Jump to content
Corsair Community

Help required. Out of ideas!


SoDaR2005

Recommended Posts

Hello,

 

Well 2 days ago i upgraded to a nice little PCI-Express Motherboard and PCI-E card from my older AGP setup just to futureproof myself a little and to get the latest chipset. Getting straight to the point, all was well i re-installed windows fine everything was looking in neat and tidy order until i installed a driver that did'nt work. Well the driver was an MSI Driver for my GFX Card called "3D! Turbo Experience" it refused to work upon completing the installtion of the driver, so i uninstalled it as you would. At this particular point i was finished with my re-installation and i decided to run SiSoftSanda for a little benchmarking.

 

The computer locked up, crashed just as i clicked the little "!" in the quicklauch bar, so i had to resent the computer using the reset switch. To my complete amazement it would not let me log back into windows, it said something along the lines of - "The Config file has been damaged, please run repair from the WinXP CD by pressing "R" at the appropriate screen". I did this and "Repair" did nothing for me, so i went into Safe Mode, backed up my stuff again and proceeded to re-install WindowsXP again. I was not too bothered it was probably caused by the MSI Driver and i decided not to use it again and put the CD away with it on.

 

Anyway, it was Formatting C:\ into NTFS File System and it got to 50% or so complete and for some unknown reason i got a BSOD (Blue Screen of Death) which stated "IRQL_NOT_LESS_OR_EQUAL" and some numbers which are complete jibberish to me. So i resetted the computer and did it again, this time i got a BSOD at 49% complete and i "think" it was stating the same error. So, seeing as i had trouble with the Audigy 2 on my Gigabyte K8NSNXP-939 (POOL Error of some sort, Souncard could be damaged) i took it out and cleared the CMOS.

 

Eventually, it Formatted i thought at this point i had cracked the problem with other BSOD's i encountered with "IRQL_NOT_LESS_OR_EQUAL" on my older nForce 3 Gigabyte setup. So i proceeded on my way with the installation again, i did everything again and it was looking great. All of a sudden the Computer rebooted for no apparent reason JUST like it used to occasionally on my Gigabyte setup. So i went into "System and Recovery" (under properties My Computer) and unticked "Automatically Reboot". Now i had to play the waiting game to get a BSOD that night ... i never got one so i went to bed and left it at that.

 

Next morning i needed to copy my VALVE Folder to my C:\ Drive so i could get my Steam games up and running so i decided to kick start that going in the morning and i left it as i went to work. I got home from work 8 hours later to descover my brother was on the computer and i asked him was my User Account still logged in when you went on it? He said there was a BSOD on screen but he forgot what error it displayed as he is not that computer orientated and didn't write it down. The computer has locked up on him once when he was playing a game of NHL 2005 and he had to press the reset switch.

 

Ok so this is very, very annoying now i still get practically the same crashes and BSOD's, if not worse, than that of my older NF3/AGP setup. Like i said the GFX Card and Motherboard are 2 days old so what can it be? I may have eliminated the Motherboard, Graphics Card and Soundcard (no longer using as i use nVidia Soundstorm now). The only other two or maybe three options now is MEMORY, Processor or Hard Disk Drive and i think it's bad memory modules quite personally.

 

Ok so my memory is tested at what? 2.75v by Corsair, granted my DFI is running it at 2.65v but i will TRY bumping the voltage to 2.80v see if that resolves anything - thats a thing to try. Also another weird problem is on the Gigabyte (older setup) i ran memtest to over 400% (Gigabyte only used a +0.2v increment meaning the Memory would be at about 2.65v, i would have needed a +0.3v to get it to about 2.80v) coverage and i have done the same with the DFI one - NO ERRORS, apparently. I just don't understand it IRQL errors is associated with RAM and i WILL take the Memory back to Corsair as i want to test everything that is not new in my system that may cause the issue.

