New World, now officially launched, still damaging cards.

[yotano211]

I am not surprised that top end EVGA cards like the 3080 or 3080ti are failing. When I was selling lots of graphics cards some years ago, I had many issues with the EVGA 2080ti black edition. Out of 12 black editions that I sold, I had to personally RMA 3 of the cards and I got messages a few weeks to 2 months later that another 4 where RMA'ed. 7 out of 12 where RMA'ed within a few months.

That made me question myself that if I ever built a desktop in the future I will avoid EVGA.

I watched this video when it hit, its indeed weird... its only very specific high end cards too, so I sort of agree with Amazon its bad design on some cards... like there are other 3090's that have no issue with the game at all, but very specific models with different power models do have issues with it. I really don't know who is in the right or wrong here, but the be honest since this is literally the only game doing it to ANY gpu... I'd say... umm yeah prob the games fault at end of day. the common user should not be forced to limit fps and lower power draw. that is not the job of the end user!!!

The video that JayzTwoCents made a few days ago, it seems its also the fault of the gpu manufactures like EVGA. I am not surprised with the issues with EVGA. I would personally avoid anything from EVGA. With my own personal experience with EVGA, I will fully avoid that company in the future with graphics cards.

 

[freeagent]

Never had a problem with them. Been using them since early 2000s..

 

[yotano211]

Never had a problem with them. Been using them since early 2000s..

with EVGA?

 

[InVasMani]

EVGA just pushes the envelope more than most other brands beyond reference in general. That in tandem with higher end chip sku's in the first place is obviously going to have higher failure rates in general. I still see this issue as a quality control problem not a software GPU killer matter otherwise it would be more consistent across more GPU's not just a sprinkling of them. There will always be a sprinkling of GPU's that fail under demanding GPU workloads is the reality though space invaders problem ring a bell and guess what New World didn't break those early RTX cards.

 

[yotano211]

EVGA just pushes the envelope more than most other brands beyond reference in general. That in tandem with higher end chip sku's in the first place is obviously going to have higher failure rates in general. I still see this issue as a quality control problem not a software GPU killer matter otherwise it would be more consistent across more GPU's not just a sprinkling of them. There will always be a sprinkling of GPU's that fail under demanding GPU workloads is the reality though space invaders problem ring a bell and guess what New World didn't break those early RTX cards.

From JayzTwoCents' video, that game might get blame on some issues also. The game code is pretty bad, watch the video.

 

[FireFox]

I am not surprised with the issues with EVGA. I would personally avoid anything from EVGA. With my own personal experience with EVGA, I will fully avoid that company in the future with graphics cards.

Only brand i have always used, never had a problem.

 

[DeathtoGnomes]

EVGA just pushes the envelope more than most other brands beyond reference in general. That in tandem with higher end chip sku's in the first place is obviously going to have higher failure rates in general. I still see this issue as a quality control problem not a software GPU killer matter otherwise it would be more consistent across more GPU's not just a sprinkling of them. There will always be a sprinkling of GPU's that fail under demanding GPU workloads is the reality though space invaders problem ring a bell and guess what New World didn't break those early RTX cards.

If you had watch the videos, Jay mentions that there were some non 30xx cards that have been reported but nothing substantial to elaborate on. As for the "not a software GPU killer", the crap software exposed the QA issue with the cards that did fail, the one with the bad solder on the VRMs. Go back and watch the videos.

 

[BSim500]

This stuff is why I lost all interest in top-end gaming. New World may have some shoddy code but the current trend seems to be "we've run out of all the easy 700-900-1000, etc, series efficiency gains so the only way we can double the speed is to double the wattage", and they've started pumping out space heaters that are pushed to the absolute limits by default and literally have no overhead whatsoever. You want a 350w GPU, then design it for 400w and put a 100% power cap at 350w. Then an accidental 110% load that breaks that power cap won't explode the card. Instead it looks like the only way they've been able to push the envelope for top-end cards is to throw away any real overhead or durability that lower end cards have and design a 350w card for exactly 350.0000w...

And New World isn't the only game causing 3090's death / instability issues (example 1, example 2, example 3), it's just the one getting all the media attention at the moment.

 

[OneMoar]

this sounds like a edge case that the driver/hardware was never designed for the entire point of all the power regulation and protection circuitry on modern NVidia cards its to prevent exactly this kind of failure
the game would need to be doing something really outside the norm to sidestep all of that and cause a failure

this is a real head scratcher because some how though shotty code and force of will amazon has managed to create a game that puts more strain on a card then even the hardest synthetic tests or mining loads

my best guess agrees with @buildzoid the only way this could be happening is some kind of wildly unbalanced or `spikey` load ands repeatedly overloading the voltage regulators faster then the IC can compensate for, a sustained load should be throttled into the ground long before it ever gets any ware near damaging something even with a overclock and a 120% power limit

 

[pavle]

2 things are wrong in this case: first what does amazon know about making videogames? And second iffy construction of all these high end 30xx cards, seems like everything's been rushed from the beginning with frequency issues, thermal pad placement, Amazon power-virus situation. Nvidia's must have really not thought this one through. Looks like they underestimated big chip's behaviour, large currents are involved and induction is bound to show its teeth on every corner (current spikes).

