[Mobo/CPU] Endless reboot

[N3rot0xin]

So a few days ago, I was messing around with flashing the BIOS on my HD7850 to unlock voltage control and higher OCs. In the process, I *thought* i bricked the card because my system wouldn't boot into windows without a BSOD of varying sorts (the latest mentioning "atikmpag.sys" as the root). So I ordered a new GPU (R9 270x) to replace the thought-was-dead 7850. Got the r9 today and installed and.... same issue! System will POST just fine and I can stay in the bios indefinitely. But as soon as i try to load an OS (either from the SSD or a Linus Live CD[which fails at the "loading initrd.sys....." step]) and reboots. Every several reboots, it will appear to work, as it wont reboot right away. But there is nothing on the screen except a single pixel at the top left of the screen. Obviously it wasn't the GPU (which suck because now I have two good GPUs...Crossfire maybe?). But I just spent a paycheck on hardware I didn't need and wont have that money again for a while...

 

Here is what I have tried so far:

 

-Reset BIOS / Clear CMOS

-Swapped the RAM sticks in and out to see if maybe one of them was faulty

-Reseated the CPU in the socket and the block for the h60

-Unplugged the HDDs and left only the SSD

 

All of these things have failed. I'm thinking maybe the bios on the motherboard is to blame? Because I'm not getting the vibe its the CPU failing..Maybe when flashing the bios of the GPU it somehow left behind trash in the bios of the motherboard that CMOS clearing wont rid.. Thoughts? I didn't have much time to work on it tonight, and before I get back to it tommorow I'd like to have a list of things to do or a general idea of if "Its dead, Jim"... 

[Benjio]

I'd run a memory test if your computer allows you without restarting, here's a link to one: http://www.memtest86.com/. I know you switched them out, but hey, at least if it works you can rule the RAM out for certain.

[N3rot0xin]

I'd run a memory test if your computer allows you without restarting, here's a link to one: http://www.memtest86.com/. I know you switched them out, but hey, at least if it works you can rule the RAM out for certain.

I didn't think of that. I mean, if it turns up errors all its really doing is telling me "Hey, you need to spend $80 on new memory kit". Because I dont think they're still under warranty. I'm *hoping* this can be solved without more $$$. 

[Benjio]

I didn't think of that. I mean, if it turns up errors all its really doing is telling me "Hey, you need to spend $80 on new memory kit". Because I dont think they're still under warranty. I'm *hoping* this can be solved without more $$$. 

Well there's no use in avoiding the problem to save money; once you figure out what's wrong there's no point in not fixing it. I can see why you'd be frustrated though.  :unsure:

[N3rot0xin]

Well there's no use in avoiding the problem to save money; once you figure out what's wrong there's no point in not fixing it. I can see why you'd be frustrated though.  :unsure:

I'm really frustrated that I spent $200 on a new GPU. I'm justifying that however by saying "Its the GPU upgrade that I've been planning on for 6 months", since the 270x can reach close to 7950 ranges if you're good and have a good card. In that sense I've accidently more or less created a situation in which I can Crossfire for the sake of saying I can since there was nothing wrong with the old GPU. 

[Benjio]

I'm really frustrated that I spent $200 on a new GPU. I'm justifying that however by saying "Its the GPU upgrade that I've been planning on for 6 months", since the 270x can reach close to 7950 ranges if you're good and have a good card. In that sense I've accidently more or less created a situation in which I can Crossfire for the sake of saying I can since there was nothing wrong with the old GPU. 

Is it too late to return the one you just bought?

[N3rot0xin]

Is it too late to return the one you just bought?

I dont think its too late, but I've already justified it enough to keep it. It is nice knowing I have a spare incase something goes wrong with either. I might try again at flashing a new bios to the 7850 since I now know that it wasn't actually corrupted. 

[Guest]

I don't think it needs an OC to reach 7950 levels; it's already there. Although it definitely wasn't that great of an upgrade given your previous 7850, but maybe you could run another monitor/fold/mine/put the 7850 in another rig?

[Benjio]

I dont think its too late, but I've already justified it enough to keep it. It is nice knowing I have a spare incase something goes wrong with either. I might try again at flashing a new bios to the 7850 since I now know that it wasn't actually corrupted. 

Btw, does the computer allow you to boot into safe mode? I googled atikmpag.sys and it seems to be related to AMD's drivers, perhaps the driver you had doesn't like the new firmware on your card or just got corrupted somehow. If this is the issue I still don't know why it wouldn't let you boot into the Linux CD, but it's still something to consider.

[N3rot0xin]

I don't think it needs an OC to reach 7950 levels; it's already there. Although it definitely wasn't that great of an upgrade given your previous 7850, but maybe you could run another monitor/fold/mine/put the 7850 in another rig?

 

If I had another rig capable of running it i would do that. But I dont atm. It wasn't meant to be an upgrade as much as a replacement since I thought it was the GPU that was dead. 

Btw, does the computer allow you to boot into safe mode? I googled atikmpag.sys and it seems to be related to AMD's drivers, perhaps the driver you had doesn't like the new firmware on your card or just got corrupted somehow. If this is the issue I still don't know why it wouldn't let you boot into the Linux CD, but it's still something to consider.