 

What does anyone suggest ... before i request an RMA from Corsair i was wondering if anyone has any suggestions or ideas i can use or try to resolve these issues which were caused on both the OLD and NEW setups. Please do not use technical Jargon i don't want to keep pestering people asking what this and that means thanks :D: . Also another piece in the puzzle i found was the Seller i got the memory modules from (back in late December 2004/Early January 2005) have had many complaints about selling bad components like Ram, Motherboards and even Graphics Cards as of late. However, they used to be very trusted but now it seems they have gone down in quality a little.

 

I hope someone can help, im well and truly stumped! :(:

 

Kind Regards,

 

Paul Dale.

Link to comment
Share on other sites

Like i said the GFX Card and Motherboard are 2 days old so what can it be? I may have eliminated the Motherboard, Graphics Card and Soundcard (no longer using as i use nVidia Soundstorm now). The only other two or maybe three options now is MEMORY, Processor or Hard Disk Drive and i think it's bad memory modules quite personally.

Do not eliminate the motherboard and video card from the list just because they are new.

 

 

I just don't understand it IRQL errors is associated with RAM and i WILL take the Memory back to Corsair as i want to test everything that is not new in my system that may cause the issue.

http://www.pcformat.co.uk/tips/default.asp?pagetypeid=2&articleid=16167&subsectionid=688&subsubsectionid=488

 

IRQL_NOT_LESS_OR_EQUAL stop errors are caused by software not hardware. When Windows is in the middle of scheduling an execution thread, it puts the processor at an Interrupt ReQuest Level (or IRQL) of ‘Dispatch’. This blocks further software interrupts from the scheduler until the process is complete, and no one is allowed to do anything that requires the scheduler. Unfortunately, accessing a memory address in virtual memory is such an activity, because when a process generates a VM page fault, it is usually suspended while the slow hard disk gets around to delivering the requested page. The upshot of this is that drivers (and everything else but especially drivers) are restricted to the contents of physical memory while the IRQL is at dispatch. If anything breaks this rule (say because a corrupted pointer tries to access a random memory address), XP flings up the blue screen and shuts that bad boy down pronto.

 

Anyone still awake? No? Good. Because, embarrassingly, the answer is to find out which driver is responsible and replace it.

In otherwords, IRQL errors don't mean that it's the physical memory, just that it's a problem with how the memory Windows sees (including virtual memory) is managed.

 

If it in fact IS a HW issue, personally I'd lean towards the memory controller on the CPU having issues. Could be a socket or heat issue as well. I'd first remove the partitions and re-do only windows with minimal HW in the system.

Link to comment
Share on other sites

Do not eliminate the motherboard and video card from the list just because they are new.

 

 

 

http://www.pcformat.co.uk/tips/default.asp?pagetypeid=2&articleid=16167&subsectionid=688&subsubsectionid=488

 

 

In otherwords, IRQL errors don't mean that it's the physical memory, just that it's a problem with how the memory Windows sees (including virtual memory) is managed.

 

If it in fact IS a HW issue, personally I'd lean towards the memory controller on the CPU having issues. Could be a socket or heat issue as well. I'd first remove the partitions and re-do only windows with minimal HW in the system.

 

Hmm i see, well i will backup my stuff on CD today and completely delete all partitions and like you said try installing windows with minimal hardware in. I will have to use CPU, GFX, 1 Memory Stick, HDD and CD Drive. Everything else can be left out of the system at that point. Also before doing so i will bump the voltage on the memory to 2.80v just to partially eliminate that it isn't the voltage. Also what i will do is make a Ghost Image of Windows XP and if it is indeed a Software issue i can roll back in less than 5 minutes and try again rather than installing from a new.

 

The thing i don't understand if it is a Software issue is why did i get the IRQL Error when i was no in windows but simply formatting the HDD with the Windows XP CD where virtually no software would be used? Also i have posted this on another forum and i get the reply as my Power Supply Unit again. How many more PSU's do i ****ing need to run a PC, christ i have bought nearly 6 PSU's in 5 months. This one seems to be the better one as it's a 420w Tagan and has 20/24pin and other goodies.

 

I will try this but are there any other suggestions please? Thanks!

