[Win7 64-bit] random BSOD

My story.
Scroll down for TL;DR version.
Alright, I'm new to these forums, I joined out of desperation.
Here's the problem, and I'm sure you've heard it before. After installing the newest ATI/AMD 10.12 display drivers my windows 7 64-bit machine won't boot properly. I can get into safe mode, but for some odd reason trying to uninstall the driver from there mucked up the install and made itself remove itself from the list. This in turn left many files behind, I then used driver sweeper (from here) to clean everything up, that worked as planned but upon logging into windows 7 under normal safe mode (not with networking or CMD) or just normal boot up causes a Blue screen of death, the error messages have varied as I changed things, from IRQL_NOT_LESS_OR_EQUAL to a recent one (after I uninstalled the HDMI audio from system>device manager and the video card driver) SYSTEM_SERVICE_FAULT (or something of that resemblence).

Now the thing here is, sometimes I can get into windows and have it crash within 10 minutes. I don't understand why it could be so buggy, I tried downloading as many updates for windows I could. As I am writing this I just got another IRQL_NOT_LESS-ect Now before today I ran my machine on a 400W power supply with the following specs,

Asus P5K SE/EPU
Intel Core 2 Duo E7300 @ 2.66mhz
2 Corsair 2GB Ram Sticks @ 800mhz
2 Geil 1GB Ram Sticks @ 1066mhz
1 Sata HDD
1 IDE HDD
Nvidia 9500GT GFX

I then replaced the nvidia graphics card (after removing the drivers via the uninstall in programs & features) and replaced it with an ATI HD 7550.
Now, I read the manual, it did say, use a 450W power supply but I thought, the least that could happen was it wouldn't run. It acctually ran fine, although it immediately bluescreened (I believe, I did not see it) after the inital reset after the install of the drivers. (10.12) I then ran just cause 2 in high quality and It ran beautifully, I then proceeded to test the card, see what it could do, I was excited (possibly my downfall?) then I went and played crysis: warhead, turned it up to gamer, fearing enthusiast might be a little too much for it. I then played the game for a good half hour, with a few stutters until the game locked up, and I'm not sure but it was just a stuttering sounds lockup so I reset my computer and left it for a few days.

I then put the old 9500GT back in and booted into my ubuntu partition which still had my Nvidia drivers (I never booted into windows with my old card again).

Today I went out and bought an Antec 902 Case and 750W Antec PSU. I connected all of these things up and ever since, I have had blue screens tormenting me. I have gone through all sorts of measures to attempt to un-install these drivers. I have noticed that LOWER resolutions used to keep the BSODs at bay, but right now nothing but blue screens here.


tl;dr
To cut a long story short, can anyone help me? I need to have my ATI 5770 working. I used 10.12 that gave me blue screens, I can only use safe mode.

Let it stand that I have had NO trouble with my computer until now.


p.s. sorry for the large block of text, I am going to swap my nvidia card in now to see if I can get into normal windows.

 

As stupid as it may sound I'd check my RAM latencies on BIOS and verify that those are not an issue by dropping them default +1T on tCL-tRCD-tRD-tRAS 1:1 DDR:FSB ratio 667-800Mhz. I doubt that the power supply ever was an issue. Boot in safe mode go drop all graphical drivers and Audio drivers down clean them up with DriverSweeper reboot to Safe mode go to device manager and 'scan for new devices' and then reboot to regular windows state.

After you get to regular windows state see that device manager has all GPUs listed as 'Standard VGA Display adapter' and then add in the 10.5a drivers or something with a
good affect on old hardware like 10.8 beta 6 which I keep repeating to people. Also 10.10e might be good, but you should start with 10.5a as it's known very good driver.

and when done hit command 'SFC /scannow' on command prompt so windows will verify that all windows files are intact and system drivers/modules are in good shape.

-edit-

This way you could basically test components 1 by 1, if an 'MemTest86+' would give you an error in first tests you would know that issue is broken RAM, if it would error on something like test 5 mostlikely issue would be on RAM or NB voltages and on something like Tests 4,6,7 issue would mostlikely be found on latencies (if testing on low latency highest speed).

As for GPU testing you should be able to boot on lower version drivers and verify that the GPU memory isn't bad (check for artifacts on something like Kombustor from MSI Afterburner).

Drop an PI or LinX on CPU a little see on large size links does the Northbridge fail while testing or small enough to fit on L2 cache for see, if CPU gives any problems and same time use soemthing like Lavalys everest (now days known as AIDA64) to see does your 12V PSU drop/raise over the 5% and same for 3.3v 5v which would tell you PSU would be culprit.

and so on..

