My New World Testing revealed something unexpected... This NEEDS to be fixed!!!
HTML-код
- Опубликовано: 22 авг 2024
- While trying to show how to protect your high end GPU from blowing up while playing New World, I discovered an even BIGGER issue!
World Wide GPU Giveaway here! - • I'm giving away THREE ...
Get your JayzTwoCents Merch Here! - www.jayztwocen...
○○○○○○ Items featured in this video available at Amazon ○○○○○○
► Amazon US - bit.ly/1meybOF
► Amazon UK - amzn.to/Zx813L
► Amazon Canada - amzn.to/1tl6vc6
••• Follow me on your favorite Social Media! •••
Facebook: / jayztwocents
Twitter: / jayztwocents
Instagram: / jayztwocents
SUBSCRIBE! bit.ly/sub2Jayz...
"I'd rather have lower fps than zero fps because the card blew." Priceless Lol
I would rather love to see a FARCRY 6 testing! The game that the majority gamers would play & not this new world bluff!
i wouldent even play the game thats a risk to my gpu
@@IronMan-jj2fd Ran well for me.
lol
lol that is priceless any other news what happen to Cole?
It's from the miner built into the game, it farms Daddy Bezos coins for you while you chop trees.
Lol
people laugh but it's true
HEHEHEHE
what cant you understand that its nvidias fault about dsr and 4K on non4K monitor? i had the same thing already 2 years ago.
@@ImotekhtheStormlord-tx2it its also new worlds fault lmfao this shit ain happening with any other games
The expression: "This game will use 150% of your GPU power" now make sense ;)
Hey, does amazon play games and mine eth for you too in-game? Drinks like a fish laughing, speaking like a drunk again.
Can you do a video of a “general computer maintenance” specifically software. You taught me how to build a PC a few years ago, but I feel like there’s things I should do for upkeep. Updating bios etc, it’d be great to have a tutorial on everything to do!
1. Windows or no antivirus at all (firrwall is a must have anyway)
2. Don't chase newest drivers without a GOOD reason
3. Control your temperatures and prevent power limits using Throttlestop and MSI Afterburner
4. Use frametime lock equivalent to fps just below tour monitor's refresh rate (for 60 hz 16.960ms works best)
The best PC maintenance is "if it ain't broke don't fix it".
Definitely don't update your BIOS unless you actually have a reason to (fixing a bug, adding compatibility when you're upgrading your CPU, etc). If it's currently working for you there's no reason to change it.
"I hope that guy wasn't playing New World..."
I WAS playing it, Jay. I was...
Lmao, did the card blew off?
Was... RIP.
Oof
F sorry bro
F
SO if its doing it in the menus it makes sense that its killing cards seeing as people have been sitting in queues for 8+ hours
In his video you did see it spiking up even when in the game, so I don't think it matters too much if it's menu or game, it's the same stress it seems.
This only happens with the 3xxx series, not giving 4k option with 2070
I have a 2070super. It revs high in the menu and once I get in game it levels out. I'll have to pull up some monitoring software while I play today and see what's up.
@@Enders1315 I have a 2080 super. I’ve had zero issues until I bought my house in game and tried to decorate it. Apparently that maxed out my gpu for some reason.
If people are stuck in a queue then they need to pick a different server to play on. I have played the game for 12 or 14 hours or so (sorry I work 12 hours a day 7 days a week) and I have never seen a queue line yet.
This is all a test for the next VR generation, to see if they can fry our brains while we game.
On some other note, imagine getting your hands on a 3090 in this day and age and having a game fry it.
I'm sitting here with my Sapphire 5700xt Nitro+ because no one wanted them because of early driver bugs (it's pretty great now)
Deffinitly going to be avoiding this game cause I don't wanna risk it, seriously, they deffinitly didn't do enough (if any) testing or optimizing before launch.
Deffinitly makes a good case to wait a few months or even more than a year before picking up new games, the serious lack of effort to *finish* a game before launch is getting ridiculous
Jay,
DSR is driver-level trickery to enable something that any game can implement if it chooses, except when a game does it, it can do a better job. If the internal rendering resolution for the 3D scene is higher than the native display resolution of the panel, you can actually get some really nice anti-aliasing. If this is being done at the driver level, it gets applied to the entire screen. However, if it's done by the game, then UI/HUD elements can be rendered separately, which removes the weirdness that can be caused by over-rendering the UI/HUD and then using a downscaling filter to get it to fit on the screen.
It looks to me like the game might just be doing what DSR does, but internally, instead of at a driver level.
Is that just a fancy new name for supersampling? If so not separating it out to its own option is a bit misleading.
@@egoalter1276 Sure, it's misleading, but AMD and Nvidia both do it.
@@Niarbeht Oh, im not knocking NVIDIA. Options menus are designrd by the program developer.
Jay: "Careful playing this game"
People: "DON'T TELL ME WHAT TO DO" *destroys 3090*
HAHAHA There goes me lmfao at poor misfortune people destroys the 3090s
so now 3090 is will go up due to started get even more rare
@@WilliamHollinger2019 meh 4090 is coming next year anyway lol
I mean there was also previous jay "this game is fine guys evga said it was just really old cards and a very small number of them I saw nothing in there testing and never did they"
My neighbor just blew up a AMD RX 6900XT running New World. I think it's a Gigabyte as well.
It seems like the app is drawing itself at whatever resolution it gets set to in menu, and using Windows' UI scaling to resize that 4K render window down to fit on your monitor. That would explain why the Afterburner overlay resizes too. Also, the GPU is doing double duty to over-render the window, and then scale it down using non-driver-based scaling.