Link to comment
Share on other sites

Well the DFI mobo's especially your type ( the non SLI type ULTRA-D) are known to have major memory issues, I would first try to do a bios upgrade ( do it with one working module ! ), then setup your ram accordingly (cas,...) Don't OC yet. See that the correct voltages are set for Ram,... and then install from scratch. Also the newer powersuppy will help your rig to be more stable, waht are the amps at 12 volt ?, You need to have a good 20 amps at 12 volt to get it smooth sailing.

Here's the link for the bios update :

 

http://www.dfi.com.tw/Upload/BIOS/NF4LD310.exe

 

Read how to do it first if you haven't done this before ask a mate do it for you , you could permanently damage the board !!

 

And as Write suggested because the components are new doesn't mean thay are working 100%. Many of them have little bugs that are resolved within a few months after it's release.

 

Here's a link at Asusbozards forum , their main Oc'ers are using DFI boards to and they had some issues with them here's the link , read first post from Sierra

 

http://www.abxzone.com/forums/showthread.php?t=88699&page=1&pp=15

 

Let us know what gives

Link to comment
Share on other sites

Well the DFI mobo's especially your type ( the non SLI type ULTRA-D) are known to have major memory issues, I would first try to do a bios upgrade ( do it with one working module ! ), then setup your ram accordingly (cas,...) Don't OC yet. See that the correct voltages are set for Ram,... and then install from scratch. Also the newer powersuppy will help your rig to be more stable, waht are the amps at 12 volt ?, You need to have a good 20 amps at 12 volt to get it smooth sailing.

Here's the link for the bios update :

 

http://www.dfi.com.tw/Upload/BIOS/NF4LD310.exe

 

Read how to do it first if you haven't done this before ask a mate do it for you , you could permanently damage the board !!

 

And as Write suggested because the components are new doesn't mean thay are working 100%. Many of them have little bugs that are resolved within a few months after it's release.

 

Here's a link at Asusbozards forum , their main Oc'ers are using DFI boards to and they had some issues with them here's the link , read first post from Sierra

 

http://www.abxzone.com/forums/showthread.php?t=88699&page=1&pp=15

 

Let us know what gives

 

Friend has same board he has no problems and no technical issues with DFI and he has had his board longer than me. I have heard of no major memory issues on these DFI Boards even from DFI themselves have denied a "major" memory issue that can cause crashes 85% of the time. Also, i will BIOS Upgrade, 1 memory stick and up the voltage if need be then re-install like you said. Im willing to try anything at this point. The TAGAN 420w has 22a on the 12v rail and there is no way in hell i am buying a new PSU as i have gone through 4 yes four ... which is a tad annoying.

 

Im going to spend tomorrow finding out what it can be i hope i find the problem and get to the bottom of this. I will let you all know how i got on and what tests were performed! Any questions/comments after the report will help greatly. I WILL find out what is causing this even if it takes me a month! Had it on my last setup and now my latest, damn things! :roll:

Link to comment
Share on other sites

Well I have my background a month ago in the German hardware mag and DFI was gonna solve the issues with a bios update, if you read the forum at asusboards it's full of Bios stuff with major stability issues ( Dimmslots not working , beta bios screwing things up,... ) it ain't a hoax !!

 

Let us know what gives but like said before it just could be that you have a defective mobo and only an RMA can save you, you can try the ram in your friends rig to test it out. If you hook up eg an 6800GT PCI express your PSU wil have it hands full delivering the juice to all your components, sorry to say that ;) Tagan had some issues with some series of PSU, I am an Enermax fan I have the 465Watt ( 34 amps at 12 volt )on the Intel and a noisetaker 600Watt ( 2X18 amps ) for the AMD64. You can swap some components with your friend's PC while he plays with your girlfriend :)

Link to comment
Share on other sites

UPDATE:

 

Ok today i spent over 8 hours trying to solve, or at least identify what was causing these annoying errors. Ok the first thing i did was update the DFI BIOS to the latest (10/03/2005) which had MemTest86+ in the Genie BIOS option. So i decided to take out a stick of ram and put it in DIMM 2 (which is used for Single Channel. NOTE: DIMM 1+2 for 2 sticks of ram running at Single Channel, Individual ram sticks needs to go in DIMM 2 for Single Channel Operation also DIMM 2+4 for Dual Channel Operation) and decided to run MemTest86+ by Enabling it in the BIOS. I ran it for 1 hour, which means it does the 10 tests 4 times (15mins to complete the 10 tests) and i got NO errors not a SINGLE one.

