Sign in with
Sign up | Sign in
Your question

Interesting things happen when i overclock my i7 2600k

Last response: in Overclocking
Share
January 24, 2011 6:23:24 PM

First off here are my system Specs
CPU: I7 2600k @ 4.7 ghz
-Core Voltage: 1.408
-VTT: 1.051
-PCH:1.059
-PLL:1.832
-VCCSA:.925
-BCLK:100mhz
-Multiplier: x47
- D2 stepping
MOBO: ASRock p67extreme 4.
RAM: OCZ Reaper @ 1600 mhz.
GPU: 6970 in Crossfire

PSU 1000w antec TruePower.

Monitoring systems: Asrock extreme tuner,Speedfan, CPU-Z, OCCT.

So here is my issue:
When i run OCCT the processor downclocks itself to 4.0 or 4.1 GHZ...the multiplier itself is being lowered while the voltages stay the same. Now here is the crazy thing. C1,C3,C6,EIST Speedstep and Thermal Throttling are all disabled. The temps max out at 60 degrees celsius while under 100% load using the Linpack test in OCCT and sits at 22 Degrees under idle. So what is going on here? The only thing i can think of is that the CPU is throttling itself, but i know i am well within the safety zone as far as Temps go. The only other thing i can think of is that the processor is not getting enough juice somewhere or that the problem resides on the MOBO itself.

My target overclock for this processor is at least 4.8-5 ghz. When i raise the Multi to 48 the computer will POST but it will not boot into the OS. So far the computer behaves proper at 4.6 ghz, The multiplier is stable and temps are safe. By all means 3.4 to 4.6 is a good overclock (especially when compared to my qx6700's inability to get past 3.2)but i know these D2 chips are capable of more than that while running on air.

I am brand new to overclocking on the core-i7 platform and so far i am very impressed with it but clearly i still have much to learn given my inability to trouble shoot this problem. That said i have scoured the internet (including these forums) looking for answers to my problems and have yet to find anyone else with the same issue.

Yes i have read all the overclocking guides available for sandy bridge and i am either missing something or they simply do not cover enough for me to troubleshoot my problem. That or i am just a total n00b who has no clue what he is doing lol.

Thanks in advance,

-Lotus
a b K Overclocking
January 24, 2011 6:49:00 PM

I am experiencing a similar issue.

I OC'd to 4.0, 4.4 & 4.5 and 4.5 is my current speed. Every speed that i tried, speedstep still kicked in and throttled to 1.6 when idle, no matter what i tried, it still did that.

set base multi to 34, set turbo multi's to 45, disabled speedstep and the like...

using 1.35 vcore and it is completely stable using prime 95 @ 4.5 and stays at that speed when stress testing with prime 95, soon as i end prime 95, it drops back down to 1.6 when idle... weird...
m
0
l
January 24, 2011 7:00:06 PM

Jonpaul,

I was having the same problem as well up until yesterday. It disappeared when i used my MOBOS automatic overclock option. I took down the voltages in AXTU disabled speedstep and C-states and manually set everything and was able to get it to run stable. The multiplier stopped changing on me. I have no clue what stopped it to be honest.
m
0
l
Related resources
a b K Overclocking
January 24, 2011 7:37:14 PM

nice, i will see if i can get the same results doing the same!
m
0
l
a b K Overclocking
January 24, 2011 7:48:02 PM

Lotus ... it shouldn't have been down-clocking when loaded up like that. I don't know what was causing your problem, but I'm glad it got fixed. However, your system will continue to down-clock when it becomes idle. That's a nice overclock for a D2 chip.

jon ... those settings are supposed to be enabled. The chips are supposed to down-clock when idle. They were specifically designed to do so, which is why the chip is ignoring your settings and doing it anyway.
m
0
l
January 24, 2011 7:53:56 PM

Thank you for your reply Leaps, However i am still having a problem with the processor downclocking when it is set to 4.7 ghz. It doesn't do it when it is set to 4.6 ghz. That extra 100 MHZ seems to set things all crazy.
m
0
l
a b K Overclocking
January 24, 2011 8:00:45 PM

Maybe that means you're hitting the ceiling -- 4.6GHz might be that chip's max good multiplier overclock.