No, no form of booting works. Safemode shows the loading of all the drivers, but as soon as it gets to the point it would normally show the login screen, it reboots. This applies to all of the various safe modes, even the 640/lowres mode. Normal start up gets to the "Windows is loading" screen, and again as soon as it gets to the point where it should show the login screen, is where it reboots. I wish I knew more about what exactly failed during the linux boot, that leads me to believe its almost for sure hardware. 

[N3rot0xin]

Uhhh... about the linux boot. 

 

initrd - "Initialize RAM disk"

In short, when loading a linux image from a live disk, initrd is used to load a temporary file system block into RAM for loading of actual root kernal from live disk. 

 

In short, simply by looking at the file that failed, I've more accurately pinpointed the cause of failure to one or both of my RAM sticks. 

[Benjio]

Uhhh... about the linux boot. 

 

initrd - "Initialize RAM disk"

In short, when loading a linux image from a live disk, initrd is used to load a temporary file system block into RAM for loading of actual root kernal from live disk. 

 

In short, simply by looking at the file that failed, I've more accurately pinpointed the cause of failure to one or both of my RAM sticks. 

Well that's a good sign I guess, just try running memtest I guess and see what results you get from that.

[N3rot0xin]

Well that's a good sign I guess, just try running memtest I guess and see what results you get from that.

I wish I would've thought to google the initrd purpose sooner, infact as soon as it failed at that point the first time is when I should've in hindsight. All part of the learning process. I will run the memtest tomorrow when I can get to my PC, and report back here. Provided I can actually do that given the unstable state of my computer. Memtest aside, is there a chance its the BIOS on the motherboard?

[Benjio]

I wish I would've thought to google the initrd purpose sooner, infact as soon as it failed at that point the first time is when I should've in hindsight. All part of the learning process. I will run the memtest tomorrow when I can get to my PC, and report back here. Provided I can actually do that given the unstable state of my computer. Memtest aside, is there a chance its the BIOS on the motherboard?

It really doesn't seem like it's a BIOS problem, as far as I know the only time BIOSs really go bad is if you're flashing a new BIOS, have bad OC settings, or the CMOS battery dies. I would imagine it's more likely that you would have a bad BIOS on your old graphics card, but I'm not knowledgeable about flashing BIOSs on gpus so I can't really say much about that.

[N3rot0xin]

I'm running memtest now. I started off running v1.65 (the version on the LiveCD i had) but realized that was probably a bad idea since it was so old and couldn't detect the CPU or chipset properly. Latest version is running from a flash drive now. 

[Benjio]

I'm running memtest now. I started off running v1.65 (the version on the LiveCD i had) but realized that was probably a bad idea since it was so old and couldn't detect the CPU or chipset properly. Latest version is running from a flash drive now. 

Good to hear, let me know how it goes.

[N3rot0xin]

Well...after 5 hrs of memtest, no errors. This only adds more ton the mix

[N3rot0xin]

Good to hear, let me know how it goes.

Correction on that last post, i reran the test and it showed two errors immediately, but somehow I hit quit and and restarted and they didnt show up again. i'm running 6 passes overnight, see if any errors show up again. Maybe this is a ram issue, given that if theres errors with ram memtest will show them pretty darn quick

[N3rot0xin]

I think I'm just going to RMA the motherboard. Its under warranty still so won't hurt. Not sure how long my CPU is warrantied for, but I may also RMA that as well.. I know where the problem isn't and that only leaves a couple places it can be. Guess I'll just replace until there aren't anymore problems.

[Benjio]

Another question: do u think temps could be the problem? My CPU got up to 60c during memtest and I think 61c is the max temps for an fx4100.

I would guess the overclock would cause more problems than the temperature,  you may want to set it back to default speed/voltage to see if that helps anything if you haven't already. I don't think 61 degrees is anything to be too worried about, it does seem a bit high for just running memtest though. And you may want to see if you have a friend or someone you can borrow some RAM from to see if that makes a difference; it's probably not a good sign if you're getting any errors at all.

[N3rot0xin]

I would guess the overclock would cause more problems than the temperature,  you may want to set it back to default speed/voltage to see if that helps anything if you haven't already. I don't think 61 degrees is anything to be too worried about, it does seem a bit high for just running memtest though. And you may want to see if you have a friend or someone you can borrow some RAM from to see if that makes a difference; it's probably not a good sign if you're getting any errors at all.

I just changed the post as you posted lol. I temps were an issue I would think it would be more unstable then it is. But it fails at the exact same point across seemingly all operating systems (windows and Linux). I ran memtest for another 8 hrs last night, not a single error. Maybe my ram is under warranty still, if it is I'll just replace that first.

[Benjio]

I just changed the post as you posted lol. I temps were an issue I would think it would be more unstable then it is. But it fails at the exact same point across seemingly all operating systems (windows and Linux). I ran memtest for another 8 hrs last night, not a single error. Maybe my ram is under warranty still, if it is I'll just replace that first.

Alright, doesn't hurt I guess. Also you did set your cpu back to stock speeds/voltage right? That would be a good thing to rule out.

[N3rot0xin]

Yes I did. I cleared CMOS and flashed the stock bios back. That inturn put the CPU back to stock