Yes, and do it with green, red and blue. Especially with Intel b/c it will be down right funny, some of those games have improved over 500%, we all know their day 1 was rough.
Use a drive a couple of versions in from the gpu launch, and compare with todays. These feels like Gameready drivers for new game launches, as it is 551.xx in both tests
they don't do planned obsolescence with drivers, they don't need to. there are plenty of other ways the planned obsolescence kicks in, vram issues, longevity of the card itself, etc.
Or just collaborate with game developers to push the envelope on what's needed to run the game. I mean, CP2077... or Avatar. Avatar's got that "unobtainium mode" (or what it's called), which goes beyond ultra. When AMD and Nvidia release their next hardware, people are going to buy their most expensive GPUs just to have their game at the highest fidelity... and also bragging rights. ^^
"You got to buy the latest GPU from us because the following games just don't work as well with your older card" BS. Maybe if the game developers made GOOD code instead of slapping it all together and relying upon the GPU drivers and the bells and whistles that comes with the expensive cards to sort it all out, we wouldn't be having this discussion.
indeed. had a 570gtx for 7 years before the Vram was too little to open pretty much anything new. but the things it did run was smooth af. so plenty of power left.
@@silverwerewolf975 that's kind of a separate discussion because that's up to game devs, not the gpu manufacturers. Then there's going full tin foil and wondering if nvidia is supporting or even paying for poor optimization on older gpu's.
You can see the lighting in CP2077 is also different. Looks like this person intentionally turned up settings to create a fake frame drop for clicks and attention.
its always what these channels do to generate negative attention / clicks. wouldnt surprise me if they intentionally lowered the core clock & vram clock speed on the gpu too. or forget to restart the machine after driver update / re-apply the overclock for the gpu or just let a chrome browser open on a 2nd screen & u will get the 10-15 fps difference i never had any driver that lowered performance except in 2 cases - 1st did lower my fps by around maybe 5% but got rid of the horrible microstutters and 2nd case was actually a driver issue they fixed after ~2 weeks that lowered performance in 2 titles i played. every driver since the gtx 9800 times gave me more performance, stability or nothing in specific titles i played but never willingly lower performance (cant speak about AMD though), some game updates however completely tanked it which got fixed mostly too.
@@squidwardo7074 One game, maybe. CP2077 is notorious for resetting your settings, but _every_ game? No way. That HAS to be intentional for either a narrative or agenda. I'm guessing that agenda is clickbait or just straight up Nvidia bashing.
What people often forget to do is double check the settings in the control panel. When downgrading the driver version, it also installs an older version of the control panel. If the version they had installed before added new settings or changed anything, it might cause that setting to change or go back to default. I even had it reset all changes made to the settings. That alone can produce either higher or lower performance, depending on what setting(s) changed. If you do not verify this, you will not know whether your results (old vs new) will even be comparable, regardless of game or test scenario.
This is a good point! I always do a clean install to zero out any issues like this or corruptible installation files from previous to newer. I know my settings and can always change them back again.
I always double check settings like some OCD crazed man with things like the control panel but even ingame. The amount of times that shit resets after an update or even the slightest file change is infuriating tbh
@@DarkNia64 This is literally me 😂 It all started with my system and how often I had to re-check shit to be in its proper place. Now I fucking check everything twice in life. It ain't even fun anymore..
The lighting in the first video @5:17 for cyberpunk2077 is vastly different from the left to the right. That's not just a driver version change. There are other changes as well.
Good catch. I could see one title being an honest mistake if they didn't notice settings changing as CP2077 is notorious for this, but _every_ title? I smell some agenda tampering here.
Saw this too. I thought for a sec that it could've been atmospheric difference... Would be pretty extreme and make the benchmark totally inconsistent. Something is definitely up
Dwindle FPS did a heap of testing on Cyberpunk and found Screen Space Reflections, Mirror Quality and Volumetrics affected the performance the most. If you look at that right side the ceiling light is shining through a thicker volumetric fog. That setting alone has around a 30% performance impact.
You and multiple others have done enough testing to show that driver updates don't bork older cards, and i appreciate yall continuing to keep an eye on the topic.
You do see this a lot when people still have much older cards that nviida just dropped support for. So the nvidia 300 series stopped getting updates in 2016 so no new driver for them. The 500 series stopped getting updates in 2018 everything after the 500 series is still getting updates.
planned obsolescence is when you build a product that will fail or be incapable of working as intended like putting 8gb on a 70 class graphics card. this will cause you to buy a new one. force obsolescnce is when you update a product, so it will run slower and causing you to consider buying a new one.
