A Cry For Help.

Valiant23

Valiant23

Valiant23

Associate
Joined
6 Feb 2011
Posts
9
Some time around Christmas I bought some components from OC to build a system. It was based on a system I had already built for my sons which was put together with the help of OcUK staff, so I was pretty confident that the items I chose were good quality and would be a pretty good system.

This is what I bought on 18-12-2010;


  • Gigabyte GA-P55A-UD3 Intel P55 (Socket 1156) DDR 3 Mobo

  • Intel Core i3 540 3.06Ghz (Clarkdale) (Socket LGA 1156)
  • Arctic Cooling Freezer 7 Pro rev 2 CPU Cooler (Socket 939/AM2/AM3/775/1156/1366)

  • 2 x Corsair XMS3 4GB (2x2GB) DDR3 PC3-12800C9 1600Mhz Dual Channel Kit (CMX4GX3M2A1600C9) *8GB total
  • Gigabyte GeForce GTX 460 OC 768MB GDDR5 PCI-Express Graphics Card
  • Coarsair TX 650W ATX SLi Compliant Power Supply (CMPSU-650TXUK)
  • 4 x Sharkoon 120mm Silent Eagle SE Modular
This went into a NZXT Beta EVO Black Midi Case (bought elsewhere).

The only difference between this system and the other system that I built for my sons is that they had a different brand GTX 460 (I think it was an OcUK brand) and they had a SSD too.

Anyway I put it together with no problems and it is installed with Windows 7 Professional.

I'm not sure exactly when I first noticed the problem. It wasn't straight away but it was certainly after a couple of days.

Anyway what happens is that I have my PC on when I get up to when I go to bed. I switch it off when I go out, but sometimes I would leave it on if I was only going to be out for a short period of time (say an hour or so). Sometimes I would turn my PC on first thing in the morning, but not actually go to work on it until about 30 minutes after I had switched it on.

What would happen is that it would freeze and be totally unresponsive. The only way to solve this would be to power down, then boot up again. I'd get the message that windows didn't close down correctly but I received no other error messages or indication that anything was wrong. What I did notice was that when the screen froze I knew what time it happened because the clock display on the right of the toolbar would be stuck at the time it happened (eg the screen would show that it was 10.35 when the real time was say 10.47).

I have left my PC on for more than 24hours at a time when I was doing a lot of downloads as part of a Christmas present and during that time there was no freeze whatsoever, so it suggest to me that it may have been something when the PC is idle, so I disabled as many power saving features as I could find. Still no luck.

I have tried to solve this by process of elimination and because I have faith in the parts I have used I assumed it was a Windows or system error. I believed that there may have been a background scan that may have caused some conflict (why or how I don't know). To this end I put a stop to System Mechanic running automatic background scans and checks so that I do it manually now. This didn't work.

I posted my problem on Tomshardware forum with some success here and I thought I had it beat just by having the screensaver kick in. The last suggestion was to update my Mobo BIOS, but it came with the latest version available.

However on Friday I left my PC for a while and when I came back to it the screensaver was running. So far so good. I moved my mouse slightly and was about to start using my PC when it froze.

I've spoken to Simon on the counter at OcUk because my gut feeling is that it is a grphics card issue and he suggested stress testing (as well as posting up my problem on here), but I'm unsure how to go about it. To do it properly I'd assume that I need to stress test Memory(?), CPU (?), GPU(?) and Mobo(?).

I'll be happy to answer any questions and provide any info in order to get this sorted.

Thanks in advance.
 
I had a problem sort of similar to this, where i could run the pc continuously for days and pass every stress test. Yet, after shutting down and booting the next day i would often get lockups and errors. After a restart it would always be fine.

For me it turned out to be a RAM problem, where the motherboard was selecting timings out of spec for my RAM. As you have 8GB this could be even more important.
 
That's an essay of a post - always good to include all info you can, If I were you, if prime and other stability tests prove to be fine I would Be considering rma ing stuff. Its obviously a fault which has reason for parts to be replaced
 
Buchanan0204 said:
You never mentioned the hard drive you were using. Have you tested this?
Click to expand...

Actually no! (:rolleyes: @ myself).

Its a Samsung Spinpoint so I've d/l HUTIL.exe from the Samsung website. I can run the test inside Windows so will do that now - I'm watching the Super Bowl atm so will run Memtest after that and then post up my results.

Thanks to all.
 
Hi there,

If you can't get hold of Prime95 directly, then I would suggest using OCCT. This app uses prime within it - and has a nice interface and useful feedback available (like displaying graphing temps and voltages).

As for a GPU stress test - I would suggest Furmark. If your card can run this for a few minutes without issues then it should be fine.

