Help with overclocking...

[blackbelter]

Hey guys. I have been trying to oc my system (i5 3570+Asus P8Z77V Pro). Originally I had 4.5GHz in mind. After like a million failures, I have changed my goal to 4.3GHz.

 

The system was stable under stress (Prime95 and IBT), but not under normal use. I encounter BSOD and freezes occasionally. So I kept bumping up voltages a little. But the situation remained.

 

So I switched to Asus auto tuning. It oced my system like crazy, to 4.6XGHz. Without me stopping it, I think it will go even further. But it is not really doing a good job. When I ran stress test, the temperature reached 105 degree. And I still have had a few BSOD. Now I think I am allergic to it already.

 

So I switched again to manual ocing. I got it at 4.3GHz, with a voltage ~1.2X under load. But it is still not absolutely stable, in the sense that when I played a little track in DCS A10C, if I tried to paused it, go to the desktop, check cpu temperature and voltage, and finally go back to DCS, it will crash. Is this a problem of my system, or that of DCS? Mind you that I am running DCS version 1.1.0.5.

 

Also, how come my system can run IBT without problems, but cannot sustain everyday jobs (surfing, movie playing, gaming)? Is it that the vcore not high enough? Or something to do with the memory?

 

Thank you! I am a little desperate now...:mad:

[Corrigan]

Not sure how much it helps you, but I'm pretty sure I've linked my game ctds described here

 

http://forums.eagle.ru/showthread.php?t=88557

 

to having an unstable overclock. So, dcs crashing for you definitely could be your system.

 

However, with a stable system I still saw one ctd when alt tabbing back in, before I switched to widowed mode. If you ONLY see crashing after tabbing, I'd guess it's the game.

Edited May 16, 2012 by Corrigan

[Mustur]

Does prime95 stay stable with the small FFTs option? If so then your CPU is probably fine.

 

My guess would be that your RAM is causing instability. Run either prime95 (with the in-place large fft option or blend) or Memtest86+. Then we will know for sure.

 

In the case that it is the RAM, try to loosen RAM timings and increase RAM voltage (to a safe point!).

[Mugenjin]

How many hours did you run prime?

 

Also prime stable does not guarantee a really stable system.

Why don't you just run with something slightly slower, like 4.0? It won't make that much of a difference.

[blackbelter]

However, with a stable system I still saw one ctd when alt tabbing back in, before I switched to widowed mode. If you ONLY see crashing after tabbing, I'd guess it's the game.

 

Sadly, I did see a few spontaneous crashing without me tabbing anything...

 

Does prime95 stay stable with the small FFTs option? If so then your CPU is probably fine.

 

My guess would be that your RAM is causing instability. Run either prime95 (with the in-place large fft option or blend) or Memtest86+. Then we will know for sure.

 

In the case that it is the RAM, try to loosen RAM timings and increase RAM voltage (to a safe point!).

 

It is definitely stable under Prime95. Wanted to run memtest86+, but could not get it run. I used usb boot version. It simply cannot boot. Ended up freezing my system at the windows logo every time. Not sure what to do about that. Maybe I try straight away loosening RAM timings? My question is that, what are the valid timing combinations? Can I just try something like 10-10-10-30? Or are there just a few valid combination of numbers? I have no idea about that.

 

How many hours did you run prime?

 

Also prime stable does not guarantee a really stable system.

Why don't you just run with something slightly slower, like 4.0? It won't make that much of a difference.

 

I didn't run Prime95 for too long. But my system did pass 20 IBT tests (maximum). I prefer IBT as I do not want my system to run for the whole night. I was hoping that is sufficient, too. Or is it? I might have to go with a lower OC... forced to...

 

I would run 3 runs of memtest to be sure ram is ok.

 

I cannot get it running... Please see above. No idea what to do about that.

[blackbelter]

How to use memtest

 

Download ISO image and burn to disc

Stick disk in drive

reboot PC , hit delete and go into bios

go to boot settings , make sure it is set to boot from CD before Hard drive in to boot order , save settings

now it should run memtest cd INSTEAD of loading windows , let it run min 3 passess.

 

I did all that, only with USB instead of cd... I'll try again later. Thanks.

[Corrigan]

I think you should assume that the game ctds are caused by system instability.

[blackbelter]

Thanks guys. I now loosened the RAM timing a little, and raised DRAM voltage a little. Now all is good so far at 4.4GHz.

