I recently did a upgrade got my self a m3a78-cm mobo, amd anthlon 64x2 7750, 2x2gb ram ocz , and ocz modxtream pro 600 watt. I have built this comp with no problems what so ever besides this problem which i think is a irq conflict. After i had built this comp i was planning on reinstalling my old 8500 GT512mb since i ran short on money for a new vidcard. After i knew all my parts were working fine i went ahead and installed it no problem what so ever with drivers and software. I decided to see how i could run cod4 now that i upgraded from a single core and 512 mb of ram, Everything was running fine for about 30-45 mins. Till i reach a point in game where my system locked up could do ctrl + alt + del or alt + F4, so i turned off my comp my comp with my power button.At first i thought maybe i was overheating but then i relized if my comp did overheat it would most likly shutoff. I uninstalled my drivers reinstalled again with the latest drivers tryed cod4 again and same result. I took out the mobo manual and started reading on the part about configuring expansion cards and read this line "When using PCI cards on shared slots, ensure that the drivers support " Share IRQ" or that the cards do not need IRQ assignments;otherwise, conflicts will arise between the two PCI groups, making the system unstable and the card inoperable" Since i didn't know what they ment by IRQ i did some googling around i now know what it is and how to view them but i cant change them within windows cuz the use automatic settings after clicking on the irq is greyed out and i dont know alot about bio settings so i cant tell if i could change it in there
You can't change hardware interrupt requests from within Widows - this is something that needs to be done in BIOS - before Windows even gets ahold of it. But, it's more than common to have multiple components all sharing IRQs, it's all but mandatory. Having your 8500GT sharing the same IRQ as the PCI standard PCI to PCI bridge really shouldn't be what you're looking for - in fact, I think you're grasping at straws here. Go into your BIOS and check that 'Assign IRQ to PCI VGA' is [Enabled]...you shouldn't have to do anything else for your graphics card. I would think that the game has an issue more than any sort of IRQ conflict. Components share IRQs, there's nothing you can do about it. I would be looking at testing your RAM via Memtest86+ with one stick at a time. If they pass, then I would be testing any overclock (CPU or graphics card). If that's not it, then I would run some stability tests on the system...24 hours of Orthos Stressprime will prove if the system is stable, and if it is, then you've got software issues and it's most likely nothing to do with hardware. Psychlone
^ Where did I miss that?? Dude, you are overheating like Morbias said...figure out your cooling situation and this problem will probably stop. Also, when posting about your temps, use Celcius. The USA is the only country in the world that uses our standard measurements (I'm from the US as well), the rest of the world uses a different standard, so when talking about temps, everyone really expects to see them in *C rather than in *F ...I know it'll take a little getting used to, but that's just the way it is. Good luck! Psychlone
Ok i relized my oc was fine its jus had my heatsink/fan on backwards now its idling at 35*C which is around 95*F. I Did some playing around today with it and this is what i found out With Nvidia 8500 GT 512 mb installed with latest drivers im able to play 30-45 mins before a lockup occurs With my onboard ati Graphics card im able to play as much as 4 hours with no lockups so far but only get half the frames Now that this is being said im assuming that its something with my vidcard P.S I dont see anything within my bio's that says irq and i even updated my bio's with asus update
That setting is usually under something like PCI-E or HT...but your BIOS may or may not have it at all. You could try pressing F4 while inside the BIOS to see if any new options open up on any of the tabs. The first thing I need to say is that your idle temps mean absolutely squat. I could idle all day long at 27*C, and then start up a game/application that's CPU intensive - and the full-load temps may soar above 60*C, at which point I'd have stability problems. In the end, it's the full-load temps that really matter - no one really cares how hot your CPU is when it's just sitting there, it's all about how hot it gets when it's getting abused! - Especially if you're overclocked. Now, just because you can game indefinitely using the onboard still doesn't necessarily mean that it's videocard related...and I can't see it being IRQ related at all for the simple fact that videocards are effectively a 'plug n play' device, only requiring the correct drivers, which even Windows can find and install for you via Windows Update - Windows and/or the BIOS itself will assign an IRQ to every device that's inserted on the motherboard... If you've got another machine that you can test your videocard in, that would help with troubleshooting. Obviously, if it fails in another machine, then something on the card is faulty, and if it doesn't then you've got a problem elsewhere in your main rig. If you don't mind, you should download OCCT and run a full stability test with it. After it's done, it will give you some reports in graph form that you can go through - and the two reasons I ask this are: 1) It will show the temps from beginning to the end of the tests which can help in determining if you're overheating at some point 2) It will show any fluctuations in the PSU rails during the tests...this is going to be extremely important to see if you've got a major 'ripple' effect going on with the PSU - if it deviates more than ~2%, then you've got a problem with the PSU itself. When you're done, post up some screenshots of the temps and the power graphs so that we can help you better. Psychlone
