FireGL GPU running at 1x

[psyg]

Hi guys, i've had my ATI/AMD FireGL V8650 for a bit over 1.5 year now and about 7 months ago or so, maybe longer, it started screwing up.

The FireGL ran fine at 16x until one day i turned the computer on and it kept restarting by itself, so i took out the GPU and CPU and reinserted them into their slots.

The restarting went away but it was replaced with another problem, the GPU runs only at 1x.

The computer doesn't boot up on first go either, when i turn the computer on it the PCIe LED blinks for about 15 seconds, it makes 3 or 4 beeps (nothing is on displayed mind you) then i have to press restart, it beeps once then the monitors turn on but the PCIe led remains on. This is the sequence i have to go through everyday to boot the computer up.

The very weird thing is that for the past month, this didn't happen at all! I turned the computer on, beep once and i was running at 16x. Now again it's acting up.

Here is a picture of the problem.

http://gpuz.techpowerup.com/10/03/31/9ky.png

I've changed my CPU, my PSU and even the took it to a computer shop who tried a different brand mobo to be presented with the same problem.

I've contacted AMD and they refuse to give me a RMA because i lost my receipt for it even though i told them about the problem before the one year warranty was up.

I'm really going insane trying to figure out why it worked for that month and sometimes by luck boots into 16x but 99% of the time only boots 1x.

I'm getting a major performance hit which is the reason why i'm complaining...

Can anyone help me out?

EDIT: Forgot to add my computer specs if it matters.

HX-1000 Corsair PSU ( I had a 620 before but was told by two other computer shops that it might be insufficient power)
Gigabyte EP45 DS4P
Intel Q9550
8GB RAM

Monitors... if they matter
2 x Dell U2711

[wingman99]

After what you have done these are the options i don't think they will help becase of your PCI E booting problems. Number 3 seems like the only solution

1 Did you run driver sweeper on your old drivers and install new drivers.

2 Did you format your hard drive and install new chipset drivers.

3 Workstation Graphics Accelerator replacement.

Can you try any video card, just to see without keeping it.

[psyg]

The problem happens before the computer boots into Windows, so it's not a driver issue... but i did format windows and even installed the latest FireGL drivers with the same problem.

To be honest, I've tried nearly everything except replacing the GPU. The price of the GPU is what is preventing me from coming to the realization that something is wrong with the GPU.

I'm just curious to know why the GPU (1% of the time, and even one last month) booted to 16x.

It can't be a PCI slot issue since i changed to another mobo and the same problem was there.

If all else fails, i guess i'll put it in the oven and hope for the best.

After all this, AMD have lost a customer.

[petteyg359]

Any abnormalities (corrosion) on the PCIE connector edge?

Those computer shops that told you a 620 wasn't enough are lying to make a sale, incompetent, or both.

After all this, AMD have lost a customer.

If you failed to take advantage of the warranty and get an RMA SEVEN MONTHS AGO WHEN THE PROBLEM STARTED, then YOU are at fault, not AMD.

[wingman99]

Any abnormalities (corrosion) on the PCIE connector edge?

Those computer shops that told you a 620 wasn't enough are lying to make a sale, incompetent, or both.
+1 620w was enough

Can you try any video card, just to see without keeping it?:burn:

I have had problems with ATI, go with evga and follow there warrantee sign up procedure, in the 30 days of purchase, they have a good warranty, you do need to keep a copy of the invoice.

LINK:https://www.evga.com/login.asp

[MARCI]

Any abnormalities (corrosion) on the PCIE connector edge?

Those computer shops that told you a 620 wasn't enough are lying to make a sale, incompetent, or both.



If you failed to take advantage of the warranty and get an RMA SEVEN MONTHS AGO WHEN THE PROBLEM STARTED, then YOU are at fault, not AMD.

Pettey, my take on this is that he requested the RMA, but was denied because he no longer had the receipt. I may be wrong... but that's what I got from the reading.

[petteyg359]

Pettey, my take on this is that he requested the RMA, but was denied because he no longer had the receipt. I may be wrong... but that's what I got from the reading.

It is also not AMD's fault if a person can't keep track of a receipt required for warranty service of an expensive item. If AMD were not willing to replace it during warranty without a recipt, that's perfectly understandable; most companies would do the same. If they were offering to replace it even after the warranty had expired, and their only requirement was proof of purchase, then they should be commended. In either case, the OP's lack of a receipt is still not their fault.