You're gonna need to qualify that a bit more. 8 gb on a 770 would've been massively overkill for no good reason. Same with a 670. And you're also missing the fact that a card that can run a specific game at a specific performance level will almost certainly be able to run that same game at that same performance level for the entire life of the card, unless updates massively change the game. You might as well complain that your AMD K5 can't run Cyberpunk 2077. Just because technology advances doesn't mean it's planned obsolescence.
@@BobBobson I'm pretty sure he was talking about the RTX 3070 and RTX 3070 Ti, not 770 or 670. The first 70 class card with 8GB VRAM was the GTX 1070 from 2016 and the last one was the 3070 Ti from 2021, finally in 2023 the RTX 4070 (Ti) got 12GB
Yet most systems running a 3070 will be designed around it and therefore couldn't really make use of the newer games on the settings more ram would allow anyway. Vram requirements have been shooting up for many years - long before the 3070 came out. If you buy a card thats only good enough for average performance on todays games why would you think its going to do that on games years later? A certain big builder who shall remain nameless used to pair them to i3 8100 and 9100's depending on application. I wouldn't expect those cpu's to run CP2077 on high everything either.
the biggest thing that happened in the 551 driver branch is the fixing of v-sync micro stutter and it's great for people who were in the 537.58 driver for a long while.
@@kevinerbs2778 I didn't experience it too much in my setup compared the others. I did see in every update pre-551 drivers in a lot of reddit posts that the micro stuttering with v-sync + g-sync/free sync was causing people to have issues while playing games (think of a jumpy fps graph in msi afterburner while v-sync is enabled in the driver type of issue). I guess it depends on the setup. A lot of the people complaining had 4090's, 4080's, and some 30 series too. My 5600x/3090 setup wasn't entirely affected since I run at 4k60 capped at 58 fps with vsync + g-sync, but I still did notice bigger stutters pre-551 drivers that were fixed on 551.52.
Without going into a bunch of detail, I have an RTX 3070ti, with the newer drivers I have been able to finally get rid of the micro stutters and get better framerates when aligning my monitor settings and v-sync. I have MSI monitors with adaptive sync and have battled micro stuttering forever. I can now finally eliminate most of the stuttering.
The only industry I know that doesn't really have any very obvious planned obsolescence is the enthusiast and professional camera market. Camera gear is expected to last for ages. Cameras like the good old Nikon D850 are built like a tank. You could knock someone out with it if they try to mug you and just keep shooting. With modern mirrorless cameras they are even removing the last couple of moving pieces (the mirror box and even the shutter curtains in some models) so the chances for a failure are even lower than before. Many brands support you with firmware updates for years and it's not rare to get your warranty extended to 5 years. I've been a photograpger for about 14 years now and I have yet to see a single piece of my gear fail on me (except for one very cheap SD card but I got that one from a crappy manufacturer). When Canon and Nikon introduced their new mirrorless systems they also sold you adapters (separately or as a bundle with a camera) so you could just keep using your old lenses. They made it clear that they would love it if you buy their new lenses because they are way better (which they actually are) but nothing is stopping you from just using your old gear while swapping the camera body only.
Nope, Cameras have "planned obsolescence" in the form of batteries. Now, I don't believe most products have planned obsolescence so much as they have planned no longer supporting the hardware/software/device. And if there is a market for it almost everything can have a 3rd party aftermarket parts. Unless you are malicious and locked out at the software level. Which most companies don't do. Well, most companies except Apple, John Deere, and car manufacturers.
Phoebus Cartel, established in 1925. Major lightbulb manufacturers joined together in agreement to not produce lightbulbs that lasted beyond the agreed allotment. Planned obsolescence.
This is not the slam dunk you think it is. The reality is that while it looks bad. There are good reasons to do what they did. And since no one can prove why they did it we have no way of knowing. The main ones are efficiency and heat. The only way to make an incandescent bulb last longer is to make it run cooler and because the way an incandescent makes light is black body radiation you get dimmer bulbs on the orange-red end of the spectrum that generate a lot of heat but not a lot of light. So you would need more bulbs to provide the same amount of illumination but you would be generating way more heat while doing that and increasing the chances of in-wall electrical fires. It could have been planned obsolescence. or it could have been to make the light more attractive to customers by making it brighter than candles. or it could have been a thing to reduce the energy required since the grid wasn't as large as it is now. We will never know.
Thanks for fact checking! This would be an interesting series to keep up - testing continuous/older generation cards and keeping a tally of improvements with modern drivers.