 

[qubit]

Any game that can make my graphics card reassuringly busted is the game for me!

 

[freeagent]

I’ve run the heck out of my card already, I’ve watched it’s temps all over, it’s a beast. Runs cool, really cool.. it might have the same problem, but for now I’m going to say I doubt it. Still not gonna run that game though

 

[Chomiq]

2 things are wrong in this case: first what does amazon know about making videogames? And second iffy construction of all these high end 30xx cards, seems like everything's been rushed from the beginning with frequency issues, thermal pad placement, Amazon power-virus situation. Nvidia's must have really not thought this one through. Looks like they underestimated big chip's behaviour, large currents are involved and induction is bound to show its teeth on every corner (current spikes).

Except this happens also on the 20-series.

 

[pavle]

We all know 20xx series cards were a joke.

 

[qubit]

We all know 20xx series cards were a joke.

My RTX 2080 SUPER works fine. It's the MSI Gaming X one with the big cooler.

 

[Space Lynx]

My RTX 2080 SUPER works fine. It's the MSI Gaming X one with the big cooler.

MSI has gone downhill in recent years. I liked their motherboard I had recently x570 tomahawk but that was the only exception.

 

[Bomby569]

Except this happens also on the 20-series.

it can happen to any card that has a problem, if it weren't for NW they would die from any other intense load, some people said 5700's too for example. But those are very small numbers. The problem is clearly the 30 series and some cards and brands in particular, EVGA.

 

[Chomiq]

MSI has gone downhill in recent years. I liked their motherboard I had recently x570 tomahawk but that was the only exception.

MSI has gone downhill... yet somehow their cards don't explode in New World.

 

[mouacyk]

MSI has gone downhill... yet somehow their cards don't explode in New World.

Perhaps their cards have proper phase load balancing.

 

[Simpleris]

400W for only approximately 110 FPS! LOL!

And the graphics are not that impressing, compared to existing games. Is this even worse optimized than cp77? I swear it might be even mining crypto at the same time.

 

[80-watt Hamster]

... first what does amazon know about making videogames?

Nothing. Which is why they bought/hired studios that do.

General response to the "Unoptimized!" comments: It runs pretty well and looks pretty good for my roommate on an RX 470 for supposedly being poorly optimized. Maybe my standards aren't as high as others'. *shrug*

 

[chrcoluk]

I have never got this crazy of uncapped FPS, I see very limited benefit from it, if any at all.

I cap my FPS every game I play, I dont like running components at max load, and I find games run smoother when they running at a cap vs a hardware bottleneck. Not to mention it saves lots of power doing so, so a greener way of gaming.

 

[DeathtoGnomes]

Nothing. Which is why they bought/hired studios that do.

General response to the "Unoptimized!" comments: It runs pretty well and looks pretty good for my roommate on an RX 470 for supposedly being poorly optimized. Maybe my standards aren't as high as others'. *shrug*

There is no optimizing poor underlying code.

I have never got this crazy of uncapped FPS, I see very limited benefit from it, if any at all.

I cap my FPS every game I play, I dont like running components at max load, and I find games run smoother when they running at a cap vs a hardware bottleneck. Not to mention it saves lots of power doing so, so a greener way of gaming.

Someone that gets it.

 

[Vayra86]

It makes me question my purchase of a 30 series card. I haven’t watched the full video yet though. I thought because mine was new it would have no problems like every other card I owned. Except my XFX GTX295.. one side died. Yay warranty lol.. But because of this and the shortage my gpu is at stock clocks all around. The slider only gives me 5% so it’s not very generous..

Ever since Turing had totally new problems that Pascal never had, I was convinced Nvidia had entered a path I don't want to partake in any time soon, and if so, at a very profitable deal only. We've seen persistent stutter issues, a relatively high number of card deaths in the first few years, the Space Invaders, and TDPs that went up across the whole stack. Ampere only continued that strategy and went even more bonkers on power consumption. Of course its not easy to copy over your Volta tensor cores to knock off consumer GPUs, but hey, writing's on the wall Huang.

High TDP cards = more prone to early demise. Exceptions really make the rule, especially if you move up to the top end SKUs (above x104). Its not like cards die left and right... but its clear the whole stack is pushed a bit further into the risky zone. We've seen this before, I recall the way Vega was clocked, for example, and how undervolting made everything better - its no coincidence this also works on recent Nvidia gens. Even Pascal has a tiny slice of that already, albeit very tiny, because you can pull a few extra boost bins out of it then.

 

[InVasMani]

Except this happens also on the 20-series.

That series also suffered catastrophic failure with space invaders so not too reassuring and that problem too was mostly tied to the higher end tier of cards.