Is it really worth OC'ing the i5 2500k?

Majcric

Golden Member
May 3, 2011
1,377
40
91
Lately, I've been thinking of buying a cooler and slapping a 4.5 or so OC on my i5. I definitely expect to see performance gains in benchmark runs and a select few of CPU limited games. And just games in general.

Then I stumbled accross this test: http://www.ocaholic.ch/modules/smartsection/item.php?itemid=1129&page=4

At 1080p there are no gains with the exception of skyrim. I play at 1440p and I expect gains to be even less.

So let me get some opinions from members here. Is the above test skewed or would the OC definitely help mins in games.?
 

Yuriman

Diamond Member
Jun 25, 2004
5,530
141
106
BF3/4 will not see improvement in single-player but should scale close to linearly in multiplayer. Some games are not CPU limited at all, others (especially multiplayer/MMO/strategy) will see improvements.
 

nwo

Platinum Member
Jun 21, 2005
2,309
0
71
Most of those benchmarks are single player or 0 player mode... So it's not really stressing the CPU much. As Yuriman said, your CPU gets stressed in multiplayer situations especially those with lots of action in small areas.

I would definitely recommend overclocking an i5 2500k if you play multiplayer games and are looking for a few extra fps. However, if you're only playing single player mode games and/or are happy with your current multiplayer fps, I see no reason to OC.
 

Dave3000

Golden Member
Jan 10, 2011
1,375
91
91
No I don't think it's worth overclocking a 2500k or any CPU for that matter. It can cause system instabilities, data corruption, risk damaging, and reducing the lifespan of you CPU. Overclocking also adds another element to troubleshooting computer problems. Also it takes too much testing and time to test for overclocks that might or might not be stable. In my opinion if you want more CPU speed even if it's not much faster, I suggest just sell your 2500k with the motherboard and upgrade to either a 4670k or 4770k with a new motherboard as Socket 1155 does not support Haswell chips. However if you are going to lose too much money in the process then stick with what you have and maybe upgrade to a faster GPU if you don't have a fast GPU.
 
Last edited:

Techhog

Platinum Member
Sep 11, 2013
2,834
2
26
No I don't think it's worth overclocking a 2500k or any CPU for that matter. It can cause system instabilities, data corruption, risk damaging, and reducing the lifespan of you CPU. Overclocking also adds another element to troubleshooting computer problems. Also it takes too much testing and time to test for overclocks that might or might not be stable. In my opinion if you want more CPU speed even if it's not much faster, I suggest just sell your 2500k with the motherboard and upgrade to either a 4670k or 4770k with a new motherboard as Socket 1155 does not support Haswell chips. However if you are going to lose too much money in the process then stick with what you have and maybe upgrade to a faster GPU if you don't have a fast GPU.

Why are you so strongly against overclocking yet still recommending K-series CPUs. It makes no since, especially since K CPUs are feature gimped and don't off any additional performance.
 

LagunaX

Senior member
Jan 7, 2010
717
0
76
You have an Intel cpu that is 2 generations from the current, in a dead socket 1155.

Buy a recommended $30 air cooler (Hyper 212 evo) and go for an overclock at 1.375v or lower.

Gaming, video encoding, and your e-penis will all improve.

Grab a 240gb ssd on sale for $150 or less while you are at it too.

If you see a deal, move to 8gb too.

Then your system will "fly."
 
Last edited:

Dave3000

Golden Member
Jan 10, 2011
1,375
91
91
Why are you so strongly against overclocking yet still recommending K-series CPUs. It makes no since, especially since K CPUs are feature gimped and don't off any additional performance.

I agree with you that non-k is better if you don't overclock. I forgot about the non-k series. I recommend the non-k series if you don't overclock.
 

Majcric

Golden Member
May 3, 2011
1,377
40
91
No I don't think it's worth overclocking a 2500k or any CPU for that matter. It can cause system instabilities, data corruption, risk damaging, and reducing the lifespan of you CPU. Overclocking also adds another element to troubleshooting computer problems. Also it takes too much testing and time to test for overclocks that might or might not be stable. In my opinion if you want more CPU speed even if it's not much faster, I suggest just sell your 2500k with the motherboard and upgrade to either a 4670k or 4770k with a new motherboard as Socket 1155 does not support Haswell chips. However if you are going to lose too much money in the process then stick with what you have and maybe upgrade to a faster GPU if you don't have a fast GPU.