I'd love to see a driver test with older cards such as a 1060 or 1070 for example. Would be out for 8 years this summer (for the 1070 anyway) so maybe do a launch/2016 driver and a driver from every 2nd year up to the current driver at the time so 2016->2018->2020->2022->current and see what the diffrences would be with the drivers.
I had some major glitching, artifacting and general graphic bugs on 551.52, to the extent that I spent a few hours troubleshooting with NVIDIA customer support and ended up downgrading to a much earlier driver. Just updated to 551.61 today and the issues seem to be gone. Whatever the issue was, it has definitely been resolved for me.
The problem with software is, that the combination of lets say your OS patchlevel, different divers and versions, software, etc. might cause strange behaviour. So "downgrading" might be the solution for a very individual problem that looks like "planned obsolescence" but is just a coincidence.
@@hw2508Oh I'm not saying it was "planned obsolescence" or anything of the sort. I presumed that the driver files I downloaded for 551.52 were either corrupted in some way or there was something else going on with my system. I was actually implying what you said, that each system can produce different results based on a wide variety of potential variables.
Yes, please do what you proposed at the end. But with older hardware. To me 30 series isn't actually that old. I have a 1070 still. That would probably be more of a target for planned obsolescence, if it exists. The test that was done makes no sense in my mind. It's basically the same driver with fairly modern hardware.
That gpu is almost 10 years old now, at this point any performance drop would be from thermal paste or something like that being old from use. Would you count that as planned obsolescence?
@@MSusername69 You're just assuming they haven't changed that old paste? I do that with my older cards, hell I did it to my 3070ti not too long ago and it dropped 5c. Such a silly condecending response to someone asking for legitmate content, be on your way.
Newer drivers just don't support old hardware, but they do continue to provide access to the last drivers that supported the old hardware. At least the 10 series is still supported by the current drivers. My previous card to my old 1080 (the great 780Ti) is 474.82 [my current card is the 3080 10gb]
@@ihateeveryone8161 His argument is pointless anyway, because Jay would test different drivers on the same card and the card will not magically age inbetween tests, so paste or no paste makes no difference.
@@ihateeveryone8161 and the same to you. I was just pointing out something that could happen to older cards. I could also point out that some of the new software might not even work on the old cards, so the tests wouldn't even prove anything. Unless you call companies stopping support for older hardware planned obsolescence. What about the silicon degrading overtime? Is that the same thing?
Of course Planned Obsolescence is a thing, Going to Electronics Engineering College 40 years ago, Mechanical engineering one of my first Engineering professors was GM Engineer in Automotive, he told he left after he designed a great Transaxle that would last a miilion miles at 99%, they told him to redesign it for 200k at 90% because they make $20k for the Car BUT 50k on Car Parts..
Agree. However, at least for me personally I’ve always upgraded video cards when I find them obsolete. Which is usually about every other generation in my case. Usually because the games I’m playing are more demanding/larger/more players and I’m not getting the performance I want.
Planned Obsolescence is ABSOLUTELY a thing in many industries. Cell Phones (and anything with a non consumer replaceable battery) are the perfect example
There are so many of these channels that claim what ever they want to claim with zero proof of the actual hardware being used. It's all intentional bait to get views.
How smooth brained are you? The point of the video is to show the amount of variables in any give test setting, doesn't mean they're fake (they could be but you also can't say for sure which just makes it conjecture) more so that some people don't put in all the forethought to make test samples as constant as possible.
@@solarix I think you're giving people too much credit user error is a very common thing, while some people sure could be nefarious they are far more dumb people out there.
A friend of mine gave me his old computer, a Ryzen 9 3900X with a 2080Super, the computer had launch drivers, and I could swear I saw a huge performance upgrade when I updated to the latest drivers both on GPU and CPU.
See if you’re an AMD gamer like me, you would know that if a new driver has issues, it’s probably “AMD being AMD and launched a buggy driver again” It’s easy to jump to conclusion with these things, remember CZcams slowing down with Adblock that turned out to be an Adblock issue? If Intel’s story is an indication, drivers are difficult things and buggy stuff happens. There’s also BIOS updates, OS updates, and even overall how good the hardware is maintained, there’s so many factors where it’s hard to find proof for this unless if it’s really blatant
If you're going to do the launch day driver comparison to recent one, I think including the specific games game ready driver would be interesting as well, although I don't think the difference would be significant regardless.
