Overclocking 2500k on Asus p8p67 help!!

[cougar1111]

so this is my build i have i5 2500k with an asus p8p67 mobo EVGA GTX 570 16gb of g.skill 1333 4x 4gb sticks 800wat Rosewill PSU and a Zalman cpu fan a cooler master HAF X case 60gb OCZ ssd 1.5TB WD black getting the cooler master v8 in 3 days

ok so my first thing is i use the Tpu setting so it overclock's it at 4.3ghz my idle temps are 30 to 40 and underload low 60s i have no prob overclocking to 4.8ghz with 1.36vc my prob is when i overclock no matter what i do my 3d mark goes down i will post links to 4.3 and 4.6 just to show you how big of a difference it makes and have in mind only the multiplier is change and i think i read on alot of websites that up to 4.6 nothing needs to be changed it can all be set on auto here are the links

Edit my GPU is Clock same in all of this at 800 coreclock 1464 shader clock 2040 memory clock temps underload low 70s at fan 85% is as high as it goes

defaults with TPU on http://3dmark.com/3dm11/1903343

same thing 4.6 with everything on auto http://3dmark.com/3dm11/1903632

boots up runs fine even stable with prime 95 over night i dont have any prob overclocking what so ever my prob is whats with 3d mark going down as i go up my cpu seems really good i can run at 4.8ghz no prob at like 1.36 1.37
and my heat does not really move it pretty much stays the same so if any one could help maybe cause i just dont get it at all im kinda tired of trying frist form iv been on in years

 

[nzaneb]

Within the margin of error. Have you tried running it multiple times in a row? It will be primarily GPU bound regardless.

 

[cougar1111]

if u mean like going from 4.3 to 4.4 or from stock from 3.3 to 3.4 yeah iv did over 21 test in 3d mark nothing i can do seems to make a difference
im just trying to get to 4.8 stable and id be happy ik my 3d mark should be over 1980 is that a bad score anyways for the 2500k and gtx 570?

 

[nzaneb]

I meant to run it multiple times one right after the other, same settings. You'll likely get score swings. Extreme preset will primarily only respond to GPU clocks (within reason)

 

[cougar1111]

oh yeah iv did that its always lower then the stock TPU 4.3 no matter what i change or how many times i run it thats what i seem to be stuck on i get in the x1970 but i did have 1 score at 4.9 almost 5ghz where i got a x1987

 

[cougar1111]

so i should be running performance to see a difference in my cpu score ?

 

[nzaneb]

so i should be running performance to see a difference in my cpu score ?

It will be more apparent, but 3dMark11 is not as much a CPU intensive bench as Vantage was.

It's entirely possible that your GPU throttles a little bit during one of the tests, or power delivery interupts the performance momentarily, though it's not too likely.

 

[cougar1111]

can u give me a good one for my cpu only iv found my max on gpu but im trying to get a good overclock on my cpu u dont think my 1 gtx 570 is bottle neck my 2500k?

 

[cougar1111]

this is performance test this is at 4.9Ghz http://3dmark.com/3dm11/1886952

this is at stock TPu 4.3 http://3dmark.com/3dm11/1906476

maybe this are good maybe not maybe im asking for more then im going to get with my sys spec's

 

[nzaneb]

can u give me a good one for my cpu only iv found my max on gpu but im trying to get a good overclock on my cpu u dont think my 1 gtx 570 is bottle neck my 2500k?

In order to see CPU gains run a 2D benchmark, WPrime is a good one. Make sure you set it to all threads available, run it at stock, then run it at you Overclocked settings.

Hwbot is a good comparison database for overclocked scores. Here are all the benchmarks included in their database, along with download links Hwbot Downloads

Additionally you can register with HWbot, and upload your completed benches. Visit us in the Benchmarking section of the forums if you have any questions or need any help.

 

[deathman20]

Looks right, I mean the only one that got any boost was the physics test really.

Same could be said with 3DMark 11 with my i7 920 @ 4.2Ghz and my new i5 2500k @ 4.5Ghz, its nearly identical. Why? Because the GPU is the one that is tested in this case vs in the past it was primary a mix of CPU and GPU tests.

 

[cougar1111]

yeah thats so True cause here is a new test i just ran tonight and its way better then the past test by a long ways iv got alot more out of my video card i put the Volt at 1.1v and overclock it to 900mhz and 2060memory to blast me to this http://3dmark.com/3dm11/1915933


edit my cpu load temps in 3d mark are 71 and under onlu 1 core got to 71c the rest are high 60s vcore is set at 1.380 not yet sure if its stable didnt really care to run 24/7 at this just for 3d mark my mother boreds TPU overclock of 4.3ghz is all i will ever need but do you think this are good temps on air

note my temps at 4.3 are 30 idle and 50 low 60s under full load tested prime 95 over 12hr

 

[nzaneb]

note my temps at 4.3 are 30 idle and 50 low 60s under full load tested prime 95 over 12hr

Yes that is fine