well The highest ive seen my temps go to is around 60 C using pc probe by asus which came on the cd. At this moment im willing to do anything cause i jus this problem fixed.. so ill run that stability test and post some sshots when im done Results Vcore Test was the 1 hour test and says i am stable
First of all, all your PSU graphs look fine, so I would rule out (for now) any PSU issues...all the rails look to be where they're supposed to be for your model and holding pretty tight. Second, OCCT is showing your CPU temps to be tapping 50*C at their highest. Now, OCCT isn't a temperature monitoring software by itself, and since you can't be excellent at everything you do, I wouldn't trust OCCT's temps to be as accurate as ASUS Probe or what you see in the BIOS - albeit that the BIOS temps are neither idle nor full-load, it's actually somewhere in between. If you want to see what your full-load temps can be, download Linpack (Intel Burn Test) - but keep a keen eye on the *CORE* temps using CoreTemp or ASUS Probe, because it will raise your CPU like nothing else!! If you don't dare run that one, you can always get Orthos Stressprime and do a run of that for a few hours using "blend" at "priority10"...that will also give you a real good idea of what your actual full-load temps are going to be. The reason I wanted you to use OCCT is purely for the graphs...I don't know of another stress program that outputs graphs for the voltages that's as easy as OCCT, but OCCT doesn't stress the CPU nearly as hard as some of the other stress programs out there. Third (and possibly the most important!), OCCT is showing your VCORE (CPU Voltage) to be taking a pretty good hit of (-)4%...this is probably due to a motherboard discrepancy. Usually, a motherboard will ramp UP the voltage to keep a system stable, but yours is dropping the VCORE by what looks to be an abnormal amount. I would check and ensure that you've got your VCORE (CPU Voltage) set in BIOS at what it's supposed to be (I believe it's 1.3V??) - and no matter what, don't take that CPU higher than 1.45V with the 65nm core! In fact, you shouldn't have to have 1.375V running through it at stock (unless I'm mistaken on the default voltage, which I can't seem to find right now...too tired and burnt from work) Odd that you're loosing that much voltage on the VCORE under full load - this is more than likely a sign of sub-par voltage regulators not doing what they're supposed to be. Try setting the VCORE in BIOS like I suggested (whatever your default voltage is supposed to be) and run OCCT again - post another screenshot of the VCORE graph and let's do a compare - we can get to the bottom of this fairly quickly I think. Also, you may want to run IBT (Intel Burn Test) for a quick run just to see how hot your CPU is actually going to get - and be prepared to crap your pants...IBT usually heats up any CPU roughly 10* further than most people have seen theirs go... Psychlone
Bro - something's wrong with your motherboard. Is it too late to do an RMA (Return Merchandise Authorization) and get another?? If it's still under warranty and you can stand to be out of a computer for a few days while they ship out a new board, you'll be in much better shape! Your VCORE is going the wrong direction - I just did a quick-and-dirty 15 minute OCCT on my machine (at 3.1GHz, 1.3V) just to see real fast, and not once did the VCORE drop below where it's set. (even though it's set at 1.3V and it shows almost 1.33V - and the temps were way off in OCCT for me - it was showing 53*C when ASUS Probe and CoreTemp were showing 48*C) - this is the first time I've run OCCT on Windows7, so it could be polling some strange registers or something... Compare this 15 minute run with your 2 separate VCORE runs and you can already see that something's not quite right. Anyway, I think there's something wrong with your motherboard - and if you can, you should send it back and get a replacement, or even better, get a higher-shelf board like the M3A79-T Deluxe for a few more dollars. Psychlone
Thx for ur help man but im past the return date have no warranty on it i took out my 8500 and im able to play games now alot i wonder if there's a difference with the vcore now ill hopefully be able to get a vid card after i make some money within the next few months cuz i got laid off and im jus looking for work now atm not a job
So you take out the 8500 and it's solid now? That means there's probably a power leak in the NB or something related to it. I don't make motherboards, I only know enough about the circuitry to be dangerous, but even with my limited knowledge of how everything is connected, removing the 8500 from the PCI-E lanes and getting stability doesn't necessarily mean that you've got a bunk card - you should test it in another system and see, but I'm betting that it's the Core/PCI-E link that's borked...that would actually make a little sense. Try testing the card in another computer if possible and see if it has the same performance in OCCT's VCORE graphing - first run a base test for 15 minutes, then drop in the card and test it again to see...I'd be surprised if it's the card itself, and I sure hope it's not since an upgrade there is going to be more than a new motherboard. Well, either way, you're looking at at least a couple hundred bucks, which sucks badly... Sorry Brother. Psychlone