 

Thanks for the speedy reply TwL! Weird thing is, when I was in safe mode uninstalling the display driver, the 5770 was listed under unknown devices but as a standard VGA adapter. But I havent tried scanning for new hardware, I just threw the 9500GT in then and it booted up fine, I'll try removing any trace of ATI drivers. Scratch that, I just got IRQL_LESS_-ect again on the OLD card. I need to get rid of these ATI drivers.

 

I didn't understand your last paragraph, can you explain in detail please?

 

I just mean that only way to test hardware issues this kind 'IRQ_NOT_LESS_OR_EQUAL' is like microsoft would be saying to you "We don't have a clue what's wrong, but something is". Would say there's some serious issue in somewhere and usually this issue appears under memory controller where you have ful memory controller full of blocks and that stresses the controller so much that some of people can't run memory at specified speed.

This create several possibilities of error location and I was just explaining how you could see where the issue is testing components 1-by-1 to identify what voltage you need to pump or drop or what is the cause of board unstability in general.

 

Ah alright then, I'll try dropping the two corsair sticks and see if that works considering they are the newest.

I just got an error I had received earlier, it's system_server_exception.
maybe that could be more related to the issue?

 

These benchmarks can tell you several things examples:

* MemTest86+ Test 5 = it's an move block test inside memory, if it errors usually the current DRAM voltage used for memory isn't enough.
* MemTest86+ Tests 4,6,7 = are through memory tests which passes checksum and so these are good to identify if the blocks are too tight latencies the checksum fill fail. So, drop the latencies.
* MemTest86+ Tests 1-3 = If memory is bad usually the first sign you see here on quick easy memory testing where some area of block simply won't pass and 9 out of 10 there's an true physical issue with the block (where with new MemTest86+ you can identify the block even).
* IBT / LinX (LinPack) - Small blocks = when you push something like 150-250 times small block through like 4-8MB which fits to L2/L3 cache it will tell you that CPU might have an issue on it's caching at least with current voltage or power safe functions.
* IBT / LinX (LinPack) - Large block usuall tests through Northbridge to memory and CPU of course, but first sign of error usually is because of northbridge is unstable or CPU is unstable depending on BSOD it drops you of course (considering, if MemTest is clean).
* PI testing would be better for CPU testing perhaps as that won't crash the whole box when it fails.
* Kombustor (or any testing which uses full GPU and memory) = On artifacting on stock setting usually GPU Memory issues. Unless specifically tested on stock to pass and then overclocked GPU core to test.

and so on.. There's alot of tools but you only need few to determ whole a lot of specific hardware errors. Some of the people believes that every test should have their own tools some even believes there's 1 tool to everything, but I haven't found that very reliable in the end.

 

The thing I don't understand is, I can boot into my linux(ubuntu) partition fine, no worries and go play around, so I assume its more driver related.

 

I doubt this there's always something on drivers linux loads a bit different style things and also uses everything very differently. Like hard drive caching on software RAID.

Still the drivers used works for other people and not on your hardware would speak against the idea of problem being at drivers in windows OS.

 

Ah but there you're wrong, if you search Blue screen of death and ATI and HD 5xxx you come up with a world of results. Although I may be wrong and these problems may not be related (not that I'm not grateful for your help, this is just chatter while I wait for things I am doing to complete).

 

First thing I did. There aren't any restore points before I installed the driver, which is strange since I've had my system running for a good while.

 

heh, yeah you can google any hardware in this world and you'll find million BSODs. The real question is how do you deal with it so that personal hardware works. Either you settle and say "hey, this sucks corporations should build better crap" or you test through hardware and know what is broken or doesn't process through correctly or trust on other people testing and search lifetime of answers on forums like this which may never be same on 'this specific hardware'.

ahh, todays lovely world nothing ever works as planned to. :bang:

 

you truely are wise

 

Alright, the ATI graphics card now displays as a "Standard VGA Graphics Adapter"
Jumping out of safe mode now.

No BSOD yet, usually it occurs directly on login to a normal session.

spoke too soon...again...got a blue screen, going to swap ram sticks around.

 

Well, that's hard to tell, between installing the card and installing the drivers there was about a 5 minute window for any BSOD-ing to occur and that didn't happen so it seems it may be related to the driver.

and sorry, no, only one PCI-E

 

I had the same problem 1 month ago, I got BSODs with various erors in various situations, the problem was in memory that was seted at 7-7-7-18, now my system works on 7-8-7-18 without BSOD's.
Hope will help....

 

All of my ram is set to auto, I know it isn't optimal but at least it doesn't break anything.

 

Then, try to update M/B BIOS