And that speed makes more sense -- D2s fall into the 4.4-4.6GHz overclock range. D1s are usually needed for higher than that.

Keep the multi at 46 and see what your BCLK can get to without doing the throttling when loaded. That will then be your max OC.
m
0
l
January 24, 2011 8:10:40 PM

Hmm, i was under the impression that D2 was better for Multiplier OCing. In what increments should i increase the BCLK by?
m
0
l
a b K Overclocking
January 24, 2011 8:19:24 PM

Not according to PureOC. Start reading just under the white graphic with the three arrows going up.

BCLK won't go above 110MHz, and 103-105 is common. Do it in 1MHz steps with a stress test in between. It should let you know right away (within minutes) if it's too high.
m
0
l
January 24, 2011 9:02:01 PM

Thank you again Leaps.
I can't get the Bclk above 101 without it freaking out. At 102 it BSOD's and at 103 it won't boot into the OS.

I guess 4.6 is the limit of this chip. Still that is a damn good overclock compared to what we were getting back in the Conroe and Kentsfield days.
m
0
l
January 24, 2011 9:19:12 PM

Hmmm. So what should i increase? Remember i am a total newb to OCing on I7 and from what i understand OCing on SB is a new game itself.
m
0
l
January 24, 2011 10:30:36 PM

From that PureOC article that Leaps-from-Shadows mentioned:

Quote:
The next major difference is that Turbo now comes into play very heavily. [...] The Turbo will boost the multi on the chip when under load, like an "insta-OC". Linked to all this is that C1E, EIST, and Speedstep must be left ON. Unlike all previous generation processors that overclocked better with these turned off, that has now changed and these "speed adjusting" settings must be enabled.

Another critical point in overclocking Sandy Bridge is that all K-series chips have a multiplier wall. Previous generation Intel CPUs responded to increased voltage and lower temperatures to achieve higher clock speeds. That is no longer the case. A chip has a maximum multiplier it will achieve; additional voltage or lower temperatures will not improve the results. As mentioned, there is currently D1 and D2 stepping, and the D1 is preferred. Each chip will have an inherently different multiplier wall. Essentially this means that voltage, temperatures, and motherboard model will not make a difference in the maximum overclock of a particular chip; it all depends on the chip now.


So I suggest you try re-enabling the Speedstep if you haven't already, and go through OC steps up from 4.5GHz again and see how you go.

Shouldn't need to bump the voltage past 1.4v really.. 1.5v I've read is the recommended max.
Work with the Multiplier and BCLK and see how high you can get it.

Also, is your RAM OC'd at all? or are they running at the 1600MHz stock? If they're OC'd this may cause other instabilities, so try it at the stock speed as well as any OC'd levels.

Hope it helps (Get my 2600K in the next couple weeks)

Slothy
m
0
l
January 24, 2011 10:42:39 PM

Nah i have my RAM running at stock settings. I rarely touch the RAM unless it is necessary.
m
0
l
a b K Overclocking
January 25, 2011 11:44:34 AM

Leaps-from-Shadows said:
jon ... those settings are supposed to be enabled. The chips are supposed to down-clock when idle. They were specifically designed to do so, which is why the chip is ignoring your settings and doing it anyway.



Turns out that I didn't have C3 & C6 disabled, soon as i disabled them, no more speedstep, I now have a 100% OC of 4.5 GHz @ 1.35 vcore. I will try to back down the voltage a little bit or up the multi and/or BCLK...
m
0
l
January 26, 2011 7:03:04 AM