[MARCI]

It is also not AMD's fault if a person can't keep track of a receipt required for warranty service of an expensive item. If AMD were not willing to replace it during warranty without a recipt, that's perfectly understandable; most companies would do the same. If they were offering to replace it even after the warranty had expired, and their only requirement was proof of purchase, then they should be commended. In either case, the OP's lack of a receipt is still not their fault.

I'm not attempting to assign fault, nor attempting to make excuses for anyone. I was attempting to clarify what may be a mistaken assumption. There's no need for combativeness or arguments here.

[w2richwood]

@ op you could try clearing cmos so your bios can pick up your hardware again hope this helps
Rich

[wingman99]

@ op you could try clearing cmos so your bios can pick up your hardware again hope this helps
RichHess tried a different board.:(

[psyg]

It is also not AMD's fault if a person can't keep track of a receipt required for warranty service of an expensive item. If AMD were not willing to replace it during warranty without a recipt, that's perfectly understandable; most companies would do the same. If they were offering to replace it even after the warranty had expired, and their only requirement was proof of purchase, then they should be commended. In either case, the OP's lack of a receipt is still not their fault.

You know, a decent company takes care of their customers and not just the ability to follow protocol.

I remember when i had a TN display from ASUS when i analysed the whites to be washed out, without a receipt they changed the panel free of charge.

When people buy workstation grade GPUs, i'd expect better service... the main reason people buy workstation grade GPUS other than the performance boost is customer service. They didn't give it and left me in the cold.

Sure i misplaced the receipt, but everyone makes small mistakes like that... still doesn't mean they shouldn't take care of a customer who paid big dollar for a GPU that could've easily been softmodded instead.

There isn't any corrosion at all, as i mentioned, it was working then all of a sudden one day it just began showing these abnormal signs.

Also the thread wasn't made to complain about AMD, it was made to see what my options are in the case of the PCI Express running at 1x. I don't want to throw it out if it's a simple fix.

It doesn't make sense whatsoever that it ran for a month (last month) at 16x then runs at 1x now... what's causing that?

Before last month, it ran at 1x but sometimes would kick in at 16x... this is where my confusion is at, the cause.

[petteyg359]

Sure i misplaced the receipt, but everyone makes small mistakes like that... still doesn't mean they shouldn't take care of a customer who paid big dollar for a GPU that could've easily been softmodded instead.

How does anybody know you paid for it, if there's no proof of purchase? Not saying you didn't, but they want proof of purchase for a reason. Depending where you live, it might be a good idea to buy online in the future, just so there's an electronic copy of the invoice you can retrieve later on :)

There isn't any corrosion at all, as i mentioned, it was working then all of a sudden one day it just began showing these abnormal signs.

It doesn't make sense whatsoever that it ran for a month (last month) at 16x then runs at 1x now... what's causing that?

Before last month, it ran at 1x but sometimes would kick in at 16x... this is where my confusion is at, the cause.

Trace(s) that might've been broken might re-connect if the PCB flexes just right in a certain ambient temperature. Any chance of memorable lightning storms before it stopped working? Could be a half-fried PCIE bridge. There are all many reasons why a PCIE connection might be at reduced speed. It could be a failing clock generator, and it can't sync at x16 speed.

[psyg]

Trace(s) that might've been broken might re-connect if the PCB flexes just right in a certain ambient temperature. Any chance of memorable lightning storms before it stopped working? Could be a half-fried PCIE bridge. There are all many reasons why a PCIE connection might be at reduced speed. It could be a failing clock generator, and it can't sync at x16 speed.

Nah no lighting strikes at all. The failing clock generator, is that due to the GPU or mobo because the latter was replaced with the same problem.

I tried the other PCIE slot in the mobo and same problem.

Yeah, next time i have to be careful but i guess you learn from mistakes in life :chair:

[petteyg359]

Nah no lighting strikes at all. The failing clock generator, is that due to the GPU or mobo because the latter was replaced with the same problem.

There's a clock generator on both sides. If the GPU's is dying, and can't match frequency to the motherboard's, the connection will slow slow down until the frequencies are close enough to not cause data corruption (at least as far as I understand; I'm not an electrical engineer, it's probably a lot more complicated than that). I think there was a post on this forum several months ago where somebody had a similar issue and actually replaced the clock on their card. Can't find it, though :(