Might seem like a silly question to ask but not long ago many people reported their 30 series cards were dying while playing Halo MCC specifically. It seemed to be happening too often to just be a coincidence. Can anyone who's played Halo MCC on an EVGA 30 series card comment?

It has been fine for me, the only game that has been killing 30 series cards is New World probably fixed by now. As the game put the cards in a state that is worse than benching/mining/rendering 3d/folding, removed all the safeties both in the hardware firmware and software.

Now if you have an old PSU, meaning one that was manufactured/built before the 30 series came out there is a good chance during the 30 series large boost clock power spikes it could cause the psu to fail and take your card out with it.
I have not had any issues with MCC or CP 2077. The only ones who I have seen losing cards their older psu's have gone pop and taking the cards with them most of the time.

donta1979
It has been fine for me, the only game that has been killing 30 series cards is New World probably fixed by now. As the game put the cards in a state that is worse than benching/mining/rendering 3d/folding, removed all the safeties both in the hardware firmware and software.

Now if you have an old PSU, meaning one that was manufactured/built before the 30 series came out there is a good chance during the 30 series large boost clock power spikes it could cause the psu to fail and take your card out with it.
I have not had any issues with MCC or CP 2077. The only ones who I have seen losing cards their older psu's have gone pop and taking the cards with them most of the time.

 
I built my computer right when the 30 series was launching. My PSU is a Superflower Leadex III 850W 80+ Gold. Apparently this PSU model first was released in Feb 2020. Would that count as new or old in your view here?

ty_ger07

Michapolys
All games are safe if your PSU can handle the card.

Finding a PSU that can power these cards reliably is another subject.

I disagree entirely. For every example of a card killed with a PSU you don't approve of for whatever reason, is another example of a card killed with the best PSUs available.

I would say it depends on the card. If it was a 3090 FTW3 Rev 0.1, I would not trust it in Halo MCC, even with the best PSU money can buy.

Halo MCC had been killing 3090 FTW3s for the better part of a year before New World came out.

I have a 3080 Ultra FTW3. Would that be safe in Halo MCC? I'm on my second 3080, so my first RMA. The first card started dying in about two weeks. The current one has lasted almost exactly 1 year so far. I didn't try Halo MCC on the first one, and I haven't tried it on this one yet.

On founders editions and kingpin editions it’s safe.

Sajin
On founders editions and kingpin editions it’s safe.

Does this apply to 3090s only, or normal 3080s aren't safe either?

nomoss

Hefbn87

Sajin
On founders editions and kingpin editions it’s safe.

Does this apply to 3090s only, or normal 3080s aren't safe either?

3080s were always fine.  Don't be afraid to play any games.  You have a 3 year warranty.  Use it if you need to.

Have you played MCC on a 3080?

nomoss
 
3080s were always fine. 

Just gotta do a google search... https://www.reddit.com/r/...us_evga_rtx_3080_ftw3/

It'll definitely help mitigate the issue. The biggest thing I've noticed in some older games is that they don't allow you to limit the fps when in menus and that is when you can see some pretty big power spikes. I have the Halo MCC downloaded but haven't tried it yet on my 3080Ti.

I would be set a frame rate limit but it should be fine.

ty_ger07

Michapolys
All games are safe if your PSU can handle the card.

Finding a PSU that can power these cards reliably is another subject.

I disagree entirely. For every example of a card killed with a PSU you don't approve of for whatever reason, is another example of a card killed with the best PSUs available.

I would say it depends on the card. If it was a 3090 FTW3 Rev 0.1, I would not trust it in Halo MCC, even with the best PSU money can buy.

Halo MCC had been killing 3090 FTW3s for the better part of a year before New World came out.

donta1979
The only ones who I have seen losing cards their older psu's have gone pop and taking the cards with them most of the time.

I think that statement should be backed with some evidence.

Please do not listen to this person. He does this contrarianism thing where he shoots down anything he doesn't personally believe in.
 

Intoxicus
There is a fan ramping issue that has been correlated with other games causing these kinds of problems. You'll see the rpm double into the millions of rpm while you can hear the fans actually trying to achieve the numbers stated.

Still this nonsense? A fan can't ramp past 100%. 100% is 12v and there is no boost converter on the card to go above 12v. The fans aren't trying to achieve the numbers stated. What you are seeing is a bugged report of the fans when the GPU goes into a protection mode.

