RX580 MSI armor OC 8gb no post

[chaos-theory]

reflow does nothing on this cards u need reball

I tested VREFC Aand MVREF for all channels and they all have that 1.09v, so the GPU it powering the vram chips, other than that I don't think I can test resistance to the data channels, you mean I need to reball the GPU, that's a hard process is there a way to confirm?

[ktmmotocross]

I tried reflowing the GPU core with flux, I don't know if the balls melted or not, but I heated it really good, but the same thing happens, I also tried pushing down on the core while looking at the image and pushing down on the vram chips, but nothing changes. Many people suggested that the horizontal lines are caused by faulty vram chips or controllers, but the other places are also distorted, can anyone with more technical expertise explain that

reflow does nothing on this cards u need reball

[chaos-theory]

balls under gpu

I tried reflowing the GPU core with flux, I don't know if the balls melted or not, but I heated it really good, but the same thing happens, I also tried pushing down on the core while looking at the image and pushing down on the vram chips, but nothing changes. Many people suggested that the horizontal lines are caused by faulty vram chips or controllers, but the other places are also distorted, can anyone with more technical expertise explain that

[ktmmotocross]

I have used that picture before but it doesn't seem very accurate, can one channel of ram interfere with other channels? Because I removed 2 chips and I still have the artifacts at the same place

balls under gpu

[chaos-theory]

MANY times. have so many spare cards because of this. it is cheap so reball is not cost effective. u can try reflow

u can try this picture to check ram channel

I have used that picture before but it doesn't seem very accurate, can one channel of ram interfere with other channels? Because I removed 2 chips and I still have the artifacts at the same place

[ktmmotocross]

Have you encountered that error before? Do you think reflowing can solve this?

MANY times. have so many spare cards because of this. it is cheap so reball is not cost effective. u can try reflow

u can try this picture to check ram channel

[chaos-theory]

mostly bad solder joints under gpu

Have you encountered that error before? Do you think reflowing can solve this?

[ktmmotocross]

Hello, I am back again, sorry for disappearing!
I was working on an Asus motherboard which might have a bad PCH.
So I had this Graphics card which seems to be showing something on the screen but it's full of artifacts, and I decided to make a new thread because I was in a wrong path!
I reballed the A0 channel, and then removed both A0 and D1 vram chips and the card is still showing the same artifacts with only a bit of change on the corrupt lanes.

I am starting to suspect the core.

I tried tserver memfa test but it gives me an error : [error][Framework] Exception during attempt to execute test: TcoreSiliconTarget: Failed to load ME microcode using toollib on Tonga/Ellesmere

I made some researches and it seems to be caused by either bad vram contacts or GPU contacts.

So I went and did ./agt -mccfg
And it's showing the vram manufacturer as UNKNOWN
Very weird since it should show SAMSUNG, I researched more and some people implied that the channel A0 vram chips is responsible for defining the information about the vram chips, so I had to reball the vram chips, and saw that D1 has discoloration so I decided to remove it to to see if it's causing the artifacting, that's how I got here, the card is going into bios normally but full of artifacting without both vram chips present, and I still get the same errors.
I didnt want to use dmgg because the results are not reliable in this case as it's showing all chips are faulty.

I really don't know how to proceed, here are some pictures to show the artifacts with and without the 2 vram chips installed.
Some of the artifacts are static and some are dynamic.

Caption

Please keep this thread as it's very important to me, I have done plenty of researches and I need experience here instead of luck, thanks everyone!

mostly bad solder joints under gpu

[chaos-theory]

Hello, I am back again, sorry for disappearing!
I was working on an Asus motherboard which might have a bad PCH.
So I had this Graphics card which seems to be showing something on the screen but it's full of artifacts, and I decided to make a new thread because I was in a wrong path!
I reballed the A0 channel, and then removed both A0 and D1 vram chips and the card is still showing the same artifacts with only a bit of change on the corrupt lanes.

