[Solved] Core i5 System - Strange Storage/Networking Issue

[weshy]

Hi All

First, I'll say that I am pretty experienced with PC's and troubleshooting problems. Now on to the Issue!

System Specs
i5 750 (B1 Stepping)
Gigabyte P55-UD3R (Various Bios, current F3m)
Gskill 2 x 2GB DDR3 [Trident] F3-16000CL9D-4GBTD
Nvidia 7800GTX
Gskill Falcon SSD 128gb (F/W: 1571)
WD Caviar 500gb SATA
LG 16x IDE DVD/RW
Antec Signature 650w
Windows 7 x64

Symptoms

1. The system intermittently locks up, sometimes going to a black screen, other times with the screen freezing as is and sound repeating.
   
2. It is readily repeatable if I transfer a large amount of data over the network (e.g. 60gb) and wait. The issue is exacerbated if I am also playing an x264 video file (with sound), as the lockups tend to occur more rapidly .
   
3. Initally it seemed the lockups also occurred randomly, but am now pretty sure that any random lockup is related to the above issue.
   
4. The system completes Intel Burn Test on High without issue (ran it 3 or 4 times in a row). It also ran Prime 95 on all 4 cores for a whole day (at least 8 hours, closer to 12 hours from memory). Memtest also runs without error for 12 hours.
   
5. The lockup occurs irrespective of whether I am using the Jmicron controller or the Intel controller and irrespective of what driver I am using. The lockups occur more rapidly when in SATA mode on the Intel, but happens in IDE mode as well.
   
6. I have been able to install Windows 7 without issue, save for one time where the system locked up 64% transfer (I have been installing Windows 7 via an SD card. Originally I was using the Install DVD and the lockups still occurred).

Possible Causes
Based on the above, I'm comfortable ruling out the CPU and RAM as having any cause. I am also comfortable ruling out drivers as a cause of issue, as the lockups happen regardless of what LAN driver or SATA driver I use.

I am also reasonably comfortable ruling out the WD HDD (has been working fine in a previous PC for a long time) and the Antec 650w PSU (Comes pre tested and also went through the stress tests without issue).

Therefore, I think that the culprit is either of (in most likely to least likely order):

1. Motherboard - specifically the link between the chipset and the LAN and/or sound.
2. Video Card
3. SSD Drive

Any Thoughts
I've been wracking my brain over this problem any time I have had to spare (not that much) for almost a week now. I've found it quite difficult to narrow down the culprit due to the initially random nature of lockup.

I have a friend who is buying a very similar system and will use his parts on Tuesday and Wednesday to hopefully get to the bottom of it!

But in the interim, has anyone else got some suggestions? Has anyone else come across this problem?

[punkerdaleski]

hi weshy!
I'm having the same problem on almost the same system;

core i5 750
Gigabyte P55-UD4P
Gskill 2 x 2GB DDR3 [Trident] F3-16000CL9D-4GBTD
7600GT
Enermax Modu82+
Win7 Ult 64bit

Random lockups where the screen just freezes.
Once only while installing from DVD, and maybe 10+ times while in windows
Happened with in a few mins if the gigabyte easytune6 program was running...

[Mitch01]

SATA Hard Drive
IDE DVDRW

Ensure you have the Win7 x64 (NOT VISTA) drivers for EVERYTHING (that includes the IDE/SATA Controllers).

My money is on either the drivers not being installed correctly for the controller (and yes Windows 7 has had a few issues with controllers and drivers already), or the BIOS not liking an IDE DVDRW and SATA Hard Drive.

That can be tested by disconnecting the optical drive and testing the machine.

Also download and use OCCT for testing. I have found it brilliant.

[FragU]

Try disabling sound and test again.

[punkerdaleski]

fixed for me i think! Cheers!
Didnt have the SATA RAID drivers installed (not using raid), but looks like they are needed for AHCI
woot!

[weshy]

Originally Posted by punkerdaleski

fixed for me i think! Cheers!
Didnt have the SATA RAID drivers installed (not using raid), but looks like they are needed for AHCI
woot!

Hi Punkerdaleski.

I hope that the drivers you installed resolve your problems. Something to note is that when SATA control is set to AHCI mode in the bios, it seems that WIndows 7 installs the AHCI driver and the IDE driver and then recognises the SSD and IDE drives as multi-word dma mode 2 devices. This does not occur when you set the SATA mode to RAID in the bios - there is no longer any mention of multi word DMA in the bios.

Also note that I can now readily reproduce my issues.

Originally Posted by Mitch01

SATA Hard Drive
IDE DVDRW

Ensure you have the Win7 x64 (NOT VISTA) drivers for EVERYTHING (that includes the IDE/SATA Controllers).

My money is on either the drivers not being installed correctly for the controller (and yes Windows 7 has had a few issues with controllers and drivers already), or the BIOS not liking an IDE DVDRW and SATA Hard Drive.

That can be tested by disconnecting the optical drive and testing the machine.

Also download and use OCCT for testing. I have found it brilliant.

Hi Mitch

Yep have made sure that I have only win7 x64 drivers and that made no difference. I have also tried with the DVD not plugged in, that made no difference either. I have tried in IDE, AHCI and RAID modes using the default windows 7 x64 drivers, the ("non-compliant") 8.8.1009 Intel matrix drivers and the new 8.9.1023 intel matrix drivers. I have tried with and without the Intel chipset drivers installed.

