Overhauled rig misbehaving, how to proceed?

wakou222 · 7 Jan 2015 at 11:29

Hi guys, I recently bought a fairly decent PSU, a Cooler Master 550W jobbie. I am very pleased with it. BUT.......
I have done a quite major overhaul of the whole PC since Xmas, including new Graphics card, (changing from nVidia to Ati), new case, new CPU cooler and a new HDD.
I usually use a linux OS, but I have MSWIN 10 tech preview. Since the rebuild I have been having some problems.....
When I use Windows, it often crashes/blue screens with a variety of messages given, none of which are very informative.
OTOH I ran Prime95 at torture settings CPU/RAM for 5 hours yesterday without issue, until I stopped the test. When it promptly crashed.
I have also run a quite intensive game for a couple of hours, which behaved perfectly, until I stopped playing. It then crashed.

All this would not perhaps be surprising, as I am using a beta/pre-release OS, but IF I start my usual OS straight after a Windows crash, it also fails to start. My usual OS (openSuse Linux) is normally rock-solid.

My first guess is that something is getting hot. or is somehow not behaving WHEN IT IS COOLING DOWN (?)

It is not the CPU, I have the new fan running at MAX, the CPU will not go above ~52C even during the stress test.

Temperatures reported from Mainboard and GPU and HDD's are also very cool indeed.

It could be a memory DIMM moved or dislodged whilst rebuilding (I hope it is!) Easy to test and resolve.

I have done the windows memory diagnostic, no problems were found, I also ran MemTest for a while, but have not done a full overnight one dimm-at-a-time stress suite, as we all know that is a quite annoying task (and my PC is quite noisy ATM !)
Also I guess a cable slightly not seated etc. But this would usually show up in normal operation of the openSuse Linux install as well as the Windows one.

The new HDD only has the Windows 10 on it, the Linux install is on a separate SSD.

If it were not for the fact that my linux install also misbehaves AFTER a windows crash, I would put it down to drivers, the Windows install itself etc etc.

So there is a possibility that the new PSU is somehow not working correctly. This would be ENORMOUSLY difficult for me to pin-point...And even if I did decide it was the culprit, it might well behave perfectly on a test bench or in another rig(?) (!!)

Any tips or advice please? How to proceed?
Any other possibilities?

I have the old PSU, but it is a cheapo and underpowered for the new GPU. Are there any ways to test a PSU? I have lm_sensors on Linux and SpeedFan on windows, all voltages are in range. I suppose it is possble to poll/log those voltages using SpeedFan?

FYI, here are some of the Blue-screen messages I have had:

DRIVER_IRQL_LESS_OR_NOT_EQUAL
IRQL_LESS_OR_NOT_EQUAL
KERNEL_SECURITY_CHECK_FAILED
APC_INDEX_MISMATCH
ATTEMPTED_WRITE_TO READ_ONLY_MEMORY
KMODE_EXCEPTION_NOT_HANDLED

I really hope it is NOT the PSU, as I am very pleased with it, and it would be a MAJOR PITA to change it!

implosiveturnip · 7 Jan 2015 at 11:47

Strip your system down to the bare minimum (ie on board gpu, one hdd one stick of RAM and reset BIOS/remove any overclocks) and test. If it works ok, and one more component and test again. Rinse and repeat until you find the problem component.

Also, could you list your full system spec to offer us a bit more insight?

wakou222 · 7 Jan 2015 at 12:44

implosiveturnip said:
Strip your system down to the bare minimum (ie on board gpu, one hdd one stick of RAM and reset BIOS/remove any overclocks) and test. If it works ok, and one more component and test again. Rinse and repeat until you find the problem component.

Also, could you list your full system spec to offer us a bit more insight?

Ty Turnip, wise advice.....
Luckily I quite 'enjoy' this sort of thing!

I guess it might be also a proper thing to begin by a DIMM-by-DIMM memtest, then in pairs then all four..

It has been a long time since I worked with MSWin. Will it be OK with changing the HW in this way? So I do boot to safe-mode (is that still f8 with > Win 8.1?) Then uninstall graphics (Ati Radeon) driver. Shut down, pull card, reboot (to safe?) install nVidia (for the IGP) reboot...
..And hope

As for system specs, Here is a brief output from inxi, shout up if you need more deets...

Code:

System:    Host: linux.site Kernel: 3.16.7-7-desktop x86_64 (64 bit) 
           Desktop KDE 4.14.3 Distro: /etc/SuSE-release corrupted, use -% to override
Machine:   Mobo: ASUSTeK model: M2N-MX version: Rev 1.xx Bios: American Megatrends version: 1004 date: 03/26/2008
CPU:       Dual core AMD Athlon 64 X2 5600+ (-MCP-) cache: 1024 KB flags: (lm nx sse sse2 sse3 svm) 
           Clock Speeds: 1: 1000.00 MHz 2: 1000.00 MHz
Graphics:  Card: Advanced Micro Devices [AMD/ATI] Curacao PRO [Radeon R9 270] 
           X.Org: 1.16.1 driver: fglrx Resolution: [email protected] 
           GLX Renderer: AMD Radeon R9 200 Series GLX Version: 4.4.13283 - CPC 14.501.1003
Audio:     Card-1: NVIDIA MCP61 High Definition Audio driver: snd_hda_intel Sound: ALSA ver: k3.16.7-7-desktop
           Card-2: Advanced Micro Devices [AMD/ATI] Cape Verde/Pitcairn HDMI Audio [Radeon HD 7700/7800 Series] driver: snd_hda_intel
Network:   Card: NVIDIA MCP61 Ethernet driver: forcedeth 
           IF: enp0s7 state: up speed: 100 Mbps duplex: full mac: 00:1a:92:07:38:62
Drives:    HDD Total Size: 2428.5GB (1.9% used) 1: /dev/sda SAMSUNG_SSD_830 128.0GB 
           2: /dev/sdb TOSHIBA_DT01ACA2 2000.4GB 3: /dev/sdc Maxtor_7L300S0 300.1GB 
Partition: ID: / size: 30G used: 13G (45%) fs: ext4 ID: /home size: 84G used: 31G (38%) fs: ext4 
           ID: swap-1 size: 4.30GB used: 0.00GB (0%) fs: swap 
Sensors:   System Temperatures: cpu: 27.0C mobo: 27.0C gpu: 30.00C 
           Fan Speeds (in rpm): cpu: 2020 sys-1: 16875 
Info:      Processes: 164 Uptime: 0:07 Memory: 625.5/3961.4MB Client: Shell inxi: 1.7.24

I just noticed two little anomalies in that list:
1) the distro says '-release corrupted', this is because I use some beta packages, not an issue here
2) Fan 2 reported at stupid RPM, I have the fans rigged unusually ATM, I am 99% certain this is not an issue either, the CPU is being cooled by a new cooler running at full-speed 2250 RPM

varkanoid · 7 Jan 2015 at 15:40

Have you checked out the error code when it crashes ?

wakou222 · 7 Jan 2015 at 16:38

varkanoid said:
Have you checked out the error code when it crashes ?

Hmm! There were so many, and this is after all a pre-release/beta OS, that, no I did not. I am sure I could dig them up from event manager or other Windows tools, but I am not a windows expert, and have found in the past that little useful info can be had from the 0ax000001 type codes. I would be happy to post them to a pastebin if you could get some info from them? And even the *.dmp files, if you know how to 'read' them? I have launched event manager after a crash, and clicked for 'more info' but usually nothing useful is returned. The 'system' works, and clicking launches an instance of MSIE, and 'asks' what the error means, but the answers are either not there, or are above my level...