I am starting to suspect the core.

I tried tserver memfa test but it gives me an error : [error][Framework] Exception during attempt to execute test: TcoreSiliconTarget: Failed to load ME microcode using toollib on Tonga/Ellesmere

I made some researches and it seems to be caused by either bad vram contacts or GPU contacts.

So I went and did ./agt -mccfg
And it's showing the vram manufacturer as UNKNOWN
Very weird since it should show SAMSUNG, I researched more and some people implied that the channel A0 vram chips is responsible for defining the information about the vram chips, so I had to reball the vram chips, and saw that D1 has discoloration so I decided to remove it to to see if it's causing the artifacting, that's how I got here, the card is going into bios normally but full of artifacting without both vram chips present, and I still get the same errors.
I didnt want to use dmgg because the results are not reliable in this case as it's showing all chips are faulty.

I really don't know how to proceed, here are some pictures to show the artifacts with and without the 2 vram chips installed.
Some of the artifacts are static and some are dynamic.

Caption

Please keep this thread as it's very important to me, I have done plenty of researches and I need experience here instead of luck, thanks everyone!

[ductranvinh]

chaos-theory

this dmgg test is crap. never get normal result from it.
u can use tserver or agt

this picture can u maybe serve better, but rx580 have more problems with balls under gpu than bad ram chips

it's ok! with me,i think rx 400 & 500 series usually i meet 80% ram chip fault when display signal out crash and 20% is fault at gpu. Thanks for your pic

[ktmmotocross]

chaos-theory

this dmgg test is crap. never get normal result from it.
u can use tserver or agt

this picture can u maybe serve better, but rx580 have more problems with balls under gpu than bad ram chips

[ktmmotocross]

So when you post picture with crash image,i think you should check with mats, and replace ram chip was fault. RX 580 so cheap. Try and get experiance

i dont know how u get courage to advice something, when u advice test AMD card with mats. U better learn something before advice any of this crap

[ductranvinh]

In my opinion,only replace ram chip for higher model begin rx 5000 series

[ductranvinh]

So when you post picture with crash image,i think you should check with mats, and replace ram chip was fault. RX 580 so cheap. Try and get experiance

[chaos-theory]

Hello, after I managed to boot my PC with the MSI armor rx580 OC 8gb with CSM disabled, i went durect and booted Linux mint in CLI mode, did both direct mem test and dmgg.
But dmgg is showing that all chips are faulty

But when I specify only 4 chips it says no problem all chips are good, now the lspci -v is showing this card as a 4gb card is that weird, was the bios tempered with or anything?
these are the results, thank you!

Card crashes when loading drivers and showing weird dynamic artifacting on uefi bios

[chaos-theory]

Thanks mate, the bios is not the culprit, I managed to boot the graphics card by throwing it into another motherboard and disabling CSM I will make a new thread to explain it in detail and to ask a new question, this thread can be closed! Thank you all for the feedback, it's greatly appreciated!

In case anyone is curious it's showing artifacts, dynamic artifacts here are some pictures

[ductranvinh]

you was tried reprogram bios chip yet ?
if you was, him was dead (GPU) or measure hertz by Oscilloscope make sure him real dead

[chaos-theory]

So the result you were mesure very beautiful. All block voltage not missing. Last work you need to do reflash the bios chip
hai-ya..... you should connect 8 pin connectors. this port very important

That's the problem, the PC is not starting, there is no post, no beeping, nothing, just the fans turning at full speed, both the pc and the graphical card, I tried to short the bios at 1 and 8 pin but nothing, still doesn't post, should I short 1 and 5 pin
?
So what should I do here, I am in a corner and I have no idea where to go next

my Bios is W25X40WLCN and this is the pinout

[ductranvinh]

So the result you were mesure very beautiful. All block voltage not missing. Last work you need to do reflash the bios chip
hai-ya..... you should connect 8 pin connectors. this port very important

[ductranvinh]

VDDCI i call it vBoot