Overclocking doesn't scare me and if I see gains it's worth it. I'm just trying to weigh out if the gains are worth it on a CPU. Depending upon the game I'm playing I push my GPU to around 1280mhz.
 

Majcric

Golden Member
May 3, 2011
1,377
40
91
You have an Intel cpu that is 2 generations from the current, in a dead socket 1155.

Buy a recommended $30 air cooler (Hyper 212 evo) and go for an overclock at 1.375v or lower.

Gaming, video encoding, and your e-penis will all improve.

Grab a 240gb ssd on sale for $150 or less while you are at it too.

If you see a deal, move to 8gb too.

Then your system will "fly."

I'm not really suffering in performance. The SSD and Ram are already noted for upgrade. I had just been googling and really started wondering if my i5 was worth OC'ing. And considering it is the K I thought I might well make use of the extra performance if the gains are substantial.
 

Majcric

Golden Member
May 3, 2011
1,377
40
91
That's a lower rez though, I knew there were gains below 1080p. I'm mostly concerned about 1440p
 

BonzaiDuck

Lifer
Jun 30, 2004
15,785
1,500
126
That's a lower rez though, I knew there were gains below 1080p. I'm mostly concerned about 1440p

If you have the time or another computer to use for your regular business over several hours and days here and there, you have a choice to "invest" in over-clocking. You may, in fact, base your choice on "gaming benefits." It's really your choice, though.

Depending on the motherboard you use for the i5-2500K, you can store BIOS setting "profiles" under different names after you have perfected them. For instance, you might have a profile for "stock" voltage and speed settings -- tweaked according to your hardware configuration, choices for RAM and other things. So if something goes "wrong" while you attempt to clock at higher speeds, you can always load this profile to handle "daily business" until you free up the machine again for the intensive testing you apply to validate higher speeds as stable.

You can then accumulate profile settings for a range of speeds. For instance, if there is common wisdom or an intel spec which says "don't run the K chip above [some voltage] setting, but you can achieve a stable overclock that seems to violate those guidelines only slightly, you could run those overclock settings occasionally while using lower (safer) settings for 24/7 operation.

Generally, entering the world of over-clocking is sort of like living in the 19th century town of Deadwood. There may be norms of behavior, but the laws aren't enforced consistently. Or in overclocking, Intel isn't "telling you what to do." Intel publishes specs on its processors -- and it used to publish a dual voltage spec -- a "safe" range and an "operable" range around the safe range. With Sandy Bridge, there is now mostly just "consensus" based on the lithography of the processor and other Intel specs.

If someone cautions against overclocking, the underlying premise of it is that they don't know -- whether you know . . . . what you're doing. You may know . . . you may be confident . . . you may not know, or you only just think you know. The remedy there is to consult several OC'ing guides for the chips you're working with.

So -- sure -- overclocking could lead to "instability" and "data loss" or other horrors. I can only say I've been doing it since 2003. Never lost any data. Never corrupted a hard disk or OS installation; never damaged a processor, although I found a couple models of RAM which had been promoted with overconfident performance prospects, and I had to replace them. I had maybe one motherboard go south after a couple years, but in those days, the mobo makers used parts that weren't up to snuff for overclocking.

My view? If you're going to do it (and in fact -- if you're going to buy "K" processors for that purpose [what other purpose would there be? . . .]) -- take a little time at the beginning to find the right motherboard and RAM when you build the system in the first place. Personally, I wouldn't buy a name-brand OEM computer and over-clock it; I'd think companies like Dell or Gateway would make it more difficult for you. So I can speak to "building" and OC'ing as interrelated choices.
 

hans030390

Diamond Member
Feb 3, 2005
7,326
2
76
You are wasting your time asking us if you should be OCing your 2500K when you instead should be spending this time doing something you should have done long ago...OC it.

You will see benefits in about everything.
 