 

I also made fool of myself. I configured the cooler to be push-pull. Today I discovered that instead of push-pull, I made it push-push... Damn! Now the temperature under load (IBT) is like more than 10 degrees lower than before.

[blackbelter]

LOL the system passed 10 rounds of ibt, 10 min of dcs a10 playback. I then posted the previous post. my computer bsod right after That. Now running memtest86+...

Edited May 17, 2012 by blackbelter

[blackbelter]

Oh good. The memtest86+ has been running for 6 hours and half, with no errors being reported.

 

But I did notice that the ram is running at 1200 (ddr3 2400) MHz, instead of 1800 MHz that I set in the bios. Maybe that's the problem?

Edited May 18, 2012 by blackbelter

[blackbelter]

I got the following information from the dmp file created after the blue screen.

 

------------------------------------------------------------------------------------------------------

Microsoft ® Windows Debugger Version 6.2.8229.0 AMD64

Copyright © Microsoft Corporation. All rights reserved.

 

 

Loading Dump File [C:\Users\WZ\Desktop\051812-15381-01.dmp]

Mini Kernel Dump File: Only registers and stack trace are available

 

Symbol search path is: SRV*c:\symbols*http://msdl.microsoft.com/download/symbols

Executable search path is:

Windows 7 Kernel Version 7601 (Service Pack 1) MP (4 procs) Free x64

Product: WinNt, suite: TerminalServer SingleUserTS

Built by: 7601.17803.amd64fre.win7sp1_gdr.120330-1504

Machine Name:

Kernel base = 0xfffff800`02e1b000 PsLoadedModuleList = 0xfffff800`0305f670

Debug session time: Thu May 17 21:11:57.760 2012 (UTC + 8:00)

System Uptime: 0 days 0:58:20.556

Loading Kernel Symbols

...............................................................

................................................................

..................................................

Loading User Symbols

Loading unloaded module list

.....

TRIAGER: Could not open triage file : G:\Program files\Windows Kits\8.0\Debuggers\x64\triage\oca.ini, error 2

TRIAGER: Could not open triage file : G:\Program files\Windows Kits\8.0\Debuggers\x64\winxp\triage.ini, error 2

*******************************************************************************

* *

* Bugcheck Analysis *

* *

*******************************************************************************

 

Use !analyze -v to get detailed debugging information.

 

BugCheck 50, {ffffff00ffffff28, 0, fffff8800f867f87, 7}

 

*** WARNING: Unable to verify timestamp for nvlddmkm.sys

*** ERROR: Module load completed but symbols could not be loaded for nvlddmkm.sys

 

Could not read faulting driver name

TRIAGER: Could not open triage file : G:\Program files\Windows Kits\8.0\Debuggers\x64\triage\modclass.ini, error 2

Probably caused by : nvlddmkm.sys ( nvlddmkm+241f87 )

 

Followup: MachineOwner

---------

 

1: kd> !analyze -v

*******************************************************************************

* *

* Bugcheck Analysis *

* *

*******************************************************************************

 

PAGE_FAULT_IN_NONPAGED_AREA (50)

Invalid system memory was referenced. This cannot be protected by try-except,

it must be protected by a Probe. Typically the address is just plain bad or it

is pointing at freed memory.

Arguments:

Arg1: ffffff00ffffff28, memory referenced.

Arg2: 0000000000000000, value 0 = read operation, 1 = write operation.

Arg3: fffff8800f867f87, If non-zero, the instruction address which referenced the bad memory

address.

Arg4: 0000000000000007, (reserved)

 

Debugging Details:

------------------

 

 

Could not read faulting driver name

TRIAGER: Could not open triage file : G:\Program files\Windows Kits\8.0\Debuggers\x64\triage\modclass.ini, error 2

 

READ_ADDRESS: GetPointerFromAddress: unable to read from fffff800030c9100

GetUlongFromAddress: unable to read from fffff800030c91c0

ffffff00ffffff28 Nonpaged pool

 

FAULTING_IP:

nvlddmkm+241f87

fffff880`0f867f87 488b4828 mov rcx,qword ptr [rax+28h]

 

MM_INTERNAL_CODE: 7

 

CUSTOMER_CRASH_COUNT: 1

 

DEFAULT_BUCKET_ID: WIN7_DRIVER_FAULT

 

BUGCHECK_STR: 0x50

 

PROCESS_NAME: firefox.exe

 

CURRENT_IRQL: 0

 

TRAP_FRAME: fffff8800b6bf790 -- (.trap 0xfffff8800b6bf790)

NOTE: The trap frame does not contain all registers.

Some register values may be zeroed or incorrect.

rax=ffffff00ffffff00 rbx=0000000000000000 rcx=fffff8800b6bf978

rdx=000000000001ffff rsi=0000000000000000 rdi=0000000000000000

rip=fffff8800f867f87 rsp=fffff8800b6bf928 rbp=00000000c1d00000

r8=fffff8800b6bf978 r9=fffffa800a9900b8 r10=fffff880009eddc0

r11=000000000001ffff r12=0000000000000000 r13=0000000000000000

r14=0000000000000000 r15=0000000000000000

iopl=0 nv up ei ng nz na po nc

nvlddmkm+0x241f87:

fffff880`0f867f87 488b4828 mov rcx,qword ptr [rax+28h] ds:ffffff00`ffffff28=????????????????

Resetting default scope

 

LAST_CONTROL_TRANSFER: from fffff80002e40ea7 to fffff80002e9a1c0

 

STACK_TEXT:

fffff880`0b6bf628 fffff800`02e40ea7 : 00000000`00000050 ffffff00`ffffff28 00000000`00000000 fffff880`0b6bf790 : nt!KeBugCheckEx

fffff880`0b6bf630 fffff800`02e982ee : 00000000`00000000 ffffff00`ffffff28 00000000`00000000 00000000`00000000 : nt! ?? ::FNODOBFM::`string'+0x43cd1

fffff880`0b6bf790 fffff880`0f867f87 : fffff880`0f85bf97 fffffa80`06713000 fffff880`0f98cc35 fffffa80`0c47f8f0 : nt!KiPageFault+0x16e

fffff880`0b6bf928 fffff880`0f85bf97 : fffffa80`06713000 fffff880`0f98cc35 fffffa80`0c47f8f0 fffff800`00000000 : nvlddmkm+0x241f87

fffff880`0b6bf930 fffffa80`06713000 : fffff880`0f98cc35 fffffa80`0c47f8f0 fffff800`00000000 00000000`00000000 : nvlddmkm+0x235f97

fffff880`0b6bf938 fffff880`0f98cc35 : fffffa80`0c47f8f0 fffff800`00000000 00000000`00000000 00000000`00000000 : 0xfffffa80`06713000

fffff880`0b6bf940 fffffa80`0c47f8f0 : fffff800`00000000 00000000`00000000 00000000`00000000 0000032f`80710f04 : nvlddmkm+0x366c35

fffff880`0b6bf948 fffff800`00000000 : 00000000`00000000 00000000`00000000 0000032f`80710f04 00000000`0000000a : 0xfffffa80`0c47f8f0

fffff880`0b6bf950 00000000`00000000 : 00000000`00000000 0000032f`80710f04 00000000`0000000a fffffa80`06713000 : 0xfffff800`00000000

 

 

STACK_COMMAND: kb

 

FOLLOWUP_IP:

nvlddmkm+241f87

fffff880`0f867f87 488b4828 mov rcx,qword ptr [rax+28h]

 

SYMBOL_STACK_INDEX: 3

 

SYMBOL_NAME: nvlddmkm+241f87

 

FOLLOWUP_NAME: MachineOwner

 

MODULE_NAME: nvlddmkm

 

IMAGE_NAME: nvlddmkm.sys

 

DEBUG_FLR_IMAGE_TIMESTAMP: 4f64ddb1

 

FAILURE_BUCKET_ID: X64_0x50_nvlddmkm+241f87

 

BUCKET_ID: X64_0x50_nvlddmkm+241f87

 

Followup: MachineOwner

-----------------------------------------------------------------------------------------------------

 

 

Apparently it has to do with firefox and the nvidia driver. I am gonna switch to chrome and reinstall the driver. Hope that will solve my problem.

[Pilotasso]

I had RAM modules passing hours of torture tests only to crash the PC ramdomly in normal use. Be sure to lock the RAM in its default settings and tweak the CPU only. Overclocking RAM yelds poor results infact its the components whose overclock affects performance the less.

Edited May 17, 2012 by Pilotasso

[blackbelter]

I had RAM modules passing hours of torture tests only to crash the PC ramdomly in normal use. Be sure to lock the RAM in its default settings and tweak the CPU only. Overclocking RAM yelds poor results infact its the components whose overclock affects performance the less.

 

I did lock the ram settings in the bios. Everything was set manually...

[Pilotasso]

and still giving problems?

[blackbelter]

and still giving problems?

 

There was a bsod. I assumed without evidence that it was caused by the ram. Now I checked the dmp file, which was posted above. Apparently it is caused by the graphics driver and Firefox.

 

I just ran Dcs for like 20 min. Ended up in blue screen again. Wtf...

[Pilotasso]

Sounds like a RAM issue, or your CPU isnt stable at that overclock. Your Dump Files might only show what was running and failed first at the time when it crashed, not the root cause.

[blackbelter]

Microsoft ® Windows Debugger Version 6.2.8229.0 AMD64

Copyright © Microsoft Corporation. All rights reserved.

 

 

Loading Dump File [C:\Users\WZ\Desktop\051812-10030-01.dmp]

Mini Kernel Dump File: Only registers and stack trace are available

 

Symbol search path is: SRV*c:\symbols*http://msdl.microsoft.com/download/symbols

Executable search path is:

Windows 7 Kernel Version 7601 (Service Pack 1) MP (4 procs) Free x64

Product: WinNt, suite: TerminalServer SingleUserTS

Built by: 7601.17803.amd64fre.win7sp1_gdr.120330-1504

Machine Name:

Kernel base = 0xfffff800`02e57000 PsLoadedModuleList = 0xfffff800`0309b670

Debug session time: Fri May 18 05:29:34.053 2012 (UTC + 8:00)

System Uptime: 0 days 1:04:30.240

Loading Kernel Symbols

...............................................................

................................................................

.................................................

Loading User Symbols

Loading unloaded module list

...................

TRIAGER: Could not open triage file : G:\Program files\Windows Kits\8.0\Debuggers\x64\triage\oca.ini, error 2

TRIAGER: Could not open triage file : G:\Program files\Windows Kits\8.0\Debuggers\x64\winxp\triage.ini, error 2

*******************************************************************************

* *

* Bugcheck Analysis *

* *

*******************************************************************************

 

Use !analyze -v to get detailed debugging information.

 

BugCheck FC, {fffff88003565180, 80000001db14e963, fffff8800351b290, 0}

 

TRIAGER: Could not open triage file : G:\Program files\Windows Kits\8.0\Debuggers\x64\triage\modclass.ini, error 2

Probably caused by : ntkrnlmp.exe ( nt! ?? ::FNODOBFM::`string'+438fc )

 

Followup: MachineOwner

---------

 

1: kd> !analyze -v

*******************************************************************************

* *

* Bugcheck Analysis *

* *

*******************************************************************************

 

ATTEMPTED_EXECUTE_OF_NOEXECUTE_MEMORY (fc)

An attempt was made to execute non-executable memory. The guilty driver

is on the stack trace (and is typically the current instruction pointer).

When possible, the guilty driver's name (Unicode string) is printed on

the bugcheck screen and saved in KiBugCheckDriver.

Arguments:

Arg1: fffff88003565180, Virtual address for the attempted execute.

Arg2: 80000001db14e963, PTE contents.

Arg3: fffff8800351b290, (reserved)

Arg4: 0000000000000000, (reserved)

 

Debugging Details:

------------------

 

TRIAGER: Could not open triage file : G:\Program files\Windows Kits\8.0\Debuggers\x64\triage\modclass.ini, error 2

 

CUSTOMER_CRASH_COUNT: 1

 

DEFAULT_BUCKET_ID: WIN7_DRIVER_FAULT

 

BUGCHECK_STR: 0xFC

 

PROCESS_NAME: System

 

CURRENT_IRQL: 2

 

TRAP_FRAME: fffff8800351b290 -- (.trap 0xfffff8800351b290)

NOTE: The trap frame does not contain all registers.

Some register values may be zeroed or incorrect.

rax=0000000000000003 rbx=0000000000000000 rcx=fffffa8006eeebf0

rdx=0000000000000102 rsi=0000000000000000 rdi=0000000000000000

rip=fffff88003565180 rsp=fffff8800351b428 rbp=fffffa8006eeec18

r8=fffff880009ef301 r9=0000000000000002 r10=000000000000001a

r11=0000000000000000 r12=0000000000000000 r13=0000000000000000

r14=0000000000000000 r15=0000000000000000

iopl=0 nv up ei pl zr na po nc

fffff880`03565180 801f00 sbb byte ptr [rdi],0 ds:00000000`00000000=??

