Sign in with
Sign up | Sign in
Your question

LGA 1156 I5 750 Overclock help

Last response: in Overclocking
Share
June 1, 2010 2:56:16 AM

I am looking for some help trying to hit a 3.8 ghz OC on my new rig. I have read many posts looking for some info on why I can not get a stable 3.8 with the average voltages that I am seeing.

My set up is
asus p7p55d-e pro mobo
i5 750 lga 1156
g-skill ddr3 1600 pc3 12800 ram 1.35volt ( switched this from a 1.65v ram set thinking that the imc /dram voltage was giving me the problem)

Right now i have a stable 3.6ghz oc with the following settings.
cpu ratio x20
BCLK x180
QPI lowest setting 5766
Speedstep tech disabled
cpu v 1.33125
imc v 1.15625
dram v 1.3500
cpu PLL 1.8125
pch v 1.0625
LLC disabled
cpu spread spectrum disabled
c1e support disabled
cstate disabled

Idel temps 32c linX tests peak at 62c
These settings already seem really high for a 3.6oc compared to what other people are getting.



I really want to hit a 3.8ghz oc mostly to get the highest oc for my cooler and getting 1600mhz out of my ram
I have read and used andy5174s thread. I was starting from his suggested voltage settings
(was not sure if i should have posted in his thread or start my own)

Andys5174s suggested voltages for 3.8
3.8GHz:
Load-Line Calibration: Disabled
Vcore= 1.264V (CPU-Z idle)
QPI/Vtt/IMC= 1.149V
PCH= 1.10V
PLL= 1.85V
RAM= Specified voltage for your RAM

I had my vcore up to 3.8v and my IMC up to 1.9v using a 19 x 200 and i was still getting crashes while stress testing it. I was afraid to go much higher especialy with the IMC being so close to intels max voltage of 1.21v. I tried bringing the Bclk down slowly trying to find something stable with no hope. I am really hoping I am missing something stupid with my bios settings. I cant figure out what else could be wrong. Unless I got a crap processor. I just hope im not that unlucky.

Anyways thanks in advanced for any help I may receive

More about : lga 1156 750 overclock

a b K Overclocking
June 3, 2010 6:56:26 PM

One thing comes to mind. Did you set the RAM ratio to the LOWEST ratio? Make sure it is set to 8x, not 10x, before you start your OC. Change the RAM ratio setting in the BIOS.
m
0
l
a c 100 K Overclocking
June 3, 2010 7:33:01 PM

180 base clock will need at least 1.21V Vtt (or IMC). I have mine set to 175x20 with turbo boost enabled and using 1.30635V Vcore and 1.20625 Vtt. This is on the ASUS P7P55D PRO mobo. On my last one, MSI P55GD65 it took 1.288V Vcore and 1.197V Vtt. At 180 bclk I found that the requirements go up quite a bit more. You might need as much as 1.35 or 1.4V Vcore and 1.25V Vtt. Just keep trying...

I'd like to suggest tho check your RAM profiles out with Everest. I also have 1600mhz RAM CL8, but within XMP1 profile it shows the settings for 1400mhz as well at CL7 so that's why I'm using 175 bclk x8=1400. RAM speeds have a very minor affect on performance, and checking my overal RAM latency in Everest the difference is about 2ns which is insignificant, really (from ~41.3ns to ~43.5)

Also, while 3.8ghz is a great OC you're not really pushing the envelop, so you can definitely turn on C1E, C States, EIST. Oh and without turbo it should technically require a little less Vcore than I'm using since mine has to be able to support up to 4.2ghz on 1 or 2 cores. Load Line Calibration is a bit of a grey zone... I've read different things on it. IMO it's good to have on for stability as it maintains a solid voltage while with it off you get some droop under load. However, the droop is within Intel specs... so it's hard to say which way to go.
m
0
l
Related resources
June 3, 2010 9:30:47 PM

Thanks for the replies.
I have a some what stable 3.8 right now with a 19x200 ram set to 1600mhz. Its passing 40min LinX tests. I will prime test it over night if it passes this next linX test

my vcore is at 1.375v and vtt at 1.200v the others set to auto ram set to the ram specs. Temps peak at 68c with LinX

I downloaded everest but I am not to sure on how to check my ram profiles. I checked under the Benchmark and motherboard/ram spots.

I tried turning c1e, cstates and turbo on with very little luck. Bios failed to post.

What pisses me off is that my 2 brothers built the exact same computers and they hit 3.8mhz with 1.3v vcore and a 1.16v vtt. They ordered their cpu's off of a site and i picked my up from a local store.


These voltages / temps look ok for a 3.8 ? My goal was a decent gaming rig that will at least last a year and a half. This has been my first OC as you could probably tell, I just hope it wont burn out to fast.
m
0
l
a c 100 K Overclocking
June 3, 2010 9:47:24 PM

It's ok, temps might be a little high what cooler do you use? Mine under load hits tops 63C with a Zalman CNPS10X Extreme. It's easier to get it stable with a lower bclk tho, you could try EIST on auto then you'll have access to the 21x multiplier... so 180 bclk could get you to the same speed and might be a little easier to get stable.
m
0
l
June 3, 2010 11:35:28 PM

I ended up trying the coolit eco enclosed water cooler. I wasnt to happy with it out of the box so i ended up adding a high static pressure fan for a push / pull set up. Honestly I should of just went with a nice high end air for the same price and less messing around with the fans. The only pro about the eco is that its very clean looking and i have a plexi side panel. Pure cooling though, i kinda regret it now.


If i end up going with the lower bclk my ram freq will come down. But you were saying I would never knotice a difference between 1440mhz to 1600mhz.


what were you using for your stress test? With prime my temps only hit around 60. I find linX was giving me higher temps
m
0
l
a c 100 K Overclocking
June 3, 2010 11:47:04 PM

Yea linpack tests like LinX do tend to run a bit hotter. I prefer Intel Burn Test to LinX but they're pretty much the same thing. I find those are good to get a general idea of stability but you should still aim to get about 8-10hrs stable in Prime95 Large FFT.
m
0
l
June 3, 2010 11:56:56 PM

Good news though, i got eist back on and its looking stable so far. will keep testing. i may try to get my cvolts down a bit more. i think my vtt was the problem
m
0
l
!