Well things get a little crazy at the 4.6-4.8 seems to be across the board with this chip. The sweet spot seemes to be around this range. In most cases if you have a lower end mobo that is usually the case of the bottleneck. Turbo boost will throttle the CPU this needs to be disabled in order to obtain the un-throttled 4.7 and above. This worked for me. I used gigabytes profile for the 4.0 overclock then increased the multi 100mhz each time from there, leaving the vcore on auto, the vcore is then adjusted accordingly with the increases in clock speeds automatically. I have my current clock at 4.5 and it is stable as a rock, i can go higher but see no point, i want a long lifespan on this chip. One problem people have had is not updating their BIOS to the latest version, chances are if you have not done this, this is one of the main factors as to why you cant get a higher clock, the latest BIOS for my board removed a problem with thermal limit ceilings. This chip is a beast but takes some time getting used to on how to overclock it , it helps when shooting for those high overclocks to have a high end heat sink (thats a given) but also for anything past 4.6 a high end mobo is also nice, people having a majority of their problems are using entry level and mid level mobos and that is usually the source of their problems.
m
0
l
a b K Overclocking
January 26, 2011 7:19:33 AM

TheLotusEater725 said:
Jonpaul,

I was having the same problem as well up until yesterday. It disappeared when i used my MOBOS automatic overclock option. I took down the voltages in AXTU disabled speedstep and C-states and manually set everything and was able to get it to run stable. The multiplier stopped changing on me. I have no clue what stopped it to be honest.
Max out the amperage and wattage limits in BIOS :) 
m
0
l
January 26, 2011 7:49:08 AM

Ok i was reading this forum a bit and see some irregularities in past things i have read, under this Overclocking guide http://www.xtremesystems.org/forums/showthread.php?t=26... he states to disable the following
" C1E-DISABLE
EIST-DISABLE
C3, C6 States-DISABLE
CPU Thermal Monitor-DISABLE"

now from what slothy has said earlier it says just the opposite

"The next major difference is that Turbo now comes into play very heavily. [...] The Turbo will boost the multi on the chip when under load, like an "insta-OC". Linked to all this is that C1E, EIST, and Speedstep must be left ON. Unlike all previous generation processors that overclocked better with these turned off, that has now changed and these "speed adjusting" settings must be enabled."

Can anyone please clarify this ?
m
0
l
a b K Overclocking
January 26, 2011 8:13:30 AM

sockpirate said:
Ok i was reading this forum a bit and see some irregularities in past things i have read, under this Overclocking guide http://www.xtremesystems.org/forums/showthread.php?t=26... he states to disable the following
" C1E-DISABLE
EIST-DISABLE
C3, C6 States-DISABLE
CPU Thermal Monitor-DISABLE"

now from what slothy has said earlier it says just the opposite

"The next major difference is that Turbo now comes into play very heavily. [...] The Turbo will boost the multi on the chip when under load, like an "insta-OC". Linked to all this is that C1E, EIST, and Speedstep must be left ON. Unlike all previous generation processors that overclocked better with these turned off, that has now changed and these "speed adjusting" settings must be enabled."

Can anyone please clarify this ?


Turbo multipliers are the ones that are unlocked, but some motherboard manufacturers found a workaround to trick the CPU into running "full turbo full time". Actual settings therefore depend on the motherboard manufacturer.
m
0
l
January 26, 2011 8:37:48 AM

I see, this guide was specifically for the gigabyte series p67 boards. Makes sense.
m
0
l
Anonymous
a b K Overclocking
January 26, 2011 8:41:51 AM

You need to remember that not every CPU is a perfect sample. They will all run at stock speed just fine, most will also run up to 4.5ghz or so, but after that the production quality of your CPU will show.

Not every D2 can reach 5ghz, thats just a fact of life.
m
0
l
a b K Overclocking
January 26, 2011 11:46:13 AM

sockpirate said:
Ok i was reading this forum a bit and see some irregularities in past things i have read, under this Overclocking guide http://www.xtremesystems.org/forums/showthread.php?t=26... he states to disable the following
" C1E-DISABLE
EIST-DISABLE
C3, C6 States-DISABLE
CPU Thermal Monitor-DISABLE"

now from what slothy has said earlier it says just the opposite

"The next major difference is that Turbo now comes into play very heavily. [...] The Turbo will boost the multi on the chip when under load, like an "insta-OC". Linked to all this is that C1E, EIST, and Speedstep must be left ON. Unlike all previous generation processors that overclocked better with these turned off, that has now changed and these "speed adjusting" settings must be enabled."