BonzaiDuck

Lifer
Jun 30, 2004
15,785
1,500
126
You are wasting your time asking us if you should be OCing your 2500K when you instead should be spending this time doing something you should have done long ago...OC it.

You will see benefits in about everything.

I agree there 100%. On the other hand, the OP needs to know there are tedious and routine methods -- the tedium and routine require patience.

There are strategies for taking a "stable voltage" to a level that assures optimum processor performance. [See for example IDontCare's dissertation on LinX testing with hyperthreading enabled.]

As we all agree, there are several OC guides for the Sandy cores out there on the web. You will also find, for instance, in a forum like Anandtech's, long-running or enthusiastic threads in which several veterans get together to compare their methods, their hardware, their settings and results.

As to the reliability of these sources . . . . you still are left with your own good judgment. There is no Cardinal Spellman to give us a "nihil obstat" and "imprimatur" as to the reliability of what we read on this topic.
 

OCGuy

Lifer
Jul 12, 2000
27,227
36
91
Yes. If you don't OC you are wasting performance potential.

I wouldn't listen to any anti-OC alarmist, they are never very knowledgeable on the subject.
 

UaVaj

Golden Member
Nov 16, 2012
1,546
0
76
for gaming at 1440p. that single 680 will choke before the stock 2500k is bottleneck.
hence the review shows no improvement with a 7970G at 1080p. 7970G is maxed out. now imagine 1440p with the same gpu.

for other cpu intensive tasks. overclock will improve performance.

overclocking or not. you decide.
 

BonzaiDuck

Lifer
Jun 30, 2004
15,785
1,500
126
Yes. If you don't OC you are wasting performance potential.

I wouldn't listen to any anti-OC alarmist, they are never very knowledgeable on the subject.

Well . . . . I also agree there . . . Figure the Big Boys at NSA probably hire a contractor to build parallel systems for their number-crunching.

On the other hand, I might run my personal business on an OC'd system. I might even feel comfortable using an OC'd system to do the operations-research for building a new Trade Center, or monitoring Mother's life-signs at the hospital. But that's my personal level of comfort, and things can always go wrong. So I'd probably just buy an i7-4960X, top-end motherboard etc. and run stock settings to control my local nuclear power-plant. . . .
 

cytoSiN

Platinum Member
Jul 11, 2002
2,262
7
81
Lately, I've been thinking of buying a cooler and slapping a 4.5 or so OC on my i5. I definitely expect to see performance gains in benchmark runs and a select few of CPU limited games. And just games in general.

Then I stumbled accross this test: http://www.ocaholic.ch/modules/smartsection/item.php?itemid=1129&page=4

At 1080p there are no gains with the exception of skyrim. I play at 1440p and I expect gains to be even less.

So let me get some opinions from members here. Is the above test skewed or would the OC definitely help mins in games.?

Do you only care about gaming applications? For the past two-plus years I've had my 2500k comfortably humming along at 4.5 at standard voltage (I strongly caution against anything higher than 1.5v for that chip), and I find it useful in just about everything I do, including but not limited to re-encoding audio and video, ripping my Blu-ray collection, etc.

If you're air cooling, it helps to have a roomy case and properly arranged fans. But the process has become laughably simple these days now that most manufacturers offer some sort of GUI and mouse function in the BIOS, and Intel has manufactured the chips to auto throttle when necessary. Your Gigabyte should be no exception. You'd almost have to be TRYING to break something.

My suggestion is to set your voltage to 1.5 and slowly increase your multiplier by a tick or two, boot to windows, run Prime95, and monitor temps with RealTemp for 15 minutes. If stable, bump it up again and repeat. You may need to slightly tweak Ram voltages (e.g., VCCIO) to keep up with your OC, but that's only at the top end of what your chip can handle. And there are lots of people to help answer questions if you have them.

I will add, however, that single best upgrade for ALL applications these days is to move your OS and most used applications to a current-gen SSD. You will be AMAZED at the performance difference, whereas you may not notice the benefits of an OC as much. If you can afford a decent CPU cooler AND an SSD, I say go for both and you'll feel like you have a new machine.
 