I have used 3 different versions of the realtek LAN driver - the base Windows 7 x64 driver, the WIndows Update driver (Which I note does not work at all) and the updted driver on the gigabyte website.

I'll have a look at OCCT.

Originally Posted by FragU

Try disabling sound and test again.

Hi Fragu

That is what I am about to do now.


Weshy.

[weshy]

Disabled sound in Bios, turned on RAID mode in Bios (drives not in raid, but this is the Intel recommended method to enable ahci).

Reinstalled Windows 7 x64, updated to latest Intel chipset drivers and then latest Intel Matrix drivers.

Started my standard network test by moving 60odd gb of files to server while playing a x264 video. Also was trying to update Avira (fyi crash still happens when there is no a/v) Crash still happened.

Restarted the machine. Started same test while deleting files that were transferred before (No x264 video).

Was typing this reply and crashed again...

Am now going to perform hdd scan on the data hdd, just to be sure.

I've been troubleshooting pc's for almost 10 years now and have never had a problem as frustrating as this! Can't wait to test my parts on my friend's board.

[karlcloudy]

Yeah, it is quite annoying when issues like this occur. I was getting random lockups on my system, although I think they were truely random (apart from logging into skype). I eventually worked out it was the motherboard not liking the dual channel RAM (incompatible with board). Try using one stick and see. It's a long shot, but since you can reproduce the problem very easily, it's worth a try

[Mitch01]

Try installing XP or Vista if you have them, out of interest.

That'll tell you if it's a driver conflict/issue with Windows 7.

[terrastrife]

increase your nbcore/p55 voltage.

i have extremley similar problems with my totally different setup. i have to run 1.5v nbcore just to get the system stable ( i can prime or linpack indefinitley but as soon as i start using my network or hdds itll stuff up), let alone stable at 445MHz fsb (which takes 1.6v)

(X38 default is 1.25v)

[weshy]

Morning all

A quick thanks for all your help so far!

I ran WDD Diagnostic tools on the 500gb caviar drive last night (quick test first and then thorough) and the drive passed all tests.

Originally Posted by terrastrife

increase your nbcore/p55 voltage.

i have extremley similar problems with my totally different setup. i have to run 1.5v nbcore just to get the system stable ( i can prime or linpack indefinitley but as soon as i start using my network or hdds itll stuff up), let alone stable at 445MHz fsb (which takes 1.6v)

(X38 default is 1.25v)

terrastrife, I will try that next when I get home. But surely needing to increase ndcore voltage on a stock non-oc setup just to get a stable system is sufficient reason to RMA a board (that is brand new)?

Originally Posted by Mitch01

Try installing XP or Vista if you have them, out of interest.

That'll tell you if it's a driver conflict/issue with Windows 7.

Mitch, if the nbcore voltage increase does not work and I don't have my friend's board to try out tonight, I'll have to try *sigh* Vista x64 next I guess.

[OhFoRkMe]

Would it be worthwhile trying a Linux BootCD and see if you still have issues transferring the data?

[weshy]

Hi all

I tried to transfer and delete files via the network in Windows 7 safe mode...and it was 100% fine. This now suggests that there is nothing wrong with the onboard networking and hard disk.

Now, what are the differences between safe mode and normal?
From what I can see, it is that the Windows action centre, video and sound drivers are not loaded.

Could all these lockups possibly be because of a suddenly faulty video card? I am going to borrow a friend's 7600gt to test with tonight (if I have time).

Failing that, Vista x64 is next up to bat.

Originally Posted by OhFoRkMe

Would it be worthwhile trying a Linux BootCD and see if you still have issues transferring the data?

I can see the merit in doing so, but in light of the above in safe mode, I'm not sure what more that will now add.

[Aetherone]

Have you tried a different NIC?
It might not be the that the onboard is faulty, but drivers conflicting between it and something not loaded in safe mode.

Plus

Originally Posted by weshy

realtek

Isn't a good thing. Realtek were a cheap and nasty network chipset that eventually mostly came good, but the last couple of years (esp on GB boards) their integrated PCI-e chips have gone back to the ways of old.

[weshy]

Hi all

Tried a new graphics card last night, same problem. With no spare NIC, so decided to try Vista out on an old hdd (I did not want to keep using the SSD for this).

Plugged in the old hdd and tried to install Vista. The 1st time the install would hang, though the system was still responding. On the 2nd, the system crashed. Upon restart, the system would not go past the bios "welcome" screen and now continually restarts after getting to the bios screen, even after a cmos clear or 2 and unplugging everything except for RAM, CPU and video.

I am getting a friend's motherboard today, which should confirm it is just my motherboard that was faulty.

The only other option now is a faulty PSU, which given the calibre and pre-testing performed on the Signature 650w, would be hard to believe.

Thanks again everyone for all your thoughts and suggestions! I will post a (hopefully) final update when the system is up and running.

Originally Posted by Aetherone

Have you tried a different NIC?
It might not be the that the onboard is faulty, but drivers conflicting between it and something not loaded in safe mode.

Plus

Isn't a good thing. Realtek were a cheap and nasty network chipset that eventually mostly came good, but the last couple of years (esp on GB boards) their integrated PCI-e chips have gone back to the ways of old.

Aetherone, I remember the days of those oh so dodgy realteks. However, they have gotten better since then. As to compatability issues, the difficulty I had with that theory is that the components in this system are common for an i5 build and I could not find one other instance of such problems. Surely a compatibility problem would result in widespread uproar?