From what I have seen people talking (rando's on the Internets) it is just the GDDR6 that was used on the RTX 2080ti's from nVidia
Looks like there might have been a bad batch and they degrade over time- especially if you are doing heavy memory OCing
 
https://www.google.com/se...0+ti+Bad+Micron+Memory

rjohnson11
Micron and Samsung memory were tested by NVIDIA as far as I am aware. Both meet the standards NVIDIA needs. However in all fairness from all of the forum posts I have read Samsung memory seems to have a higher overclock capability than micron.

Being able to Overclock Higher on one does not make the other BAD or prone to Failures.
If members had been posting that you can overclock Samsung Memory Higher than Micron Memory that would not be an issue as just saying the Micron Memory is just bad.

bcavnaugh

rjohnson11
Micron and Samsung memory were tested by NVIDIA as far as I am aware. Both meet the standards NVIDIA needs. However in all fairness from all of the forum posts I have read Samsung memory seems to have a higher overclock capability than micron.

Being able to Overclock Higher on one does not make the other BAD or prone to Failures.
If members had been posting that you can overclock Samsung Memory Higher than Micron Memory that would not be an issue as just saying the Micron Memory is just bad.

I didn't say Micro was prone to failure. As far as I am aware there is no known issues with Micron memory. However if someone had two RTX 2080ti video cards and wanted to give me one of them I would be more willing to choose the one with Samsung memory should one have Samsung memory and the other micron memory.

Mine has Micron Memory and is solid so far...

I do not overclock it tho.
 

rjohnson11

bcavnaugh

rjohnson11
Micron and Samsung memory were tested by NVIDIA as far as I am aware. Both meet the standards NVIDIA needs. However in all fairness from all of the forum posts I have read Samsung memory seems to have a higher overclock capability than micron.

Being able to Overclock Higher on one does not make the other BAD or prone to Failures.
If members had been posting that you can overclock Samsung Memory Higher than Micron Memory that would not be an issue as just saying the Micron Memory is just bad.

I didn't say Micro was prone to failure. As far as I am aware there is no known issues with Micron memory. However if someone had two RTX 2080ti video cards and wanted to give me one of them I would be more willing to choose the one with Samsung memory should one have Samsung memory and the other micron memory.

Yes I know, it is more about what others are posting.

rjohnson11
Micron and Samsung memory were tested by NVIDIA as far as I am aware. Both meet the standards NVIDIA needs. However in all fairness from all of the forum posts I have read Samsung memory seems to have a higher overclock capability than micron.

my RTX 2080 Ti FTW3 ULTRA with micron runs +1200 all day NP :  )  

There is at least one person on this forum who feels sure that Micron memory is bad. I asked him to provide proof. So far, he has not come here and provided proof.

https://forums.evga.com/2...ifacting-m2919395.aspx

ty_ger07

MikeBGeyer

ty_ger07
Oh, I am sorry. Samsung memory is 100% reliable and will last forever. Of course.

The reason you saw so much talk about cards with Micron memory failing is because 100% of the cards had Micron memory.

ty_ger07
This seems like a good place to provide proof for either claim:

https://forums.evga.com/T...-is-good-m2919525.aspx

So far, I have not seen proof that Micron memory is better or worse than Samsung memory in regards to this artifacting issue. All I know is that 100% of the initial cards had Micron memory and therefore it's easy to claim that the Micron memory was the culprit; even though there is no proof (that I know of) that the memory was the cause of the artifacting. I think someone would need to have an artifacting card, desolder the memory and solder on new memory, and prove that the artifacting was solved by doing so. Has anyone done that and published the results? For all I know, it could have been a GPU core issue, a firmware issue, a voltage controller issue, a PCB trace issue, or any other source of interference and corruption.

You reference a series of posts that provide no actual evidence that the micron memory is not the issue and furthermore no one has provided any information on cards with Samsung memory issues.

Like I said, feel free to post there with proof. After all, that is the purpose of that thread.

No proof that Micron memory is NOT the issue, is NOT proof that Micron memory IS the issue. Again, 100% of the cards had Micron memory, so you are falling into a statistical trap. Proof is not opinion. Feel free to provide proof. I don't think proof exists either way. Without proof, the statement should be left unsaid.

...

None of the threads you provided prove that the initial batch of artifacting RTX cards were caused by Micron, or not caused by Micron, or were caused by Samsung, or not caused by Samsung, or were caused by NVIDIA, or not caused by NVIDIA, or were caused by aliens, or not caused by aliens.

I have the 2060 black edition and it has Micron and works fine and i dont overclock. I saw that many people reported having problems with Micron and 2080 model butg it might not apply to the 2060.

I started OC'ing my Micron...works flawlessly ...
 
I still say It's anecdotal on the whole "micron is bad" garbage.
 
Anyone with a card that works usually doesn't come and post about it.
 
 

This all started during the Pascal era when reviewers discovered that ASUS was shipping their Nvidia GPU based cards with two memory types. The negative feedback regarding Micron surfaced when it was discovered that ASUS released the Strix to reviewers with Samsung memory which overclocked significantly higher when compared to its Micron counterpart, presumably to ensure outstanding performance benchmarks. The Samsung models achieved anywhere between 1000Mhz and 1300Mhz whereas Micron models could only muster 400Mhz above stock.
 
A VBIOS update resolved the issue by allowing the Micron memory to close the gap, boosting overclock capability to 900Mhz and even 1000Mhz.
 
Micron memory is not bad. They faltered and I believe since then many internet patrons who are still unable to define the term research have been spreading suppositions based on this ancient fable.