Last edited:

coffeejunkee

Golden Member
Jul 31, 2010
1,153
0
0
1.5V is way too high for these chips, for longevity try to stay under 1.35V would be my advice. Also if by standard voltage you mean auto: it's not standard voltage.

OP, just try a small oc like 4GHz (stock cooler should be able to handle that although it will get hot when stresstesting) to see if it helps. If it does: great. If it doesn't: set back to stock and try undervolting.
 
Last edited:

cytoSiN

Platinum Member
Jul 11, 2002
2,262
7
81
1.5V is way too high for these chips, for longevity try to stay under 1.35V would be my advice. Also if by standard voltage you mean auto: it's not standard voltage.

OP, just try a small oc like 4GHz (stock cooler should be able to handle that although it will get hot when stresstesting) to see if it helps. If it does: great. If it doesn't: set back to stock and try undervolting.

Not sure where you're coming from, but the common wisdom (for the last several years) is that 1.5v is a comfortable maximum for a 2500k. As I said, I wouldn't go higher than that. That's what I meant by standard. If you want to play it safe, then stay below that, of course. In fact, I realize that I had a typo in my previous post. I've been stable at 4.5 for 2+ years at just over 1.3V, so you shouldn't even need to get up to 1.5.
 

coffeejunkee

Golden Member
Jul 31, 2010
1,153
0
0
Well, my common wisdom says 1.35V. 1.4V if you like pushing it and don't mind possible degradation.

Anyway, even with cool running Sandy Bridge chips you'll need monster cooling for 1.5V.
 

BonzaiDuck

Lifer
Jun 30, 2004
15,785
1,500
126
Not sure where you're coming from, but the common wisdom (for the last several years) is that 1.5v is a comfortable maximum for a 2500k. As I said, I wouldn't go higher than that. That's what I meant by standard. If you want to play it safe, then stay below that, of course. In fact, I realize that I had a typo in my previous post. I've been stable at 4.5 for 2+ years at just over 1.3V, so you shouldn't even need to get up to 1.5.

I'll try to be polite, and I could even be wrong -- so I encourage "correction." But this 1.5V "maximum" is a myth. The 1.5V spec is simply the maximum VID in a range of VIDs that intel chips can come with.

The lithography of these chips has been decreasing with (almost) every generation. At one time, some Intel cores were 95nm, and the impetus was to decrease this value and pack more transistors and circuits within less and less space. Doing that, and aware that the design of CPUs is getting close to the "molecular" level, means a greater potential for leakage at voltages exceeding what is lower and lower with each exceeding die shrink.

At same time, Intel had been providing spec guidelines for two ranges of voltage -- one within the other. They specified an upper limit for a "safe" voltage range, and above that a maximum for an "operable" voltage range. And I believe Intel had themselves noted that running a CPU between the safe and operable limits -- still in the operable range -- would cause it to degrade over time. They also noted that the CPU might then run at stock speeds and voltage, but you couldn't expect anything more.

I could be off by one chip generation, but the last spec offered for "safe" voltage came with the Nehalem socket-1366 CPUs. And I think the "safe" limit at that time was about 1.375V. After that, enthusiasts are "left on their own" to sort it out. There was a consensus here that 1.35V was a good target for the Sandy Bridge cores, and since IB and Haswell represent a die shrink from SB's 32nm to 22nm, you would expect the last two generations to be "safe" at a lower voltage.

Now . . . so-called "safe" limits are just a number that represents a probability of damage. In fact, there are various probabilities of failure and damage as voltage is increased. You could interpret that to mean that you could exceed the prudent voltage by -- say -- 20 millivolts while not expecting damage anytime soon. Or you could interpret it to mean that running the processor well within the guidelines has some small probability of failure as well.

Some well-respected among us -- our veterans and gurus who work in the "industry" -- argue their belief that Intel has really made these CPUs capable of some longevity and performance above our cautious, consensual speculation. And we see posters and "authorities" on various other forums pushing numbers like "1.40V" or "1.45V." And that may be fine -- you can really overclock your system with such settings if you can deal with the increased heat. But since these numbers are "higher" among various recommendations, you would also expect accelerated degradation or failure -- with "greater probability."