May I ask, with your system - have you gone into the BIOS and set the memory frequency, voltage and timing manually - or have you left them at system defaults?
 
Hi mate thanks for your input.
:D
May I ask, with your system - have you gone into the BIOS and set the memory frequency, voltage and timing manually - or have you left them at system defaults?
Click to expand...
I haven't touched a thing in the BIOS.

I have had a look in there to see if anything was amiss but I'm pretty careful not to change settings unless I'm advised to.

I do plan to start 'playing' with my PC and getting into overclocking and generally expanding my knowledge base and these parts were picked with that in mind, but I haven't done anything about it yet.

I ran SeaTools diagnostic on my HDD last night as Windows 7 didn't like the the utility offered by Samsung. It reported no errors (HDD is a 1TB Samsung F3 I think).

Today I ran Memtest. Left it running for about 4 hours and it reported no errors (3 passes).

Next thing I'm going to do is try Furmark on my graphics card, then I'm going to run either OCCT or Prime95.

Question re Prime95:

I have read that for best results I need to leave it running for 12-24hours. :eek: Is this correct? I was planning on starting it just before I go to bed and then checking it/stopping it after 12 hours. Is this sufficient or should I go for the full 24hours?

Thanks again.
 
prime/memtest/occt are generally regarded as best to run for 12-24 hours, but most people just do it overnight, or maybe overnight + day at work. Problems with a CPU will tend to show up within an hour though, its RAM thats harder to detect.
 
Hi there, If you are running the RAM at default - then you 1.65V, 1600MHz rated RAM will most likely be running at 1.5V and 1066MHz. I can't say whether this is a part of your issues, but I would certainly go into the BIOS ASAP and set the RAM to the correct settings.

As for how long to prime the system - personally I think 12 hours is sufficient. Some people like to do it a bit longer - but I reckon a good 12 hours of priming should be enough. I would also suggest running Intel Burn Test as well - this test doesn't take nearly as long and is even more stressful - it is a great test to run before or after you have done the prime test as if your system passes both then you can be pretty confident that your CPU is fine.
 
just to add to what cmndr andi has suggested, you could check your sons system bios settings, paying particular attention the memory settings and voltage for memory, if was set up by OcUK, you will most likely find them running at the required voltage and timings, with you leaving everything on auto in your system bios it will most likely be lack of volts/settings for the memory modules
 
cmndr_andi said:
Hi there, If you are running the RAM at default - then you 1.65V, 1600MHz rated RAM will most likely be running at 1.5V and 1066MHz. I can't say whether this is a part of your issues, but I would certainly go into the BIOS ASAP and set the RAM to the correct settings.
Click to expand...

Oh mate you've opened a can of worms with this post (especially as I don't know how to take a screen grab of BIOS :o).

So, I went into the BIOS;

Selected Motherboard Intelligent Tweaker where I noticed in ADVANCED FREQUENCY SETTINGS that MEMORY FREQUENCY was 'grayed out' and had the value "1333". So I did nothing.

I then went into ADVANCED MEMORY SETTINGS. It said;

*EXTREME MEMORY PROFILE (X.M.P.) [DISABLED]
*SYSTEM MEMORY MULTIPLIER (SPD) [AUTO]
MEMORY FREQUENCY (Mhz) [1333]
*PERFORMANCE ENHANCE [STANDARD]
*DRAM TIMING SELECTABLE (SPD) [AUTO]
PROFILE DDR VOLTAGE 1.5V
PROFILE QPI VOLTAGE 1.1V
>>>>>Channel A
*Channel A Timing Settings [PRESS ENTER]
>>>>>Channel B
*Channel B Timing Settings [PRESS ENTER]

* = The only items that provided further options, none of which seemed relevant to anything you said, so I left them unchanged.

I then went into ADVANCED VOLTAGE SETTINGS and saw that "DRAM Voltage" was 1.500V and was set to [AUTO]. I chose that option and saw a list of various voltages. Although not the lowest, from 1.600 it went up to 1.620, 1.640, 1.660 which was in pink text. I saw much higher voltages 2.000+ and they were in red text. To be on the safe side I changed nothing.

By way of comparison I chose to load the "Optimised Defaults" from the main menu, and that is the current state of my BIOS. I did notice that the version of BIOS/Mobo driver was F11, which I believe is the latest version.

