Linux error - Help please

[mrybicki]

I had a mining rig that had a power supply take a dump, and now none of my GPU's mine. If I boot into Windows, the cards appear but the drivers wont load. In Linux I get this message ... Linux Error . To date I have completed an atiflash -f -p with the original bios from within Linux (EthosDistro), I have not don't it straight from boot either. I'm holding off on doing the Pin 1 to Pin 8 trick, only want to use that as a last ditch effort. if anyone has any ideas that would be great.

 

[jboydgolfer]

is iit possible for you to install one of the GPU's @ a time, into a different PC, and test them separately?

 

[AltCapwn]

Power Supply taking a dump often result in capacitor being blown up (the top become round instead of flat)

You'll need to replace them first on your GPUs AND your mobo before you start anything else because, sometimes it seems to work properly but can cause unwilling consequences and strange glitch or even more damage.

 

[OneMoar]

it probably fried the current sensing shunts in the cards

bad caps would't prevent the driver from running

it should not have fried the vbios rom
its also possible it nuked the motherboard

 

[mrybicki]

@Jboydgolfer: New mobo in mail today, I can report back as soon as it's in my hands. I have gone through all cards individually in a system that was not affected, and they all show the same error, however I didn't pay attention the the addressing it might be different.

@[B]altcapwn[/B]: The mobo and GPU are all new no, axial electrolytic caps all surface mount size E or F, they do not appear damaged, as none are separated.

@OneMoar: The mobo is different from the one that was affected. Do we know for a fact that the ROM data the image is talking about is in the VBIOS?

 

[OneMoar]

the error is generic whenever it can't locate the correct address for the rom image
cards are likely fried in some manner
shorting the pins is unlikely to help but its worth a try

if you are using pcie extender cables try without one card at a time

 

[kruk]

@mrybicki: The errors you see are harmless and are also present on working cards, you can't really extract anything useful from them. Paste link to your dmesg output here or GPU-Z screenshots ...

 

[jboydgolfer]

I think this goes without saying, but just to cover myself, be sure not to use the same power supply

 

[mrybicki]

Lol, it's already gone...

 

[eidairaman1]

Try cards separately in another machine, if they fail, take the coolers off along with the back plates and pci brackets and sell those the pcbs are junk at that point. If they happen to work your motherboard maybe shot.

 

[mrybicki]

Well that took a quick turn for the worse, all cards are less than 3 months old, I'll probably send them in for repair.

 

[eidairaman1]

What make and model of psu were you using???

 

[OneMoar]

What make and model of psu were you using???

not relevant dead psu is dead dead hardware is dead

short of getting schamics and probing it with a meter there isn't much you can do
RMA /thread

 

[mrybicki]

I had a corsair and I swapped it to a Thermaltake, I just swapped the box not the cables, and well, the PSU didnt like the old cables, FWOOSH...

 

[eidairaman1]

I had a corsair and I swapped it to a Thermaltake, I just swapped the box not the cables, and well, the PSU didnt like the old cables, FWOOSH...

There you have it, psu makers lay pins out differently so 1 cable at the psu side is not the same as another cable. Hard lesson to learn here. When switching brands or even between series sometimes, you definitely switch the cables.

Read up on jonnyguru.com for psu reviews.

What wattage were your psus?

What GPUs were you using as of make and model?

not relevant dead psu is dead dead hardware is dead

short of getting schamics and probing it with a meter there isn't much you can do
RMA /thread

Actually it is, I was asking questions while performing a R.C.A.

He just told us why this happened.

 

[mrybicki]

As your root cause analysis yes it was Operator error but as an EE that is really poor form and function as the cables are keyed the same. There has to be a standard on that. Otherwise there is no pokey-oke. Which is just bad business.

 

[eidairaman1]

As your root cause analysis yes it was Operator error but as an EE that is really poor form and function as the cables are keyed the same. There has to be a standard on that. Otherwise there is no pokey-oke. Which is just bad business.

That's that's how businesses make money is by accessories. They are standardized at the end you plug into a mobo or gpu, hdd/ssd etc. The only one to blame here is yourself, take responsibility for this and learn from it. Complacentcy/laziness gets you burned/killed, you are fortunate you didn't have a fire or even flatligned yourself.

 

[BadFrog]

This is the reason why I’ve never gone and purchased sleeved power cables. I’ve only purchased directly from a large manufacturer like Corsair. Too much of a chicken to try the smaller manufacturers. Lol

 

[R0H1T]

So you're gonna RMA the PSU & GPU+whatever else is fried? Are sure you didn't overload the PSU with mutliple GPUs et al

 

[eidairaman1]

This is the reason why I’ve never gone and purchased sleeved power cables. I’ve only purchased directly from a large manufacturer like Corsair. Too much of a chicken to try the smaller manufacturers. Lol

Here's an example, Mass Air Flow Sensor from a Chevrolet, Ford, or RAM Truck, they all perform the same task but are different physically, and even electrically. You wouldnt try a maf from a different maker to fix a different make of truck.

 

[BadFrog]

So you're gonna RMA the PSU & GPU+whatever else is fried? Are sure you didn't overload the PSU with mutliple GPUs et al

why not try it? lol. You'd be surprised what people will say if you just ask the question!

 

[R0H1T]

why not try it? lol. You'd be surprised what people will say if you just ask the question!

Because if that's the case it'll probably happen again with the same PSU & GPU replacements.

 

[mrybicki]

I have a watt meter installed it was only reading 650 watts when the 850W was installed. So the 1200W should have been sufficient. I'm not sure where you got the idea that I wasn't claiming responsibility, I just said that as an engineer, I would never utilize the same key pattern as a competitor if the signaling wasn't the same. And your analogy for the MAF sensors is not a valid one, as the connectors are all different. You want to say a head temp sensor, maybe. Also as a fellow airman, I'm going to ask you to back off a bit, I didn't come here for wit or sarcasm, I came here for help.

Just going to RMA the GPU's power supply I just returned to the store. Fry's returns anything.

 

[BadFrog]

Because if that's the case it'll probably happen again with the same PSU & GPU replacements.

OP mentioned he didn’t switch cables when he switched power supplies prob causing the problem. If he gets new equipment and uses the correct cables. Things should be okay

 

[eidairaman1]

I have a watt meter installed it was only reading 650 watts when the 850W was installed. So the 1200W should have been sufficient. I'm not sure where you got the idea that I wasn't claiming responsibility, I just said that as an engineer, I would never utilize the same key pattern as a competitor if the signaling wasn't the same. And your analogy for the MAF sensors is not a valid one, as the connectors are all different. You want to say a head temp sensor, maybe. Also as a fellow airman, I'm going to ask you to back off a bit, I didn't come here for wit or sarcasm, I came here for help.

Just going to RMA the GPU's power supply I just returned to the store. Fry's returns anything.

Well you screwed up your rig by using the incorrect cables, same goes for automotive parts too, use the correct part for the make and model you are dealing with or expect it to not work or do additional damage to equipment, injure or kill yourself.

I'm pretty surprised an "Electrical Engineer" would make such a critical mistake.

You know what to do at this point so the thread is done.

/thread