link to any review site that is using a intel ES cpu.
ES cant be bought in any store and are Propetry of Intel.

If it wasnt a sample there would be noting wrong with the link
I also realized you may have missed ES mark in cpu-z
http://www.anandtech.com/show/4083/t...-2100-tested/3
__________________

 

[sep]

See what you started timoseewho ; )

 

[grkM3]

Since Im switching motherboards I gave my golden 2600k a shot with stock heatsink.I am stressing this cpu to the absolute max and finally got it to throttle and it does it at 98c.over clock went from 4.8ghz to 4.6 for a few seconds then back to 4.8 and would throttle back down again once it hit 98c.

I want to also point out that I am pushing 4 sticks of 2133 ram and running the memory voltage at 1.67 so its maxing out the memory also.

this chip does 5.6ghz on water and did 5.2 100% stable with a ninja heatsink.If the system didnt shut down and it only throttles 3x multies down and still pushed almost 4.6ghz at 98c I think we are fine at 80c for short prime test.Honeslty tho,my cpu would never hit these temps on regular daily usage.

I was running intel burn test and running super pi 32mil at the same time to stress both the cpu and memory controller.

here is a pic

PS,this is also with HT on and PLL overvoltage disabled.I can prolly get 5.0 ghz intel burn test stable with stock heat sink and HT off and PLL on

 

[toyota]

whats the reason in using 46 x 104.4 instead of 48 x 100 to get to 4.8?

 

[grkM3]

whats the reason in using 46 x 104.4 instead of 48 x 100 to get to 4.8?

I was overclocking my ram to test my sticks.I run 47x103 for 24/7 use

 

[RussianSensation]

That's a nice 2600k chip you got

But what do you think is going to happen if you run that processor at 98*C for 24/7 for say 3-6 months? Just think about it: You just spent $320 on the CPU + prob $150 on the mobo. That's $470 at least? And you won't buy a $28 heatsink?

And I bet your heatsink fan sounds like a rocket ship? If you keep that CPU for 2-3 years, what's another $30-50 on a good heatsink? To me that guarantees no heat issues, no failed CPU/motherboard/burned VRMs & mosfets and ability to run my CPU at 100% load 24/7 for 3 years with minimal noise. That's priceless.

Honestly, if your board goes, that's 2-3 weeks of downtime. So much for saving $30-50. It will bite you in the arse

I should also mention with top brands like Noctua, Thermalright, Scythe, you can always buy a $10 bracket and reuse that awesome heatsink for another 2-3 future builds on future sockets. So the long-term value is certainly there.

Maybe I am biased though, because I had an overclocked Athlon XP1600+ and Asus mobo died on me due to blown/overheated VRM/Mosfets. Ever since I have always invested in a good aftermarket cooler.

 

[blackened23]

Since Im switching motherboards I gave my golden 2600k a shot with stock heatsink.I am stressing this cpu to the absolute max and finally got it to throttle and it does it at 98c.over clock went from 4.8ghz to 4.6 for a few seconds then back to 4.8 and would throttle back down again once it hit 98c.

I want to also point out that I am pushing 4 sticks of 2133 ram and running the memory voltage at 1.67 so its maxing out the memory also.

this chip does 5.6ghz on water and did 5.2 100% stable with a ninja heatsink.If the system didnt shut down and it only throttles 3x multies down and still pushed almost 4.6ghz at 98c I think we are fine at 80c for short prime test.Honeslty tho,my cpu would never hit these temps on regular daily usage.

I was running intel burn test and running super pi 32mil at the same time to stress both the cpu and memory controller.

I think you need to get your head checked because you're insane. And you get ridiculous overclocks not for gaming, not for computer intensive apps, but as a hobby to get good benchmarks to thump your chest to (with stock HSF at 90+c) Sorry for the tone, but I really do think you're crazy

 

[grkM3]

That's a nice 2600k chip you got

But what do you think is going to happen if you run that processor at 98*C for 24/7 for say 3-6 months? Just think about it: You just spent $320 on the CPU + prob $150 on the mobo. That's $470 at least? And you won't buy a $28 heatsink?

And I bet your heatsink fan sounds like a rocket ship? If you keep that CPU for 2-3 years, what's another $30-50 on a good heatsink? To me that guarantees no heat issues, no failed CPU/motherboard/burned VRMs & mosfets and ability to run my CPU at 100% load 24/7 for 3 years with minimal noise. That's priceless.

Honestly, if your board goes, that's 2-3 weeks of downtime. So much for saving $30-50. It will bite you in the arse

actually the board was 350 lol and I only did this to see with the stock heat sink.

I will never hit these temps on daily tasks,actually right now the chip is at 35c and will maybe hit 70c tops running games or doing anything other than linpack that uses 100000% of the cpu.

Im installing this chip on my ud7 board that is coming in the mail this week with a new water block.I am going to try gigabyte next to see if there 24 phase vrms will clock this chip a little higher.

these new boards run so cold its not even funny,the ud7 board has 2 cpu drivers that it switches back and forth to let the other cool down.My board is cool to the touch and the ram is cool also.

 

[grkM3]

I think you need to get your head checked because you're insane. And you get ridiculous overclocks not for gaming, not for computer intensive apps, but as a hobby to get good benchmarks to thump your chest to (with stock HSF at 90+c) Sorry for the tone, but I really do think you're crazy

a stock clocked 2500 will hit 90c running linx

you think a 4.8ghz overclock is insane.man Im going back to extreme systems,you guys are kids lol.

Ohh look at me thumping my chest!!!! if I wanted to brag I would post up 5.6ghz 3dmark runs

 

[blackened23]

guess I misread, i coulda swore I saw 5ghz+ earlier. My bad!