 

So i thought to myself, maybe the other stick is a faulty one and can cause these errors. So i whacked that in DIMM 2 and did the same, i ran MemTest86+ for 1hour. Again not ONE error absolutely zip. I was pleased and a tad relieved that one stick was not faulty. However, i decided to use the rams sticks in DIMM 2+4 to enable Dual Channel Mode and run MemTest86+ again. After 1hour of running MemTest86+ with Dual Channel Mode it spat out 12 Errors in total mostly tests 3 and 4 where it failed. Some passed, about ONE, once it re-did the same test but about 98% of the errors were outright failures.

 

Ok at this point i was worried, i phoned up a good friend of mine and we kept it short and sweet. He still suspects my Memory Controller on the AMD die maybe be causing the errors - hence he thinks the CPU is not right clever at the moment. He thinks this as it did it on my NF3 setup and now my new NF4 setup. Getting back to Memtest86+ i decided to try again with the memory in DIMM 1+2 for 1GB Single Channel operation. I ran the test for 1 hour again and not ONE error was spat out. To be honest it will not run in Dual Channel, so i re-installed windows on Single Channel 1GB and here i am.

 

I have had a few errors here and there, 1 or 2 to be exact and im able to play games and benchmark the computer at last! I had 1 lockup and 1 BSOD which it gave no error inparticular otherwise i would list it here. So at the moment it looks to be about 80% stable which is better than less than 50% stable. Woo bloody hoo!

 

So whats the deal? Can anyone give any views on this because i have abolutely no idea, i know now my memory is not faulty - individually tested they get 4 Passes, 0 Errors which is about 1 hour testing. In Dual Channel they get 4 Passes, 12 Errors (errors mostly in Tests 3 and 4). Any ideas? And anyone have any suggestions on more tests to perform?

 

Also that PSU with the 30a+ on the 12V rail i want it ... i want more power and i want to eliminate the fact that my PSU is not in the error list. HOWEVER, my DFI requires a 24 Pin Power Connector not a 20 pin and i would need one with a 24 pin. If i was going to get a PSU could someone be so kind as to point a good one out for me? I will give a web link they are like 25min away from me - http://www.scan.co.uk/Products/Products.ASP?CatID=33&Category=Power+Supply&Thumbnails=yes also i have these guys to buy from too - http://www.dabs.com/uk/channels/components/caseandcaseaccessories/products.htm?catid=201 .

 

I hope someone can shed some more light on this issue, thank you!

Link to comment
Share on other sites

Does sound like the memory controller. Try on a different CPU.

 

I wish i could, believe me. I probably would have tried this if i had access to another socket 939 processor/motherboard but i don't. The only other person i know who has an Athlon64 is the "good friend" i talked about earlier. However, he lives a 2 hour drive away from me which is a headache in itself.

 

I do want to RMA the processor, regardless whether it is or is not the problem simply because i want to get rid of these problems and i will do about anything. Problem is, when i first got my CPU and the stock heatsink had the square piece of conpound on the bottom, i used it for about 2 weeks. I decided to take off the heatsink for a new one (which was incompatible) and i got the dreaded CPU coming out of the socket when clamped in. Now this bent a god few pins on the CPU and i had to bend them back. It took me over an hour to painstakingly bend them back to get it back in the socket but it did it and it worked.

 

Now i heard somewhere that AMD will not RMA CPU's that have bent pins which is a load of crap because this is hardly my fault but the cheap sockets they use on the motherboards. I mean i always take care when taking heatsinks off and this has been reported by my friend and a few other people on forums, even the intel's did this. So the pins are very,very slightly bent BUT the CPU still fits and works so how can i RMA the thing as AMD won't accept it? Hope someone can help out.

 

Any more ideas on the problems i still get occasionally?

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...