Honestly sounds like it's just super-sampling, trivial to do; just render scene at higher resolution and sample that down to desired resolution and display to user. That is effectively what DLSS does except without the whole deep-learning bit to prevent you from requiring the full 2x/4x required resolution.
DSR does that, which is why he made a point of showing that DSR was forced off in the control panel. SSAA will render portions of the scene at higher resolutions then scale it down, patching those portions into the rest of the scene that was rendered natively. DLSS does the exact opposite, it renders the scene at a lower resolution then scales it up to what you choose using machine learning. Most games communicate with the gfx card, which has already communicated with the monitor, and only display resolutions accordingly, with higher than native resolutions marked as DSR if available and enabled. The game is either missing that bit of code that does that, or it's somehow overriding the nvidia control panel, which isnt supposed to happen.
"app" - makes it easy to spot someone who knows nothing about computers
@@Justin-ee3im how does app indicate lesser knowledge? It's just easier to say and type and less confusing (application and program have other meanings)
@Jabba The Hutt I have a Dell S3220DGF monitor that is 165Hz 1440p natively, but it accepts a 2160p (4K) signal and auto-scales it to display as 1440p. It does this so that game consoles such as PS4 (which may not support 1440p resolution) can connect to this monitor and work as if it was a 4K monitor. I suspect that Jayz's Corsair monitor may have a similar feature.
400w in the menu!? And to think some people say "I'd never buy a mining card, I only want cards that have been gamed on!"
That's such a bullshit statement anyways. Real mining cards are undervolted to ensure ideal power to power draw ratio. The rigs usually offer good cooling too.
Meanwhile my gaming card is pushed to it's limits for the last 3 years every day starved from fresh air in my sleek looking case.
And drawing 400 Watts while on a 5 hour queue. Perfect Storm here...
I'm miner and gamer at the same time, and I definitely take better care of my mining cards than one that I'm gaming on. But there is one contributing factor, in eu MSI has warranty stickers and they're valid unfortunately. But I'm changing pads and paste every 6 months in mining cards to ensure longevity.
Hey Jayz, I'm getting rather different results setting the PL at 90 on MSI Afterburner vs the X1 Precision. Precision seems to be doing more than MSI> Not sure what is happening here
This was my thoughts. Seems logical that the msi card runs exactly as it should using msi software. I wonder what the evga card would do using x1.
Plays Civilization 2 MGE on my old mac OS 9 iMac.
INB4 we all find out that New World runs a crypto miner while you're playing.
I wouldn't be surprised. i think i guessed something similar when this first kicked off
LHR?
LMAO
Considering my OC 8600k is maxed out usage on medium settings... yea, this really wouldnt surprise me. but then again I have "old" CPU by todays standards.
Honestly waiting for that video lol
“Bordered borderless full screen windowed” XD that killed me
I like how correct that probably is when looked at from a programming point of view, I wouldn't be surprised if it looked something like "windowmode = (bordered(borderless(fullscreen(windowed)))))" I don't know much about programming so this is probably still incorrect but I hope you get what I'm saying xD
Same lmao
I made a reddit post on this , and everyone made fun of me and said “you clearly dont understand anything about graphic cards”
Sometimes you gotta let people fall
innit bruv
Hahaha, a reddit post
You know, with every modern card I've had, I used afterburner to give a small negative voltage bias and limited the temperature to 60c. You're seeing that behavior because the power % slider doesn't really do much, but the linked temperature slider does. Every card I've seen favors reaching the temp limit instead of respecting the power limit.
It is easier to predict temperature. Just monitor the power draw. It takes time for the increased power to heat the system. So you have more room to compensate. The heat can be controlled preemptively. But power draw cannot really be predicted. So managing it is limited to being reactionary.
Yep I suppose a nice undervolting reduces max power draw and eventually prevents the card from burning with NewBoomWorld...
You can un link the temperature/power-slider scale and tell it to prioritize temp or power limits you set
@@owenbar5055 They usually ignore power limits anyway and spike around as needed. But yeah it's something to try. I just unlink and leave max temp at 60c.
@@pascaldifolco4611 it's not really new world's fault. Yes they left gpu spikes in, but a lot of other games do too. Nvidia didn't give 3rd party manufacturers enough time to really test the power draw these GPU's can have and how they spike. Evga and others should know by now and newer revisions hopefully will handle spikes and power draw better.
adjusting wattage on modern gpus its like adjusting temps in shower, one slightest touch to the knob and temp... ↗stonks!
The shower thing is accurate af
When they started to release 'self overclocking' cards and drivers I remember thinking it was a matter of time before this situation would arrive. Took longer than I expected though. No safety margins anymore, now you have to underclock to get stability. Still not letting off the devs or hardware, end users should not have to be engineers.
We're not engineering anything clown, at most troubleshooting. Which every consumer has to be eventually cause nothing is going to last forever. Idiotic statement you made
Rather than respond to your insult in kind I'll just ignore you after this response.
TLDR: it is unreasonable for end users to have to know about power delivery, watts or load spikes to ensure their hardware does not die at the home screen of a game.
Software and hardware engineers are responsible for making this kind of stuff work out of the box. They have all the information end users do not have access to, to ensure their goods work properly in the field. Through negligence they have pushed that responsibility to the end users.
Software Engineer: "I wrote a game that can cause your gpu to draw more power than it is capable of using without damage just at the home screen"
Hardware Engineer: "we created a gpu that can destroy itself when the razor thin safety margins are exceeded and put very few measures in place to prevent damage"
End User: "I just wanted to play a game"
If my car overheats while idling in the driveway I am a 'troubleshooter' to the extent of telling the mechanic that "it overheats when idle". It is not my job to figure out the problem is a faulty coolant pump, because then I become the mechanic at that point.
Replace mechanic with engineer, car with gpu, and coolant pump with power delivery and here we are.
My 3060 at stock settings will run itself up to 85°c and run the fan at 100% unless I lower the power limit to 70% lmao. I lose a tiny bit performance for a pretty substantial drop in power usage
I’m gonna have to try that on my 2070 Super. Almost always at 100% usage and at 83c.
Yup. It's easy to blame a game on hardware that can't handle it.
I took your recommendations during the beta test with my 3090 on a 4k monitor. I set my power limit at 82% and undid all my overclocks. The power still spikes up to 110% every now and then, but sits stably around 92-95% most of the time. I get around 80 FPS in game with these settings. Before changing the power limit from where I had it at 104% and with my old overclock I was seeing spikes as high as 135%.
“135%” that’s crazy man o_O
@@_Scythe1 Yep, That spike was on the menu screen. I would usually see a max spike of around 120% in game. That was with my overclock settings though and power limit set at 104%.
It might be interesting to hook an oscilloscope across the power rails. I wonder if the EVGA card in particular has a power control law design problem.
sadly nobody goes that deep into this. not even the devs
@@gorbgorbon5698 that's -free- very expensive content right there
Yeah there are literally some nicely priced scopes that could be used for that in a video. But like others said, sadly going into that kind of detail isn't within the scope of a pc building youtuber (see what I did there lol).
The modern "Can it run Crysis" is "Can it run the New World's menu"
Except Starcraft 2 did this EXACT thing long before New World was a thing. So I'd say this is the new Starcraft 2.
@@williameldridge9382 no WAY you just said starcraft 2. My 2010 HP craptop was able to run it real well
funny enough its based on almost the same engine, new world is built on LumberYard which is a fork from CryEngine that Crysis used
"I hope he isn't playing New World!"
Thanks Jay, I've now scalded the inside of my nose with my morning coffee 😂
You must not talk to other people much if you though that was really funny...don't watch a stand up show, you might explode
@@samgoff5289 ?
@@samgoff5289 Wow talk about capping, person can't even comment they laughed about something? If those other people were the likes of you I wouldn't be surprised if they don't talk to them.
@@samgoff5289 I'll take your advice to heart 👍
@@samgoff5289 🤣
Jeff coded this game while flying to outer space. Mistakes were made.
Don't forget as well that people have been sat in the menu screen for hours waiting in the queues 😀
That's likely a massive key to this as well, the insane wear on cards even at 60FPS in just a damn menu...
Fyi those queues havent been a thing since the first few days
@@StevenMussels I've seen similar wear in other game menus in the past, but those ran at like 600fps at that point and kicking VSync in solved the issue.
@@Jesse-cw5pv i would say the weekend has made a difrrence on the high pop servers as of Satarday i have not seen any queues for my server and that was a consistant 2/4 hour wait
Jay: "Hey GPU reduce your powerlimit to 80%, ok?"
Jayz GPU: * now listening to: Rage Against the Machine - Killing in the Name *
How about "Take the Power Back" : P
Metallica - "hardwired to self destruct"
“Would rather have lower fps than zero fps because your card exploded” 😂
Not to mention that replacements could take awhile since there are no GPUs available.
They are using our gpus to mine
Join Atraxia
@@ELBoomerTV not gonna lie, the way the power behavior acts with pulling the TDP down on the EVGA card is something very reminesint in the mining community.
There are gpus Available out there, it just you got to look hard and you will see prices are sky high the fact is that it out of your price range.
If you have this issue with "borderless border window" when the settings say fullscreen, change to window-mode AND change your resolution. Then change back to fullscreen and set your correct resolution. This fixes the weird bordered borderless fullscreen bullshit.
It does not fix. It just fake as it was fullscreen, but it is not.
@@pinkolino ?
makes it easier to go from monitor to monitor without having to hit the windows key for multi-monitor setups
@@voodoobluetaco But you lose performance for that. They are simulating a "fullscreen". But when you select "fullscreen" what you really get is the "borderless window", which will share more resources with OS and will lose performance inside the game. If you have low specs this a huge issue.
The borderless thing happens because you guys have custom resolutions and got a new driver but forgot to reactivate "Enable resolutions not exposed by the display" that is what was happening to me in multiple games before I found out.
Really appreciate you taking the time to test this stuff and provide advice!
Does this mean we're going to switch from "Will it run Crysis?" to "Will it survive New World?"
Very underrated comment
I was thinking the exact same comparison and I love how cavalier some of the defenders of Amazon are getting about people who don’t want to chance it which they constantly say then go play something else it’s like the first thing you need to do before you purchase new world is you need to price out your video card and ask yourself is it really worth it to play a $40 game and possibly have to completely replace your video card when he has these videos he’s like well yeah they’re covered on the warranty sure yea if you just bought the card what if your card is not still under warranty your shit out of luck
at least i didn't hear crysis killing hardware, at best it would lag to a halt
at this point its safe to assume amazon is mining cryptocurrency with our gpu's while we play the game lol
9:15 thats not a bug. the great mayority of games do that today and i absolutely hate it because i don't need to press the windows key to notice it.
amazon are using millions of gpus to mine their new AI system while yur on the queue screen
I made a comment like this. I honestly think they are. No joke.
ahhh jay... everyones favorite internet dad. he speaks to us like were his sons and i love it
Yeah, except dad has been lying to us the whole time, and I've been telling him ever since his first video about this topic...
The only thing I ever got in return was trolls and haters, well guess what, now Jay is saying what I've been saying for months and he doesn't even credit the people that's constantly been shoving it up his face the whole time.
@@z1mt0n1x2 Ever considered with 3.500.000 subscribers he might not have seen your messages? Like he's willingly ignoring good input from his followers. Get outta here man
@@z1mt0n1x2 hey man maybe he's late but u can't discredit the fact that ur literally one in multiple million. he won't read your comment and he sure as hell won't read mine
@@_-TC What does his subscriber count has anything to do with how many comments there are? Today, the number of comments on that video has spiked to 6k, it wasn't even 1k back when I posted my pinged comment, taking all of that into account it doesn't seem very far fetched that he actually saw my comment and decided to make a video about it just 2 days later.
So what are you on about 3.5mil subs lol?
@@kevinjay8051 As I said to the other guy, 3.5 million what? The video was some odd 20 minutes old and had less than 1k comments at the time of writing, so what does 3.5mil subs have to do with anything?
Many games are stuck in borderless nowadays. It's an issue that often gets overlooked
It's not an issue, that's the standard and is by design. Fullscreen has become Borderless since a few years ago, since Windows had an update which made them basically the same.
Exclusive Fullscreen is the old Fullscreen, and Fullscreen is borderless windowed mode.
And that's a problem.
For example, on CoD:BOCW my FPS increases by nearly 25% by switching from borderless to fullscreen.
@@dyent This is not the norm, borderless and exclusive fullscreen are more or less the same performance wise in Windows 10/11. But it sucks that there are games out there which perform much worse for some reason.
@@Mycrosss how is that "standard"? By which design? Microsoft trying to force something on users doesn't become standard all of a sudden, just as many other dubious or hostile to user changes introduced by that company.
Haven't updated my system in a while, and can't find information about update that makes borderless and exclusive fullscreen "basically same". A source of that would be appreciated
Man Im so lucky!! was playing the game and noticed the room was heating up...this game turns your pc into a furnace !!
I’d still argue it isn’t the game. Everything the game can tell the graphics card to do has to go through 1st the graphics library API (e.g. DirectX), 2nd the graphics card driver, and 3rd the graphics card’s BIOS/firmware and firmware-controlled communication chips.
It’s a hardware problem, sure, but the software side of it is the firmware and/or driver’s fault.
I feel like the fact that this shit only happens with like a grand total of 2 or 3 graphics cards is all the proof you need it's not the game.
Indeed. Doesn't matter if it is 3090s or 6900XTs blowing up, it shouldn't happen through usage of the public APIs.
100%
Imagine if you could just make a game (or any program that uses the GPU) that could just make it blow up.
That just doesn't fly, it's 100% a hardware and driver issue.
It’s important to understand that I really hate Amazon and Jeff Bezos with an eternal passion, and I’d be happy to blame Amazon if they were at fault. The problem is that that premise makes no F-ing sense. Seriously, has nobody studied the rate of 3090s dying when doing deep learning neural network stuff 24/7?
Ok...so while running the game, the card is allowing it's own power draw to go well above what it is designed for. How is that the fault of the game exactly?! Seems to me that the cards software and bios should be limiting that and it isn't. Video games don't limit the power draw. The hardware does.
Bingo. You set a limit on the hardware part and that should be it.
the game is still coded weirdly pushing the power limit well above its limit, but i do agree that the power limiting on the cards themselves should be better
then why does power limiter work on everything else but this game?
@@aussieguy1593 but the point is it shouldn't be able to do this due to a hardware limiter which apparently isn't doing its job which is why the gpu manga of these particular cards...are replacing them free of charge as it's Their fault this is happening.
false
Jay: "Careful playing this game"
Me " Crypto mining + new world on my 3090" Let's see how long it last
The way they are emulating DSR in game (in software) is the issue. Doing a blit/copy of a large amount of data, the background frame buffer to the screen buffer, on every frame is going to work the memory controller and VRAM really really hard but leave most of the GPU idle during that time. This is the same even at native resolution. Normally you would just let the DSR _hardware_ handle the down scaling and emulating it in software is obviously the wrong way. But this is where you get higher than max power draw even though the GPU usage is only at about 70-75%. The GPU renders the frame to the background buffer, then most of the GPU goes idle while it blits/copies the background buffer to the next available screen buffer but the memory controller and VRAM light up in power usage during that time. Even at native resolution this is the wrong way. So the memory controller on the GPU is drawing a lot more power _and_ the VRAM is also drawing more power. Since the VRAM chips are accessed sequentially like raid 0 they will all be hit by the blit... In theory, this all makes sense...
This Checks out, I reckon they're doing something Sus in their rendering pipeline too
"Yea, but can it run New World without popping"
"Yeah! My GPU survived New World, noob!" I can see it, now lol.
Hey Jay, if the 3090 FTW3 is a 375W, 107% is a total of 401W, which is what it was doing more or less. I think the problem is that whenever you use the slider on MSI afterburner, the Watts matched whatever you were inputing, 90% is 337W for example, which is the total power that it was being monitored, but the percentage seems completely broken.
Also theres a difference in how the percentages are calculated, theres uhhhh Total GPU Power and Total GPU Power as percentage of the TDP limit.
The FTW3 is not 375W, the MSI card is. The 1st version FTW3 is 325W and the current one is 350W.
well, there is also the issue that, why is it pulling 400w on the menu?
probably means its pulling way more power than it should at any given time in the game
@@ZoneXV On normal bios and the oc bios the pl is 420 watts. With 107% it is 450 watts. The xoc Bios is 500 watts.
@@malonysaintpierre1599 is not the only game that pulls 400w on the menu...
when i couldnt find the 3090 ftw3 but got a gaming trio instead last holiday season i was upset. hearing jay call it rock solid really warmed my heart XD
The resolution thing is interesting. I like that they went with a borderless window because Windows is weird and a real fullscreen app has limitations, but having done that, I think what's happening with the resolution is that the Windows desktop window manager or the rendering pipeline is downscaling the window to the desktop resolution. I bet if you set it to a lower resolution, it would upscale. The size of a window is in a unit that's not pixels, so they are most likely setting the window to the desktop resolution to avoid doing math or asking the DWM to resize the desktop, and then the 3D rendering surface is being stretched to the size of the container - which means up or down depending on the configured resolutions. If you went to 1080p but had a 4K desktop resolution (which is reasonable - lots of people, myself included, have GPUs that will do 4K on the desktop, some games, and normal productivity applications, but could never run this game at 4K), I would expect all the text to be super blurry. Or maybe they worked around that by using the FidelityFX algorithm (which is open source and can be implemented on NVIDIA cards).
Jay, record the max watt on the card and you can see the max wall you spike, if a 3 sec refresh is 115 for 3 seconds, the top is probably 130 at least...
I feel like a pet dog whilst watching Jay's videos. I don't have the technical knowledge to know what he's talking about and go off his facial expressions and tone of voice to know when something is good and bad.
You have just perfectly put in to words how I feel watching Jay, I never knew how to word it before 😁
Jay doesn't have much technical knowledge either. He usually has a rough idea about what he's talking about but especially when talking about electronics he often doesn't seem to understand what is causing what and why.
I mostly watch him for food entertainment and less for technical explanations.
Hahaha !!
Question: is the over draw on power a display error in MSI Afterburner? Or maybe because the it's is doing something incorrectly? Have you tried the EVGA tool? I ask because this is totally weird and may be relevant. Do you see the same behavior using the manufacturer's tools rather than Afterburner? Don't get me wrong, I can see New World obviously has severe coding issues. Just the voltage and wattage going over the setting chosen seems like a software glitch, not necessarily a hardware issue (could be specific to EVGA cards not being treated the same in Afterburner ... non-standard way Afterburner normally polls data from the card/writes settings to card?)
The evga tool is the same thing... msi afterburner and evga precision are identical
It's a combination of hardware and crap software. Some hardware component is going out of spec while running buggy software. The combination is bad and should be patched properly, it clearly moved from one manufacturer to the other after a software patch :D I have a feeling that manufacturers did not tell the people everything about the cause of the problem too.
@@cruxer666 it's due to how Nvidia created boost algorithm based on power limit, it's limiting average power, so spikes are limited only by silicon and how much power it can pull, and in new world it happens to be more than vrm can provide and transistors blow up.
I bet New World was coded by Yu-Gi-Oh fans.
Engineer: We need to lower the power draw the game is using 150% of the cards.
Devs: No, believe in the cards!
*ahem* heart of the cards 🤓
mfw yugi pulls a 3090 out of a stack of yugioh cards
I'd like to see a test with the evga card running evga precision.
Same, Precision is shit, but I'm wondering if Afterburner isn't able to handle / interpret EVGA's implementation of power scaling, or maybe MSI themselves began doing something different (doubtful, as the EVGA cards higher tendency to self-suicide is independent of any OC software).
This was my thought as well. I know that theoretically, it shouldn't matter what software you use but I'd be curious to see if it makes a difference.
I was thinking exactly the same thing
I use Precision and when I lower the power target down to 80 it will actually put it down to 80 and not let it go over. I keep it open on my second monitor and watch it as I play New World and it only jumps to 81 every so often. Also if I lower the power on Precision the GPU usage goes down from 95-100 to a normal number around 80 or a little lower and GPU temp goes down too.
@@SlothJohnson Odd that, when I use Precision and limit power to 73% or something, I often see it sit around 75 and bounce up to 83. Really don't like this software, but having 10 profiles is handy.
I like how when he set the game to “Fullscreen” the border it used was one from Windows XP
95/98/ME actually; with that said, XP up to 10 still uses the old window style if you ask Windows nicely in code (or just demand it haha)
@@kalvino3515 Watch me do a burnout with my semi truck
Sorry, windows 95 i guess, all the old styles look about the same to me
It's a whole new world, for someone, who hasn't got a working graphics card anymore, lol. Thanks to new world, you can go outside now, touch grass, hug a tree, explore the real world. Amazing stuff :-p. You can't put a price tag on this.
What a wonderfull masterpiece 5/5 IGN
You inspired me to launch new world, thank you sir now I’m free from the clutches of my gpu
@@aiden5272 Glad i could help, pal.
Jay, I have one question: 16:58 why are the fans of the msi card not moving despite full load? The cooler cannot have that much thermal mass to introduce such a delay, right?
Don't worry about it. ;)
Could we see this testing with non 30 series cards - to see if the same thing is happening across the range of cards?
Both in New Worlds and other games / benchmarks
Been new world on my 1080 and it runs it beautifully. My card doesn't go pass 67 Celsius. My 1080 runs hotter when playing fortnite.
@@reecekordzikowski3671 same here actually. runs really well, but i also dont get into these deep tests so maybe it aint running that good anyways
@@kregman6928 I got msi afterburner on and my Temps are amazing, my gpu is overlooked already so it hits like 102% on the power which is normal for overclocking. I run it 80-100fps on very high.
if it helps my rx 6900 xt spikes randomly between 50ish and 300+w usage, my idle power usage is about 30-45w
Im using a gigabyte 1070, most of the time i can run very well, but every now and then the game will hiccup and freeze, increasing power available resolved the problem, hiccup still occurs but the game recovers and i dont have to task manage to game the close, could be a similar issue affecting my card
I'd like to see some tests using hardware watt-meters like p3 (or what the UPS shows) with a 2nd PSU connected just to the video card. This will help align what the software is showing to what the card is actually pulling from the wall.
that will only show you the difference at the wall which comes down to PSU efficiency. however they do have inline power readers for PCI connectors(PCAT) they can test draw with at the cables for whatever card is plugged in. i think that would be more what you're after.
it also close enough to use current meter(clamp style not inline) on the cables and regular multimeter for voltage and do the math for watts. -75W for the pci bus
you cannot power the card with a separate source from the pcie slot its drawing power from. meaning youd have to power the mobo and gpu from the same psu unless you use a powered riser which would limit the cards performance due to the reduced number of pcie lanes available with such risers.
Tons of factors if you measure from the wall, the software reading will be the closest you can get since it is measured from the core
I can confirm that my PC with 3080Ti FTW3 is creeping right up on 600W draw from my UPS power monitor before I turned down the power limit on the card. I have never had a game cause my UPS to freak out over power draw before this, but it throws constant warnings if I don't turn it down.
Tbh I think its high time Puget Systems gets a blog going on this. If you're going to look for actual science on the matter, and not user-experiance focused content, then turn to the actual science guys over there to give us a true breakdown of what is happening to the hardware.
I wish devs would go back to optimizing their games.
P.S.
I'm not just talking about this game.
The cancer of PC gaming. The "Just buy a better PC" mentality of devs
@@Reac2 Good thing most games are made for consoles then...
@@movement2contact ah yes the golden experience of cyberpunk on consoles.....
@@kenker_ape Was it good on PC?
We get undercooked games as it is.
Practicly all games i own, show resolutions the monitor does not support, why call something a bug, that isn't one?
8:57 seen this behavior in a lot of MMO's I've always had higher resolution options for WoW in particular before DSR existed.
This is called super sampling. Normally this is under a different setting, named "render resolution". I was unable to get what JayzTwoCents got for 4K, on AMD RX 6900 xt in New World or anything higher than my monitor's resolution. Same thing on my GTX 1080ti as my secondary card.
Yeah exactly. I really don't understand what he thinks he is proving here other than stating the obvious.
@@breakbend7707 because a 4k should not exist under the settings unless you have a 4k monitor same for me by default...4k was enabled I do not have a 4k this could be a major issue that's causing crashes
This game runs like trash at 4K and 4K is the default option enabled on install regardless of monitor. The fullscreen windowed mode also means that v-sync, g-sync, and free-sync isn't going to work properly. (Edit): New World also doesn't let you select your monitor either, so if you have a separate gaming monitor you have to swap it using hotkeys. The way it deals with displays is completely broken.
@@RPD-SWGBeyond this is nothing new, other games behave the same way.
It could always be the way MSI afterburner is interacting with the evga card. We always kinda assume it just works with everything, but maybe that specific card doesnt properly interact with MSI afterburner? It might be worth it to test it with the evga utility you mentioned, just as a check. Of course the evga card also was having kinda weird behaviour when msi afterburned was on stock settings, but you never know.
amazon is smart.
selling an item that brakes a previously purchased item which causing you to buy another item to continue to be able to even play the item which will eventually re-brake your item and the loop continues
The peak power draw is probably even higher than MSI reports. By up to 50%. It looks very much like a power line resonance. But that'd be only a theory.
As for a mass on a spring, the higher the mass the lower the resonance frequency. Here we have a humongous current to draw, but fairly weak power lines.
What New World does is it's constantly switching graphics context in order to draw it on screen: it first draws the scene which brings GPU to performance mode, but then it switches to window context (that being the Windows UI) and just copies the image over it. Since the last one is a fairly simple operation GPU drops the power state to low(er).
That cycle repeats every frame, switching back and forth from high current to low current. That causes resonance. In most cases that wouldn't be a problem, but this is not a typical case, we're dealing with a small room heater class GPU. For a low power card that resonance frequency is fairly high and the GPU is set up such that it will never get close to that frequency by switching power states. But for 3090 with a giant number of coils and capacitors (considering f_r=1/LC) it looks like that resonance frequency is in the vicinity of 200 Hz. Maybe lower.
Why's that? I've seen a case when some streamer REDUCED graphics settings so his FPS shot to 200s and the card instantly popped. Because the power lines resonated. Why don't other games fry those cards then? Because they don't force the power state by changing drawing context.
How to resolve it if that's the case: manually lock power state to P0. Program called nVidiaInspector can do that. It requires a reboot and it makes the GPU run hot on idle, so... it's a temporary solution.
Again, it's only my guess, but I'm fairly certain it might be true.
The reason why the power draw on the EVGA cards may be inaccurate could be based on how they measure the cards power for their own reference board. When I use EVGA's precision X1 software it never goes past the power target I set it to. With afterburner it seems to be inaccurate.
Yeah this probably explains why I can pull higher stable overclocks with PX1 than afterburner.
using the intended software to increase power limits would be my next step of testing too. please do a follow up Jay
I tried x1 and msiafterburner for 3d mark. Pretty much the same results and average clock. I use hwinfo for my gpu power draw though.
It also happen on Gigabyte cards, it's basically happening to every card with a 400-500W bios, it's not about precision software.
A GPU should't explode for an OC software anyway, there are protection measures in place at the hardware level, those are clearly failing on 3090/3080
@@lupintheiii3055 Also the countless AMD cards that have been reported to get smoked by New World.
You could try to run Unreal5 Demo AncientWorld and bring the console command (~ key) and put t.maxfps 144 (by default it's 60) and after r.screenpercentage 300 to run it a 8k (150 if you want 4k) on your 1440p monitor and look if your EVGA card burst into flame!
The thing is though. There was absolutely no issue in their beta builds. So they messed something up in the code on launch day
So, funny thing with my experience. Specs are an R5 5600X and 3070 Founders Edition with 1440p monitor for reference.
I was watching the 'Something is still terribly wrong with New World' video after just having bought the game cause I remembered seeing it on my recommended. Installed the game, launched MSI Afterburner and set the voltage to 90%, then launched the game. Seemed fine, got to the menu, went to settings, saw it was at 1080p. Set it to 1440p, turned on the FPS counter, and was getting 62 fps. Setting were defaulted to high with no max FPS, so I left it at that.
Went back and to start playing, got to the server browser, was looking through them for a minute and was about to join the one that a couple of friends were already on, and crashed to reboot. No 100% fan or anything. After reboot, I set the power limit again to 90%, then went to Nvidia control panel and set the max fps for New World to 60, and tried launching the game again.
I got to the menu and crashed immediately to reboot, like the first time. After another reboot, set power limit to 90%, and then put settings back to default in nvidia control panel for New World, and tried launching the game once more. Loaded to menu, went to the settings menu, and then crashed again.
BUT. This time, it was powered off, no automatic reboot. After a couple seconds to process what just happened, I went to power it back on, and got 100% fan for a few moments. I about shit my pants. Luckily, the mobo logo popped up, then the Windows loading circle, and I was back at the login screen. Fan speed went back down to normal, logged in, and everything seemed fine.
Immediately uninstalled.
If I was in a scenario where the power delivery of my GPU potentially was underspecced to a level where a load of a game such as New World could make it 'popp', then I would want to play the game to see if it actually does 'popp'. Even if it's just a small batch or whatever of faulty cards, I would want to run my card to ensure that I don't have one of those bad batches.
I'd rather have the card fail while it's still under warranty, rather than have it explode a year or so after the warranty has run out.
problem is there's a massive gpu shortage currently
@@emiliskog I have a hard time to believe that warranty issues wouldn't get queue priority though. Anything else would severely hurt the customer-brand trust.
@@mackan072 they proably are prioritised but that doesn't mean it'll be quick due to limited supply due to said major current shortage
@@mackan072 It could still take months until you get a new card.
@@HeavyMetalGamingHD I'd rather be months without my GPU, than have it break outside of its warranty period, especially given the costs of these GPUs
You are a legend Jay. Thank you for all the years of work and information
Amazon: Buy a 3090 from us
Amazon: Buy new world from us
Amazon: Buy a new 3090 from us.
Bezos: That's what we call a pro gamer move.
Jay comes across as much more authentic than Linus and co.
This video is a reminder to those whom love to think, "That'll never happen to me. Pff!"
Life will tend to creep up and surprise you. Spoiler, it's rarely the good kind of surprise.
Moment:
I forgot toilet paper.
Store:
We are out.
Thats exactly why im not even touching its store page with a 1̶0̶ ̶f̶t̶ 100 ft pole
Interesting how the stability differs from the EVGA to MSI. Only thing I got to wondering while watching this is, what if MSi Afterburner is what is causing the inconsistency with the EVGA card? I have an EVGA card, I don't like using EVGA's software either, but have you tried EVGA's precision Overdrive software with the EVGA card for tests to see if it has the same inconsistency? I mean, I know it shouldn't really matter, but maybe the OC software between MSI and EVGA has their own way of "efficiently" communicating with the card? Idk? As much as we have seen software's influence on how the hardware behaves, it's just a thought. I honestly think there would not be any difference between using one OC software over the other, but who knows?
huh how weird how an MSI VIDEOCARD!!!!!!!!!!!!!!!! works better with software like MSI!!!! afterburner. then say something like a non msi card.
yes it would make seans
I had the exact same thought
I share your opinion.
They are both cheap pos s , never see problems on Asus products. You get what I pay for
MSI video card behaving better using the MSI app then on a non-MSI card? How amazing! lol
7:32 You can already hear him becoming very aware that his GPU might die in front of him
"But also *Linus"*
I died lmao
I reckon this confirms my thoughts on the firmware power management... the firmware polls the shunts on the card on a regular interval and manages load not by reducing input, but by throttling the GPU internally. If poll rate is too slow, it will fail to detect the peaks that occur as spikes. The overshoot is either bad firmware, slow polling, inaccurate power monitoring, or a combination of all 3.
i have watched 60 percent of your videos im running into an issue where i liked a video and go back later to watch the video again the like is unchecked
I started playing New World around the time this video dropped. If we only knew then the true extent of the “coding problems” 😂
Tested this last night on my 3080Ti FE. with everything at stock, the power limit never went more than 105%, averaged about 85%. When I maxed out the power limit slider (114%), I saw it spike up to 122% once.. I'll leave the power limit slider at stock just to be safe.
You got balls for testing it.. I would never put over 100% on this game
Damn, I have played whole time with 115% power limit with the same card. Istantly turned off. Thanks for testing!
As a former QA engineer yourself you should have tried the EVGA tool when you noticed Afterburner settings working sort of as expected with the MSI card just as a sanity check.
I guess that the problem is on how those tools interact with the firmware of the card so setting it to 90% on software, for some reason, might not be 90% at a firmware level.
I agree. Fellow QA here, people usually don't think that way so those findings are just one piece of the puzzle.
@@kenneth4683 I'm a developer myself and that's one of my pet peeves. People have no idea how to report/test issues...
@@jvmbatista Yes lots of assumptions are always made, even on these type of videos (not saying that these videos are bad, still informative). it's really not clear what is going on.
For what it's worth, New World has not yet fried my GPU on stock settings (MSI RTX 3090 Suprim-X), but I have had to replace TWO Corsair RM1000x PSU's that previously had no issues at all. They began wining when the game started, first thought it was the fan, but then boom, dead.... TWO of them.
Big Goldenseal on the box and, "This 3090 card is certified to play New World at max settings and will not harm the card." That's what I want to see.
8:27 Because you don't need DSR/VSR if the engine implements a custom framebuffer internally for rendering into before final output.
Its the same thing as Resolution Scale on most other games, no?
@@GucciGoblins Yeah it is! DSR is just for games that doesn't natively implement it.
@@GucciGoblins Basically, only real difference is whether it happens on driver level or game engine level, but the idea is indeed the same.
Engine level has the advantage that the HUD can be scaled and rendered separately after the game so you won't end up with tiny text on the HUD. So can't really blame new world for doing it better in that sense.
Reason why they default to 4k on a lower resolution screen with a beefy GPU is likely since it also gives the best possible AA.
@@Megalomaniakaal Yep, on RPCS3 I have resolution scale set beyond the native resolution. Exactly because it smooths out the image, even if the particular game doesn't support anti-aliasing
Yes but it's very weird there isn't an option to choose output resolution. Game is locked in to whatever res it believes your monitor is in, and if that's wrong for whatever reason, you can't change
I'm curious whether or not the spikes being seen on the GPU's power monitors are actually telling the full story, or whether we're just getting averages/point in time numbers and seeing much higher spikes that are in excess of the numbers being displayed. I imagine that'd need far more sophisticated test equipment to determine.
You can change the settings for faster measuring to get a higher 'resolution'. Jayz measuring is fairly slow
sounds to me like good 'Ol Ame'zon made a game just to break video cards / computers so they could sell us more ... good job it is "working as designed"
If you've become aware of this New World problem, WHY would you continue playing at all? Amazon and Google both have the same game dev issues, as reported by devs, they run game dev like it's NOT game dev, but like everything else business-wise, and it doesn't work, it leads to shit like this. They won't fix this, or care, so why would you continue supporting the game with your playtime, while playing GPU roulette?
Can't tell you enough how amazing I think it is that you are taking the time to test all of this. Thank you.
I would like you to see if the EVGA X1 does the same overshoot of the power draw. I would like to know if it the software of the program causing the instability?
Jay: reduces the power limit
EVGA GPU: fans go BBRRR
BuildZoid got one gigabyte 3090 "fried by new world", and though he notes that he has just a sample size of one, it seems that the asymetric design of the vrm on the reference board is causing the issue, because on the card he has, the issue seems to be that one of the vcore power stages got blown. He is going to try to fix it. 10 Power stages for 8 phases seems that the controller is having issues balancing the load, because there should be no need to balance that load.
by the way, that doesn't seem to be the issue. the issue is that the card violates the spec so much that vrm's eventually give up and die.
I swear that game feels like most of the devs were hired directly out of ITT Tech and have never worked on a game before. I have a good time playing it, but JFC it's got so many bugs.
Many many many bugs. I've plowed a stupid number of hours into this game. There is a gem of a game hidden in all the faction mechanics, PvP, and crafting/economic systems... Just jeez the bugs.
@@Flecco "It just works"
In my experience all of amazon software is buggy, so I would have expected as much
Welcome to Amazon's programming on EVERYTHING they do. It never improves. Ever.
Game is trash
Honestly better than what I was expecting when Amazon's venture into the gaming industry was revealed.
You realize that the problem was in GPU manufacturer's right?
This is 100% gpu problem. No matter what a game does, the card should not self destruct. Same with CPUs, no matter what the software does, they cannot self destruct.
are you really saying that Crucible exceeded your expectations? That game died so fast nobody even knows it existed
The ability of scammers to report people and get them banned is obscene
As soon as I start mining in new world, my 3090fe goes crazy, it’s like it’s literally Bitcoin mining 🤣
go try fishing.. My card crash every time there
Jay that Corsair Monitor actually does support 4k, the specs of it lists "Native Resolution 2560x1440 (16:9)" and "Max Resolution 3840x2160 (16:9)"
"Downscaling" is specified in the monitor's quick start guide appendix.
I was going to say that. Some 1440p monitors support 4K for using with a console such as PS4/PS5 where you can either choose output at 1080p or 4K, no 1440p.
It usually only works via HDMI though.
Jay not knowing this shows how much research has gone into this piece..... LOL
So the panel is 4K?
That's not 4k. That's a 1440p image stretched up to 4k, aka upscaling. So no, it doesn't actually support 4k resolution. Imagine trying to insult somebody and being so clueless you make yourself look ridiculous.
Jay this monitor supports 4K and downscales it to 2K. Mine does too.
you are right
He really should cut this part out of the video and republish. clearly not a bug in that case.
Thanks for that. Was wondering why it was a selection if it's not a 4K monitor.
The question is would he know that. Would the average user of that monitor know it?
@@HappyBeezerStudios It's not a well known feature of these monitors indeed
So apparently this situation is really being swept under the rug huh?
Not sure how using high-level graphics APIs would lead to bad code burning a card out. Considering the prevalence of indie developers playing with free game engines, you'd see this happening more often.
The EVGA card seems to be computing the power percent based on the stock power limit, not the boosted limit it's built with.
"Poorly written game..."
No.
Well. Rather...
No matter how poorly something is written, software should never *ever* be able to cause permanent hardware damages. If said hardware do break down due to some "unnatural load", that hardware is flawed.
The exception is of course when the software is set to overclock the hardware. That is the users own fault though.
When you use software to overclock hardware, the hardware level protections should kick in, that's why you usually have to tamper with way more to modify the behaviour, like having a modified bios and then doing stuff like shorting the shunts. Whenever you use stuff like PBO in say amd CPU's, the software will tell you about the warranty, but will still make the cpu work on safe parameters, provided your configuration supports them.
Amazon develops a game that kills video cards, and then people buy more video cards on Amazon.
Maybe board manufacturers need that specific code set to use for testing to try and find weak components.