Please beware, New World MMO beta can cause permanent damage to GPU

[ubuysa]

It's just the reality of any bug, hardware or software.

But it's very clearly a hardware bug. The argument that was made in the video (and apparently by some of the hardware vendors?) that if the offending game's code was more efficient then the problem wouldn't have happened does not make this a software issue.

As long as the game is executing legal instructions in the proper way, and is not taking shortcuts or using undocumented features for example, then any and all failures are the fault of the hardware.

We shouldn't even be saying 'the game breaks the graphics cards' for example, we should be saying 'the graphics cards are not robust'. The game is an irrelevance. Software is an irrelevance. The graphics cards have a design flaw.

 

[SpyderTracks]

But it's very clearly a hardware bug. The argument that was made in the video (and apparently by some of the hardware vendors?) that if the offending game's code was more efficient then the problem wouldn't have happened does not make this a software issue.

No one is saying it's a software issue, that's never been suggested, it's unanimously agreed it's a hardware bug.

The point is, until Amazon liaise with all the parties involved, they can't understand what procedure call or group of calls is producing the flaw, and therefor identify where the flaw is at the hardware level, and so understand which models are affected, or if it's a general architecture design flaw.

When he's saying that if Amazon rewrote their code, it's purely from a perspective of the fact that it's easy for them to help avoid the issue by applying a simple code update which is what they did in the BETA (but didn't communicate what they'd done).

As long as the game is executing legal instructions in the proper way, and is not taking shortcuts or using undocumented features for example, then any and all failures are the fault of the hardware.

Again, no one is arguing that. No one is pointing at Amazon and saying they've written anything wrong, they're simply saying they've unearthed a design flaw and without them communicating what section of their code is triggering it, there's no way to understand the cause. And also that they currently are the only ones who are able to bypass the issue, until they communicate how this flaw is being triggered.

We shouldn't even be saying 'the game breaks the graphics cards' for example, we should be saying 'the graphics cards are not robust'. The game is an irrelevance. Software is an irrelevance. The graphics cards have a design flaw.

Perhaps, and if Amazon had been more helpful in liaising with manufacturers and API coders, perhaps no one would have said that. But people are understandably angry at Amazons lack of cooperation in this instance.

The be all and end all is that no one understands how exactly this is happening without Amazon getting over themselves and working with the parties involved. Until then, the onus is entirely on Amazon, weather it's their issue or not, they are the only ones that can help mitigate it and for whatever reason, are refusing to do so.

 

[ubuysa]

The be all and end all is that no one understands how exactly this is happening without Amazon getting over themselves and working with the parties involved. Until then, the onus is entirely on Amazon, weather it's their issue or not, they are the only ones that can help mitigate it and for whatever reason, are refusing to do so.

So you're saying that the hardware vendors don't know how the software is breaking their cards? That's a pretty scary thing to be saying. Unless of course, there is a suggestion - which nobody wants to actually say out loud - that the software is not performing within the published limits of the hardware and is taking shortcuts or using undocumented features?

If the software IS performing within the published limits of the hardware then the hardware vendors should be able to reproduce this issue themselves. How large is the GPU instruction set for Heaven's sake?

 

[SpyderTracks]

So you're saying that the hardware vendors don't know how the software is breaking their cards? That's a pretty scary thing to be saying. Unless of course, there is a suggestion - which nobody wants to actually say out loud - that the software is not performing within the published limits of the hardware and is taking shortcuts or using undocumented features?

I'm surprised your finding such issue with this? This is how every day mitigations are found, this is what "White Hats" and Pen Testers are paid extraordinary sums to find out, basic flaws in hardware and software. There's nothing new about this, it's the way it's always worked.

Once a flaw is unearthed, it's pretty standard practice for the source who found it (unless obviously nefarious) to then communicate their findings in a white paper with the manufacturer / software vendors so that they can address it before going public.

Standard practice in any IT field.

If the software IS performing within the published limits of the hardware then the hardware vendors should be able to reproduce this issue themselves. How large is the GPU instruction set for Heaven's sake?

If, until Amazon communicate what instructions they're running, it's like looking for a needle in a haystack.

 

[ubuysa]

I'm surprised your finding such issue with this?

I'm having an issue because the idea that software can break hardware is absurd. If running some software breaks the hardware it's because the hardware is faulty.

If, until Amazon communicate what instructions they're running, it's like looking for a needle in a haystack.

That's ridiculous. The GPU instruction set cannot be that large that the hardware vendors can't reproduce the problem. The haystack just isn't that big.

 

[SpyderTracks]

I'm having an issue because the idea that software can break hardware is absurd. If running some software breaks the hardware it's because the hardware is faulty.

Yes, obviously it points to a fault with the hardware, that's in full admission, no one is saying otherwise.

But software is often known to cause hardware defects, that's why there are loads of exploits that can brick phones, DDOS attacks bringing down networking infrastructure, they're hardware bugs, but companies work in tandem to mitigate them initially at a software level, and then design them out of hardware on the next implementation.

That's ridiculous. The GPU instruction set cannot be that large that the hardware vendors can't reproduce the problem. The haystack just isn't that big.

I can't comment on that, I have no idea about GPU instruction sets. If it were that easy though, I'm sure one of the several hardware manufacturers or NVidia or AMD would have found it by now.

 

[jamiephillips909]

I just limited my frames to my monitor for new world and it dosent sound like my gpu is taking off anymore

 

[robin h 25]

This is a game i would like to get and play, but not bought it yet due to having an EVGA 3080Ti, with reading and hearing the horror stories of EVGA 3080Ti/3090's being bricked due to the game,

I'm hoping this issue get resolved soon, as it's a game I've been wanting to play.

 

[SpyderTracks]

I just limited my frames to my monitor for new world and it dosent sound like my gpu is taking off anymore

As in the linked videos, just limiting FPS doesn't prevent the power surges.

 

[robin h 25]