Sign in with
Sign up | Sign in
Your question

A few questions about my i5-2500k@4.5GHz please!

Last response: in Overclocking
Share
March 6, 2013 6:20:44 PM

Just have a few questions on my extremely long and extensive overclocking on my i5-2500k.

Right now my hardware that pertains to this is:
i5-2500k@4.5GHz & 1.335 Vcore
Thermaltake Contac 29 heatsink with one fan (waiting on my other fan for push/pull)
750w thermaltake psu
gigabyte z68 mobo with touch bios (all settings related to throttling the cpu when not under heavy load are on auto for the time being)
EVGA GTX 560 Ti (overclocked from 860 to 910 core clock)

1) I have been using OCCT CPU (large data set), and OCCT Linpack for the first tests (cpu@2hrs & linpack@3hrs) and then switching to a 10hr prime95 blend test. If all these pass, I would only imagine that it would be safe to call it a stable clock. Would you? My temps dont even hit 60c with this heatsink by the way.
2) After I can call it stable on the above mentioned overclocking settings (voltages and clock speeds) I will be lowering the voltages, running the tests again, and seeing how much power consumption I can save before it becomes unstable. With lowering the voltages, if I just change the values of my "VCore" will that change the voltages running to the cpu? Sounds like a dumb question, but I have been curious as to if my voltages are actually being changed to the value that I set on the vcore. After I get to the lowest value on the voltage, I will be running the same sets of tests as mentioned in question one.
3) What are your feelings about auto settings for the VCore? I have seen a few opinions, but still haven't came to a conclusion on my own about it. If I set the vcore manually, will this be the constant voltage applied to the cpu? I wouldn't mind having that manual set voltage set as a max, but in times of not having full load, it would be nice to conserve a little power if the manual setting doesn't throttle depending on load.
4) While testing, I have realtemp and cpu-z running. Cpu-z shows what my voltage is when its under max load, could I base my voltage changes on this number that is shown? Or just stick with lowering the voltages by .005 and then testing again?
5) Would this high of a clock potentially bottleneck my 560 Ti? As mentioned I did also overclock my gpu a bit.
6) Would you also say that the clock for my cpu mentioned is a high/good clock?
7) Lastly, how do you feel about leaving all the settings that people say to turn off during overclocking (c1e, eist, c-states etc) on auto? I have definitely disabled the turbo feature as I don't really want to be running my cpu harder than what I set it to at any point even after a successful clock.

I would be very grateful if anyone with good knowlege can help me with all this. To me it seems like a lot to ask, but I figured that I would try and see what kind of help I can get on this. It doesnt take rocket science to up the multiplier on the cpu (which is how I have overclocked, along with vcore changes) but I am still learning a lot on this. A lot more goes into this than overclocking the gpu haha.
I will take any advice and other questions that you may have in order to help me out.
Thanks!

More about : questions 2500k 5ghz

March 6, 2013 7:27:08 PM

Also. Would it be better to raise my base clock to 200 instead of 100 and my multiplier to 23 instead of 45? Or can I just keep raising the multiplier and leave the base clock to 100. Sorry for all the questions v.v
m
0
l
a b à CPUs
a b K Overclocking
March 7, 2013 2:55:53 PM

Your asking too many questions. On this site, hardly anyone will answer posts like this. You need to ask one or two specific questions that are the most important. And after those are answered then maybe ask another question to the person who answered your first question.

As for you last question. No you cannot just raise the BCLK. It doesn't work like that. With 2500k's you can only adjust the multiplier.

Yes you should disable eist, c1e, etc. Unless your really anal about saving 25 cents on your electricity bill per month. Leaving those settings on have never proved that your CPU will last longer. Typically I achieve better overclocks disabling those settings. So disable them.

4.5Ghz is a good overclock. 4.5Ghz seems to be where most people will stop because of excess heat from requiring too much voltage. I have a 3570k and I'm at 4.6Ghz. But I can get to 4.8Ghz while still being under the max temp barrier.

For stress testing you must use prime95 and Intel Burn In Test to truly test stability. Those are the most proven programs. I definitely recommend running Intel Burn In Test. So run Intel Burn In Test and then let me know what your temps are and I'll tell you if they're good or not. Typically staying below 85c is the way to go. Some people like to stay around 70c. It's good to run prime95 and IBT because that's what everyone else runs. Then you can compare your results. With OCCT, hardy anyone uses that so you'll have no one to compare with.

m
0
l
Related resources
March 7, 2013 4:01:47 PM

ericjohn004 said:
Your asking too many questions. On this site, hardly anyone will answer posts like this. You need to ask one or two specific questions that are the most important. And after those are answered then maybe ask another question to the person who answered your first question.

As for you last question. No you cannot just raise the BCLK. It doesn't work like that. With 2500k's you can only adjust the multiplier.

Yes you should disable eist, c1e, etc. Unless your really anal about saving 25 cents on your electricity bill per month. Leaving those settings on have never proved that your CPU will last longer. Typically I achieve better overclocks disabling those settings. So disable them.

4.5Ghz is a good overclock. 4.5Ghz seems to be where most people will stop because of excess heat from requiring too much voltage. I have a 3570k and I'm at 4.6Ghz. But I can get to 4.8Ghz while still being under the max temp barrier.

For stress testing you must use prime95 and Intel Burn In Test to truly test stability. Those are the most proven programs. I definitely recommend running Intel Burn In Test. So run Intel Burn In Test and then let me know what your temps are and I'll tell you if they're good or not. Typically staying below 85c is the way to go. Some people like to stay around 70c. It's good to run prime95 and IBT because that's what everyone else runs. Then you can compare your results. With OCCT, hardy anyone uses that so you'll have no one to compare with.


My apologies. I appreciate the input and will take your advice into action. I have heard about intel burn but I didnt really see the point in running prime95 and other programs. Wish there was just one that would suffice. Ill let you know what I get on that.
m
0
l
a b à CPUs
a b K Overclocking
March 7, 2013 7:19:20 PM

The programs are very small in size so it's not really a big deal to have multiple programs to test with. And sometimes one or two programs will pass and the third will fail. Which then tells you to raise or lower your voltage or multiplier. So it's good to run more than one, always. Intel Burn In Test is normally the best though, although it gives you extremely high temps. I use prime95 to see what my real world max temps will probably be and I use IBT to test what my synthetic temps will be.

As far as my saying your asking too many questions... I was just giving a a really good tip on how to get answers on this forum. As you can see I'm the only one who responded. So I was just letting you know what to do next time you post as I've done the same thing as you just did and time or two and had no answers to show for it.

Let me know if you have any more questions.
m
0
l
March 7, 2013 7:23:59 PM

oh yeah for sure, I wasn't taking any offense to that :D 
Yeah I noticed that IBT was giving high temps, Im getting about 78c right now. Is that normal? Not gonna lie, that freaked me out and I ended the 15cycle test. Prime95 is giving me a max of 71c on my highest core, so its good to know what you mean by synthetic/real time temps.
So tell me if this sounds like a good plan.
IBT-15 cycles
Prime95- small ffts for 2hrs, then blend for 12hrs all on stock settings for the program

I mean, this is what I have came up with on all the different methods of testing.
m
0
l
!