On the more optimistic side for Sandy Bridge, if your severely-loaded voltage is just about 1.35V, it probably is set to an unloaded (idle) value of 1.38V, and the loaded value is the one to watch. Further, if you over-clock and allow the EIST settings to be enabled, more or less of the time the CPU is used may exhibit these higher voltages. Someone "Folding at Home" or running Prime95 "searches for primes" on a 24/7 basis will run their systems at the upper loaded voltage much longer over the CPUs lifetime; others playing games or performing a variety of tasks my show an "average voltage" over time influenced more by the EIST speed and voltage used.
 

cytoSiN

Platinum Member
Jul 11, 2002
2,262
7
81
I'll try to be polite, and I could even be wrong -- so I encourage "correction." But this 1.5V "maximum" is a myth. The 1.5V spec is simply the maximum VID in a range of VIDs that intel chips can come with.

The lithography of these chips has been decreasing with (almost) every generation. At one time, some Intel cores were 95nm, and the impetus was to decrease this value and pack more transistors and circuits within less and less space. Doing that, and aware that the design of CPUs is getting close to the "molecular" level, means a greater potential for leakage at voltages exceeding what is lower and lower with each exceeding die shrink.

At same time, Intel had been providing spec guidelines for two ranges of voltage -- one within the other. They specified an upper limit for a "safe" voltage range, and above that a maximum for an "operable" voltage range. And I believe Intel had themselves noted that running a CPU between the safe and operable limits -- still in the operable range -- would cause it to degrade over time. They also noted that the CPU might then run at stock speeds and voltage, but you couldn't expect anything more.

I could be off by one chip generation, but the last spec offered for "safe" voltage came with the Nehalem socket-1366 CPUs. And I think the "safe" limit at that time was about 1.375V. After that, enthusiasts are "left on their own" to sort it out. There was a consensus here that 1.35V was a good target for the Sandy Bridge cores, and since IB and Haswell represent a die shrink from SB's 32nm to 22nm, you would expect the last two generations to be "safe" at a lower voltage.

Now . . . so-called "safe" limits are just a number that represents a probability of damage. In fact, there are various probabilities of failure and damage as voltage is increased. You could interpret that to mean that you could exceed the prudent voltage by -- say -- 20 millivolts while not expecting damage anytime soon. Or you could interpret it to mean that running the processor well within the guidelines has some small probability of failure as well.

Some well-respected among us -- our veterans and gurus who work in the "industry" -- argue their belief that Intel has really made these CPUs capable of some longevity and performance above our cautious, consensual speculation. And we see posters and "authorities" on various other forums pushing numbers like "1.40V" or "1.45V." And that may be fine -- you can really overclock your system with such settings if you can deal with the increased heat. But since these numbers are "higher" among various recommendations, you would also expect accelerated degradation or failure -- with "greater probability."

On the more optimistic side for Sandy Bridge, if your severely-loaded voltage is just about 1.35V, it probably is set to an unloaded (idle) value of 1.38V, and the loaded value is the one to watch. Further, if you over-clock and allow the EIST settings to be enabled, more or less of the time the CPU is used may exhibit these higher voltages. Someone "Folding at Home" or running Prime95 "searches for primes" on a 24/7 basis will run their systems at the upper loaded voltage much longer over the CPUs lifetime; others playing games or performing a variety of tasks my show an "average voltage" over time influenced more by the EIST speed and voltage used.

I absolutely agree. All I was saying is that 1.5V is generally a safe place to stop, in response to the previous post that suggested that that was too high.
 

schmuckley

Platinum Member
Aug 18, 2011
2,335
1
0
You are doing yourself a disservice if you do not Overclock the 2500K.
You can get 15-25 more FPS just by OCing it and the RAM.
PS:Ignore that stuff about "Oh don't go over 1.5v on the RAM"
It causes no damage to volt the crap out of your RAM for Sandy Bridge to Haswell.
You can either believe all the benchers, or someone that read some false article sometime in the past.I ran ddr3 2133 on SB with 1.73V for months on end..no damage.
No one has ever reported any damage from doing it.It's been years now.

Furthermore SB runs cool, so you can get a nice, high OC on air/water.

I wouldn't go over 1.53v on air/water.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |