Overclocking help i7 975

It not that hard to run MemTest...

It took ~15min. for me to run one pass (100% coverage) of MemTest while using 6GB memory.

You should test each kit by itself to help narrow down the identification of any possibly bad stick(s).

Run one instance for each core... Or if running HyperThread... One instance per thread.

Open Task Manager and divide up most of your available memory... Target 90% to 93% memory usage. I usually run 3 to 4 passes... (300%/400% coverage).

Back with an update.

I tried to run the memtest myself, I had about 4.3GBs available, I separated those into 8, as I went to turn on number 7 & 8, the CPU load hit 100% and the PC crashed.

I took it into the PC repair store today, got them to hook up a different PSU, was only another 500W tho, and it did last a little longer in Prime, maybe 2-3 more minutes, but still it ultimately shut down.

Its at the store now, getting memtested overnight, they updated me when they were closing, and so far no errors, and it was running for about 2 hours at that time.

Still could be any part, but I'm really hoping its not the motherboard, or more importantly the CPU...

If it was the PSU, would testing it on another 500W PSU have been enough? Considering this one also wasn't brand new, I assume they've tested it that its working well enough, but I'm not sure now that what I'm trying to do is exactly falling under regular PC repairs.

Thoughts? Whats it sounding like to you guys? Not sure I exactly trust the PC store lol I feel like I know more than them sometimes.

AJTheGamer said:

I tried to run the memtest myself, I had about 4.3GBs available, I separated those into 8, as I went to turn on number 7 & 8, the CPU load hit 100% and the PC crashed.

Click to expand...

Were you running/testing one kit (3x2GB) or two kits (6x2GB) at that time with MemTest?

Was only running the one kit (3x2GB) at the time with MemTest.

If both of your 3x2GB memory kits test without errors then it is most likely not the memory.

The repair shop's 500 watt PSU should be plenty to test your combo at stock speeds.

If your repair shop can get your combo to run one kit with their PSU at stock settings and no resets then it would seem to point to your PSU. If they can't then it may be something else... Motherboard/CPU?

Usually the way to pin something down like this is to isolate and test the parts individually... So install the cpu in another working setup and test... Or install the motherboard with other known working "good" parts and test.

The norm for X58 setups back in the day was 3x2GB... And your system should be able to run it with all default stock settings and be Prime stable without having resets.

Running with 6x2GB might require an additonal VTT voltage adjustment while running at stock settings if your BIOS auto rules happen to set the VTT voltage to low for your particular combo? With my board the auto rules default VTT voltage at stock setting to (+.000v) for both the DDR3-1066 and DDR3-1333 memory settings. However, when the memory is set to DDR3-1600... My BIOS's auto rules bump up vtt voltage by +.200v (~1.4v VTT).

When I recently attempted to dial in a 4.0 overclock on my EVGA X58 setup using 6x2GB Corsair 1600C8... I also experienced "hard resets" when stressing with Prime until I successfully dialed in my vcore and vtt voltages. The VTT setting for my combo when it stabilized was at +.100v (~1.3v).

On a side note... With my old EVGA motherboard I sometimes have trouble getting the board to fully recognize/train all dram sticks in BIOS and Windows after installing them into the motherboard slots. It seems to occur more often with 6x2GB but occasionally it will do it with 3x2GB. I have to actually swap the individual sticks out to different slots and reboot... Sometimes with multiple attempts to get it to finally work correctly.

I have some Corsair 1600C7 v2.2 sticks that won't seem to train correctly with my EVGA board while running 6x2GB? And yet some of my other sticks... Corsair 1600C8 v7.1 will successfully run in a 6x2GB configuration.

Screenshot of 6 sticks installed in my setup and only 4 sticks being recognized in BIOS/Windows. See Windows Task Manager... It shows just 8GB total and in BIOS it also reported the same amount:

I have a strong feeling its not the memory, I've never had really a problem in the slighest with the RAM, and just one of those feelings lol will find out shortly I hope.

The repair shops PSU SHOULD have been able to work, but if the problem is it hitting 500W or something, then the 500W wouldn't be enough right? Not that it should be getting near that....

So far with 1 reset to my knowledge with the shops PSU, and stock settings. I'll have to confirm with them the DRAM settings were back on default, as I had manually set them again since that seemed to work fine. (besides stress testing)

I think the shop should be able to isolate most of my PC, but with my CPU being socket 1366 I'm not sure how likely it will be they have a spare board, or CPU, to use to test either my board or CPU, and with the prices of these things being about the same as buying brand new (In my case, my board is practically impossible to find, and the CPU runs at $200+... I'd be better off investing in a new AMD CPU and board.. At least for the money.

I'm fairly certain my PC does need a little more VTT to run the whole 12GBs, but yeh, the 6GBs seems to be fine on stock (again besides stress testing) - I'll have to double check but I don't believe my BIOS auto sets the VTT when the RAM is installed above 1333 or 6GBs.

I think I had that issue with the board not recognising the RAM the other day, but its the one and only time I've noticed. It was weird, it said I had 14 GBs installed only 10GBs usable, I restarted and it told me I had 8GBs of DDR2....

Fiddled around with the sticks, and it sorted itself out.

So seems if the memory is fine, I'll have to play around and make sure its definitely not the memory not getting enough voltages, or the VTT not being raised enough for the changes.

At this stage I'm just hoping its something random like my HDD and not my CPU/Board, was they were the 2 parts I was hoping to avoid replacing at this point.

Thanks again for the very insightful reply Roger, helping me heaps in understanding what is happening, and a lot of the issues that are happening, as to why they might be happening, I don't feel so stupid about all this anymore, definitely don't feel smart, but I think I'm getting there.

EDIT - Not sure what the language rules are around here but ****ity **** **** ****. Apparently its the board or CPU, and they have no gear to check it. Memory was fine, HDD fine, GPU and all the rest causing no issues, leaving the board or CPU...

Guess I'm looking at new AMD builds

Well I'm back, after some more testing, and a new PSU, I believe I found the problem.

After installing the new PSU, the problem still occurred, I disconnected the fans and the DVD drive, it worked.

I reconnected and problem again, I took out the DVD, still problem...

I disconnect the fans, and it works again, so I'm down to it being the fans, I'm testing each individually now, but it would seem I'll be buying some new fans for my case at this rate.

Thanks for the help guys hopefully I'll be back shortly knowing which fan is the problem, can order some new ones (hopefully get some great suggestions from around here) and be ready to oc this thing soon.

Back with somewhat of an update.

Running Prime + RealTemp with optimized defaults works, BUT the CPU speed drops to x12 instead of x25, I was able to up the blck to 160 with x12 multi and run Prime for 10 minutes.

But when I turn the defaults back on, and turn my PPM function and Virtualization Technology off, the multi stays at x25 (3.33GHz) but crashes... I've tried upping some of the voltages but to no result

Any advice?

EDIT - Yep more testing, can't seem to get tbe bclock above 160...

I decided to try and keep the bclock at 160 and move onto my DRAM timings, I was able to set all 12GBs to 9-9-9-24 and run them at around 1300, raised the DRAM Voltage to 1.650 and the CPU VTT was at 1.23v, it loaded into windows but Prime ran into an error straight away, I raised my CPU VTT and windows failed to load?

I previously had this issue so I tested loading the 160 settings with everything on default/auto besides the block to 160, which runs fine, and raised the CPU VTT to 1.31v and it ran fine also, so where in the DRAM timings and CPU VTT together is causing this windows unable to load issue?

What should I be doing next? Any advice please, just when I thought I knew what was going on I run into another wall

I was able to raise my bclock to 178 with a multi of x12, with CPU VTT at 1.26v.

I proceeded to raise set the ram timings below specs, at around 1300 9-9-9-24, it worked, I then went on to start raising the multi.

I was able to get a bclock of 178 with a multi of x16 (around 2.8GHz) and ran Prime with no errors or restarts, when I raise the multi to x18, it restarts after about 2-3 minutes, I tried raising the CPU VID from AUTO to 1.3v, but it seemed to restart quicker, so I changed it to around 1.0250-1.0500v and it seem to run for around the same time as AUTO, am still testing, any advice on any other voltages I should try and adjust at this point?

It definitely restarts quicker with 1.3v over AUTO or 1.1v, Prime seems to finish the first pass with no errors I think its when it starts the '8k pass (?)' that it shuts down now...

With 6gbs (3x2GBs) installed, my settings right now are:

CPU Features/Power saving:

All off besides Intel HT Technology to Enabled and I left Execute Disable Bit to Enable. Rest is Disabled.

DRAM Timings:

All AUTO besides the 9-9-9-24 Manual timings.

Voltage Settings:

CPU VID: AUTO
DRAM Bus Volt: 1.605v
CPU VTT: 1.26v
CPU PLL: 1.80v
IOH: 1.31v

Everything else is on optimized defaults.

Clock/Multi:

CPU Multi: 12x (up to 16x tests on these settings - 18x causes restart)
Turbo: Disabled.
QPI: 18
CPU BCLK: 178
DRAM Frequency: 8 (1424MHz)
UnCore Frequency: 16 (2848MHz)

Click to expand...

Those are the settings I've been using up to a 16x multi works, but 18x causes it to crash. Prime has nothing in the results log, seems the 8k is as far as it gets tho.

On the few times I had errors when I was raising the BCLK it said:

FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.

Click to expand...

EDIT -

Event viewer events.

Critical said:

- System

- Provider

[ Name] Microsoft-Windows-Kernel-Power
[ Guid] {331C3B3A-2005-44C2-AC5E-77220C37D6B4}

EventID 41

Version 2

Level 1

Task 63

Opcode 0

Keywords 0x8000000000000002

- TimeCreated

[ SystemTime] 2014-08-31T10:07:48.711624900Z

EventRecordID 8624

Correlation

- Execution

[ ProcessID] 4
[ ThreadID] 8

Channel System

Computer AJ-PC

- Security

[ UserID] S-1-5-18


- EventData

BugcheckCode 0
BugcheckParameter1 0x0
BugcheckParameter2 0x0
BugcheckParameter3 0x0
BugcheckParameter4 0x0
SleepInProgress false
PowerButtonTimestamp 0

Click to expand...

Error said:

- System

- Provider

[ Name] Microsoft-Windows-WMI
[ Guid] {1edeee53-0afe-4609-b846-d8c0b2075b1f}
[ EventSourceName] WinMgmt

- EventID 10

[ Qualifiers] 49152

Version 0

Level 2

Task 0

Opcode 0

Keywords 0x80000000000000

- TimeCreated

[ SystemTime] 2014-08-31T11:56:50.000000000Z

EventRecordID 1967

Correlation

- Execution

[ ProcessID] 0
[ ThreadID] 0

Channel Application

Computer AJ-PC

Security


- EventData

//./root/CIMV2
SELECT * FROM __InstanceModificationEvent WITHIN 60 WHERE TargetInstance ISA "Win32_Processor" AND TargetInstance.LoadPercentage > 99
0x80041003

Click to expand...

Error said:

- System

- Provider

[ Name] EventLog

- EventID 6008

[ Qualifiers] 32768

Level 2

Task 0

Keywords 0x80000000000000

- TimeCreated

[ SystemTime] 2014-08-31T11:55:18.000000000Z

EventRecordID 9540

Channel System

Computer AJ-PC

Security


- EventData

9:42:07 PM
‎31/‎08/‎2014


28


DE07080000001F0015002A0007001502DE07080000001F000B002A0007001502600900003C000000010000006009000000000000B0040000010000008A000000


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

Binary data:


In Words

0000: 000807DE 001F0000 002A0015 02150007
0008: 000807DE 001F0000 002A000B 02150007
0010: 00000960 0000003C 00000001 00000960
0018: 00000000 000004B0 00000001 0000008A


In Bytes

0000: DE 07 08 00 00 00 1F 00 Þ.......
0008: 15 00 2A 00 07 00 15 02 ..*.....
0010: DE 07 08 00 00 00 1F 00 Þ.......
0018: 0B 00 2A 00 07 00 15 02 ..*.....
0020: 60 09 00 00 3C 00 00 00 `...<...
0028: 01 00 00 00 60 09 00 00 ....`...
0030: 00 00 00 00 B0 04 00 00 ....°...
0038: 01 00 00 00 8A 00 00 00 ....Š...

Click to expand...

Attached a file of the blue screen viewer.

Seems the higher I go on the CPU VID the quicker the PC shuts down...

Found this link and thought I'd try his settings, since he also got stuck at 178 BCLK, but at 1.45v my PC shuts down before I even get into Windows... 1.3V seems to shutdown about a minute into Prime, 1.1v takes about 3, so the less voltage the longer it lasts?

Seems I can't get over the stock speed of 3.33GHz

EDIT - Yep it doesn't seem to like me playing with the 'CPU VID' I change it and it wants to restart, even if the rest of the settings are stable.

I'll keep trying, hopefully someone has some advice for me, not sure what else to do at this point.