Can anyone please clarify this ?



I have the ASUS P8P67 Pro and i have "tricked my CPU into running full turbo mode with 1.33 vcore, has been @ 4.5 GHz for the past 36 hours. Using the Hyper 212+, temps have not gone past 41c and i have played games, burned CD's, and watched videos within that 36 hours. That is what i call a "Sweet Spot" 4.5 GHz is just fine (if not overkill) for everyday use.

I did the following to "trick the CPU"

C1E-DISABLE
EIST-DISABLE
C3, C6 States-DISABLE
CPU Thermal Monitor-DISABLE
Speedstep-DISABLE
m
0
l
a b K Overclocking
January 26, 2011 7:21:29 PM

jonpaul37 said:
I have the ASUS P8P67 Pro and i have "tricked my CPU into running full turbo mode with 1.33 vcore, has been @ 4.5 GHz for the past 36 hours. Using the Hyper 212+, temps have not gone past 41c and i have played games, burned CD's, and watched videos within that 36 hours. That is what i call a "Sweet Spot" 4.5 GHz is just fine (if not overkill) for everyday use.

I did the following to "trick the CPU"

C1E-DISABLE
EIST-DISABLE
C3, C6 States-DISABLE
CPU Thermal Monitor-DISABLE
Speedstep-DISABLE


Actually I think the board did most of the tricking for you: All you did is disable the downclocking, the board is doing the upclocking based on your selection of turbo multiplier.
m
0
l
January 26, 2011 8:47:26 PM

One thing that will help past the 4.5-4.6+ clocks will be update the the latest BIOS F7A (IN MY CASE) and enable the option "INTERNAL CPU PLL OVERVOLTAGE" .
m
0
l
a b K Overclocking
January 26, 2011 10:29:53 PM

I have that same board and I'm trying to figure out the same thing basically. I think a lot of this has to do with the Long/Short term duration wattage limit which is set to Auto, and should be upped to at least 210w

Possibly Core Current limit too Haven't messed with that one yet. That will be my next step if this manual Prime 95 test fails that I'm running now. May be need to increased from default of 150

Also, using 'Offest' for you manual voltage may help keep it stable. Use the +
m
0
l
January 27, 2011 6:56:55 AM

Also enabling LLC to level 1 or 2 will also help.

LOAD LINE CALIBRATION
m
0
l
a b K Overclocking
January 27, 2011 8:19:41 AM

Something for those who have decided to disable the C States settings in the BIOS (and those who are thinking about it):

From ASUS Technical Marketing Specialist Juan_Jose on the HardOCP forums:

"An important note to keep in mind is that disabling CStates can considerably affect HD performance ( especially SATA6G )"

These CPUs really were designed to keep the SpeedStep and throttling and such enabled for maximum power savings. Disabling them evidently does some negative things for hard drive throughput. How much degradation, and whether you'd actually notice it, is unclear at this point.
m
0
l
a b K Overclocking
January 27, 2011 12:39:08 PM

Leaps-from-Shadows said:
Something for those who have decided to disable the C States settings in the BIOS (and those who are thinking about it):

From ASUS Technical Marketing Specialist Juan_Jose on the HardOCP forums:

"An important note to keep in mind is that disabling CStates can considerably affect HD performance ( especially SATA6G )"

These CPUs really were designed to keep the SpeedStep and throttling and such enabled for maximum power savings. Disabling them evidently does some negative things for hard drive throughput. How much degradation, and whether you'd actually notice it, is unclear at this point.



So far, I've not noticed anything all in regards to SATA HD performance. It has been the same on stock and OC with the C's disabled. I am not using SATA6 yet but soon...
m
0
l
January 28, 2011 5:46:34 PM


My SATA performance has remained unchanged regardless of C state.
m
0
l
!