Resetting default scope

 

LAST_CONTROL_TRANSFER: from fffff80002e7cb20 to fffff80002ed61c0

 

STACK_TEXT:

fffff880`0351b128 fffff800`02e7cb20 : 00000000`000000fc fffff880`03565180 80000001`db14e963 fffff880`0351b290 : nt!KeBugCheckEx

fffff880`0351b130 fffff800`02ed42ee : 00000000`0000000a fffffa80`069f6c98 fffff880`009ed180 fffffa80`06eeec18 : nt! ?? ::FNODOBFM::`string'+0x438fc

fffff880`0351b290 fffff880`03565180 : fffffa80`0dfdeb50 00000000`00000000 fffffa80`0676c3f0 00000002`00000000 : nt!KiPageFault+0x16e

fffff880`0351b428 fffffa80`0dfdeb50 : 00000000`00000000 fffffa80`0676c3f0 00000002`00000000 00000000`00000006 : 0xfffff880`03565180

fffff880`0351b430 00000000`00000000 : fffffa80`0676c3f0 00000002`00000000 00000000`00000006 00000000`00000000 : 0xfffffa80`0dfdeb50

 

 

STACK_COMMAND: kb

 

FOLLOWUP_IP:

nt! ?? ::FNODOBFM::`string'+438fc

fffff800`02e7cb20 cc int 3

 

SYMBOL_STACK_INDEX: 1

 

SYMBOL_NAME: nt! ?? ::FNODOBFM::`string'+438fc

 

FOLLOWUP_NAME: MachineOwner

 

MODULE_NAME: nt

 

IMAGE_NAME: ntkrnlmp.exe

 

DEBUG_FLR_IMAGE_TIMESTAMP: 4f76721c

 

FAILURE_BUCKET_ID: X64_0xFC_nt!_??_::FNODOBFM::_string_+438fc

 

BUCKET_ID: X64_0xFC_nt!_??_::FNODOBFM::_string_+438fc

 

Followup: MachineOwner

---------

 

Can anybody decode the above information for me? Thanks a lot!

 

After this crash now I further reduced the ram frequency to 1600.

[blackbelter]

Ok. I will reduce the CPU to 4.3Ghz. This is really painful.

[blackbelter]

Sounds like a RAM issue, or your CPU isnt stable at that overclock. Your Dump Files might only show what was running and failed first at the time when it crashed, not the root cause.

 

I see... That's the most sophisticated method that I know to identify the problem...

[Mugenjin]

Just put your CPU on stock speeds and see if it'll ever BSOD (which I doubt).

Pretty sure your OC isn't stable and not faulty RAM.

[mmaruda]

RAM should be set to what the producer recommends. For Ivy/Sandy Intels, you should not exceed a voltage of 1.5.

 

1.65 is max (not really safe) and anything above might fry you memory controller.

 

As for the CPU itself, BSOD means too low voltage usually. The auto OC is not good, since it either giver too much voltage or not enough. I run my 2500K @4,4 with a 1,3 voltage. The ASUS auto mode goes to 1.4V at 4,6Gh and it's stable in Prime and IntelBurn, but not in games.

 

OC is a tricky process, there are lots of ready to use recipes, but the sad truth is that each CPU is different, so the only way to do this right is the hard way - set the speed to what you want, set expected voltage, stress test with various tools and normal use. I you get a BSOD - increase the voltage by little and start testing again, remember to monitor temps. This is a long, boring and irritating process which might take days before you get what you need, but eventually it's worth it.

[pacotito]

Try using auto config for your ram and load the xmp. That will set the ram to the factory specified specs.

[Fleshpiston]

Prime is good but remember to run one instance for each core also try LinX

 

Good for benchmarking and stressing http://www.xtremesystems.org/forums/showthread.php?201670-LinX-A-simple-Linpack-interface

Edited May 18, 2012 by Fleshpiston

[blackbelter]

Thanks people for all the help! I found a guide specific to my motherboard. I am following it strictly, hoping to gain either a stable oced system, or at least some insight into how oc works.

[blackbelter]

Can I ask , did you install windows with this unstable overclock on ?

 

No, I didn't modify anything before the os installation. Thanks a lot for the info.