Apologies for the lengthy description - I'm trying to be as clear as possible.
------------------------------

This is the result of the Furmark Benchmark Test.
r78vvt.jpg


hyperst said:
just to add to what cmndr andi has suggested, you could check your sons system bios settings, paying particular attention the memory settings and voltage for memory, if was set up by OcUK, you will most likely find them running at the required voltage and timings, with you leaving everything on auto in your system bios it will most likely be lack of volts/settings for the memory modules
Click to expand...
Sons components were bought from OcUK but it wasn't an overclocked bundle. I used the same mobo/memory/processor so I'll check them while running CPU stress test.
 
turn on the xmp profile, this should set the RAM to at least 9.9.9.24 and 1600Mhz. Change the DRAM voltage to either 1.65v or if it doesnt have that, then the closest one.
 
Ch3m1c4L said:
turn on the xmp profile, this should set the RAM to at least 9.9.9.24 and 1600Mhz. Change the DRAM voltage to either 1.65v or if it doesnt have that, then the closest one.
Click to expand...

I did as you said and selected Profile1. This set RAM to 9.9.9.24, 1600Mhz and Voltage to 1.65. Thanks!

I've just ran OCCT for 8 hours and the test completed with no errors. There are four .PNG files that show this.

Does the fact that all components have been stress tested mean the system is stable and that this is a software issue?
:confused:
 
Apart from all the good advice above if that doesn't solve it grab the latest intel chipset drivers; there are some issues with harddrives coming back alive after the pc's sleep mode is triggered by windows.

Or if you can't be bothered doing that just turn sleep mode off in power settings.
 
Yea, since its now confirmed stable by that, id tend to lean towards it being a sleep issue. If you manually put it to sleep andwake it up, does it freeze? check your bios power options and see what standby/sleep mode it is set to, S1 and S3 is what you want (I think, been a while since ive looked). Other than that, check your intel drivers as mentioned above, maybe get the Intel RST ones on there as well, and then see if it still does it.
 
Thanks gents.

Firstly, I thought it was some kind of sleep issue at the very beginning. As such I disabled anything I could find that was remotely connected to power saving or sleeping/hibernating. I even managed to find some advanced options in Windows 7 that I haven't seen in other versions. I appreciate the advice but I'm pretty certain that I've disabled them already. One thing I haven't done is to try and force the issue to show itself by manually putting PC into sleep mode. I'll try that in a little while. Thank you.

Secondly, Chipsets and the like is starting to go into unknown territory for me, I'm afraid. However I'm a willing and eager (if somewhat simple) pupil and so I went to Intel's site and ran their Driver Update Utility here. This is what came back;

Graphics Driver
Product Detected NVIDIA GeForce GTX 460
Current Driver Installed 8.17.12.5856
This device is unknown or unsupported. Please contact the manufacturer for possible updates. More information>>

Intel Chipset Software Installation Utility (Chipset INF)
Product Detected
Current Version Installed 9.1.2.1007
This version is valid.

Audio Driver for Intel Desktop Board
Product Detected NVIDIA High Definition Audio
Current Driver Installed
This device is unknown or unsupported. Please contact the manufacturer for possible updates. More information>>

Wireless Networking
Product Detected D-Link AirPlus G DWL-G122 Wireless USB Adapter(rev.C) #4
Current Driver Installed 4.0.8.0
This device is unknown or unsupported. Please contact the manufacturer for possible updates. More information>>

Wired Networking (Newer Driver Available)
Product Detected Intel(R) Desktop Board D945GCLF2: Realtek* RTL8111C Gigabit Ethernet
Current Driver Installed 7.17.304.2010
Newer Driver Available: 7.027.0920.2010
Download Now
File Size: 5.622 MB
Download Time: Time @56Kbps:13.34min
Driver Description & Driver Documentation
Click to expand...
Does this mean I have the latest chipset? If not, then I'm sorry but I have to ask what it is I'm looking for.

Thirdly, I'll go into my BIOS and check the sleep settings there.

Thanks again.
 
If all else fails, you might want to try reinstalling the OS. It should make a difference, but it could just be a weird bug that occurred during installation or something... Especially as it seems there is absolutely nothing wrong with your hardware.

kd
 
Looks like a re-install is the only option.
:(

Yesterday I put the PC into sleep mode. It woke up perfectly. Put it into Hibernation mode. Same again.

After reading the last post here, with an agreement that it appears that my software is working perfectly I thought the tweaks suggested by you all would have made a difference, so last night I disabled my screen saver (the one workaround that seemed to prevent the problem occurring) and this morning I turned my PC on at 9.01am.

I left my PC alone until 10.25am. When I went to it, the clock was showing 9.59 and was totally unresponsive so I had to power it down by the case button.

I really appreciate all the good advice I've been given, and I've followed it all to the best of my ability. However it seems the only option left is, as the post above says, to re-install Windows.

The question is, what should I do if the problem remains after a fresh install?
 
You must log in or register to reply here.
Back
Top Bottom