Did you ever repaste your GPU? That is the solution others have used. If the GPU develops a hot spot, the fans will ramp up to true 100% which exceeds the default fake 100%, to cool down the hot spot. No I am not going to explain "fake 100%" to you a 4th time; all that matters is that it isn't "ramping to millions of RPM" and is not THE CAUSE of any issue (it is a symptom which you keep neglecting).

A hot spot issue is something you can solve. I don't know why you keep ignoring it.

A hot spot issue isn't what is killing cards. What is killing cards is the VRM being overloaded by a power spike. The VRM can have transient power spikes because NVIDIA decided to use average power instead of peak power as a limit. Then, NVIDIA effectively disabled the hardware VRM protection (set it so high that it is pretty much useless, so that it wouldn't keep triggering annoying blackscreens). So, that is what is killing cards, not a fan ramp issue. When the cards are at high boost and suddenly jump from an easy task to a harder task, the spike can instantaneously kill the VRM. It's able to draw big power spikes while the average still appears momentarily low. The power spike is the GPU drawing power. It isn't the PSU forcing it to draw power (not possible). It's a NVIDIA GPU design issue. It's a design problem which has existed for a long time and is getting worse and worse as cards get more powerful. I wonder if NVIDIA has a change/solution in a future design, or if they will continue to just try to overbuild it by enough margin that it hopefully doesn't blow up (and then let people like EVGA push the power limit on reference designs to the point that they blow up left and right).

Intoxicus
Please do not listen to this person. He does this contrarianism thing where he shoots down anything he doesn't personally believe in.

That isn't my intent. I mostly want to shoot down false and information before it spreads. The fan thing was just a theory that quickly got spread in click-bait articles before the theory had a chance to be tested and vetted. Once it was vetted, it was determined to be incorrect. I don't especially like spreading false information. We have had deep discussions about this many times and the stuff you have been repeating is pure nonsense and not based on facts about how PWM fans operate. You were saying that the fans are being over-volted and all kinds of nonsense which was physically impossible.
Don't hold it against me that you can't let go of it.
The fan ramp thing is something you can try to fix. Why don't you try? As you have experienced, it hasn't killed your card; because it isn't the thing that has been killing cards.

Intoxicus

Please do not listen to this person. He does this contrarianism thing where he shoots down anything he doesn't personally believe in.
 

Yeah, although he seems to know some stuff the conclusions he jumps to defy the laws of physics.

It is a solid proof that knowledge and intelligence are different things.

Lots of educated fools out there.

ty_ger07
What I said is based on evidence of many respected in the industry. I didn't come up with it.
What you stated in one thread, on the other hand, lacked evidence and logic. Do you care to share it with us again?

You learn a bit of electrical engineering, use an oscilloscope on the PSU FETs to see their switching frequency and realize the problem yourself.
It cannot get simpler than that, really.

ty_ger07
That's how the stuff I mentioned was discovered by those respected people. I'm just stating what was discovered through research; it's not something I came up with.
What was your proposal again? The PSU craps out, the voltage drops to nearly zero, the card somehow draws a power spike from the zero power the PSU is producing, somehow that causes it to kill itself, and the CPU and memory and everything else hooked up to the PSU continues to operate just fine throughout that whole event and you can even hear the video game sounds continue to play. Ok, makes sense.
It's a fine theory. Most of it makes sense. But why invent a new theory? There are all kinds of strange theories we could make up. Is it that important to try to make something other than the GPU the culprit? What is the motivation? Where is the evidence?
And, for fun, can you please provide an example of an ATX power supply operating at a 100hz switching frequency?

OK this is it. People like you make me thankful that the ignore list exists.

My FTW3 3090 died from Halo MCC Dec 2020 and I had it paired with a Corsair AX1000 PSU. People that reported failures had there cards paired with high end PSUs. I've been lurking around these forums since Dec 2020 and I've seen many theories ranging from "bad" PSU choices to the orientation of how people mount there GPUs (horizontally/vertically). It gets pretty ridiculous I tell you.
 
The GPU controls its own power, that is a fact. The problem is not with PSUs but the GPU itself !
 
If only a handful a PSUs work for 3090s then EVGA and Nvidia MUST state it on the requirements, but they cant because its not the problem.