I doubt they're purposely _slowing_ down older cards, but they absolutely stop optimizing for them long before EOL. It's ridiculous that AMD EOL'd Vega when they literally still sell laptops with5000 series Vega APUs in them *_TODAY._* That should be a class action level offense.
Obsolescence is certainly something some companies try to do, take phone batteries for example, they made it much harder to replace your battery, the first thing that starts making your phone much more annoying to use after a couple of years. They could keep the same great looking backing, but have multiple tiny holes, like the one you use to get the sim card out, around the device, when you push them all in the back comes off. My phone I have now I considered getting a new one because the battery isn't as good as it used to be, but then I ended up buying a new battery instead... The battery is still sitting here because it's a huge hassle to change it.
I remember when i had the 980ti classified when the 20 series came out and omg idk if it was just me but every other update would cause major issues and i would have to roll back sometimes 3 updates for it to work
Not so much planned obsolescence and end of life with no future support. Eventually, the time to maintain the code base or manufacture parts is not worth the money it takes to make it. And if there is money to be made by supporting these older products you will get aftermarket solutions. My d5000 no longer has OEM batteries. But I can still get batteries for it from after-market manufacturers. same with my car and most things. Almost everything today can still be fixed if you have the know-how unless it's made by Apple or John Deere and maybe some car manufacturers. Apple really likes to use serial numbers for everything to keep you from fixing your own stuff.
Billion percent. People are brain dead if they can't understand this. It started with light bulbs in the 50's. They've been pulling this shit for half a century.
Planned obsolescence does exist in the GPU market, it's just in the long game. When they give people small bus width's on cards knowing that not everyone has made the 1440p or 4K jump yet, or when they offer VRAM capacity that can be an issue later down the line. Or when they purposefully nerf a platform for future releases. Things like that constitute planned obsolescence in my opinion. Driver/software side I heavily disagree that AMD or Nvidia have any other interest but that of improving performance or keeping it up to par at worst. Considering their are a lot of times they without improvements on the driver side because of new generation GPU releases that they want to have exclusive features on, etc. What makes it planned obsolescence and not just evolving product lines, is that there is intent in nerfing specs for that very purpose. Intent matters.
I'm old enough to remember when Nvidia had issues with older generation cards getting worse and worse over time. I'm also old enough to remember when they sold a product advertised as having 4GB of VRAM and oopsy they shipped it with 3.5GB.
@@prima164 The iPhone 4 last iOS version was 7.1.2 in 2014. Another reason it wouldnt stay connected to a cell network is because 3G was being shut down, which is what the 4 primarily used. Also, phones tend to slow down with time is just lack of optimization from app developers/the manufacturer. Apple, or any other company that does software development is going to spend tons of man hours to resolve issues for a device/product that an ultra small population of their users end up using. Does it suck? Yeah, totally. But for any business, they can't support devices in perpetuity. There comes a point where you have to rip off the bandaid, even if that means upsetting some users.
@@RAZGR1Z Old enough, come on, it's not like it was in the 80s xD I had that card and it was only 10 years ago. It was GTX 970. They didn't ship it with half GB less. One memory chip was defective and was operating at only 10% speed so later on when the bug was discovered they had to turn that module off via drivers hence the 3.5 GB total memory. Never had a problem with that card and memory was never the issue btw. In fact it lasted me 7 years, swapped to 3060ti like 3 years ago.
Only had to roll back a driver once and that was when my frame time tanked in Horizon 5 on my 4090 and tripled. Rolled back and all was good, since then I've kept the driver up to date and using DDU each time and all has been good. Lately seen an FPS boost in Horizon 5 at 4K.
this is a clear test that these driver update did not brick a recent (last gen) graphics card. but as other people are saying, what about an actual older card? like a 1060 or 1070 with current drivers. also what about last year drivers or something like that. i understand it would take a lot of effort to test that but it would make it for an actually comparable planned obsolescence test
The channel thegamrone has a video on their channel testing a 1080 Ti in November 2022 comparing its 378.78 release driver to the then current 526.98 driver across a variety of games, the current driver actually showed in an increase in performance in some titles while others performed basically the same. This lines up with my personal experience with my 1080 Ti I've had for over 6 years now where performance feels very similar, if not a bit better, with current drivers (well technically my drivers are 6 months old being version 537.13, I only like to update when I feel I need to, but still relatively recent) compared to the old ones.
I have also found that as time goes on the performance of old NV GPUs improves. I ran a 280GTX system right up until I replaced it with a 980GTX and did not come across any performance drops in older titles that came out around the time of the card. The same when I had my 980GTX running until I managed to get my 3080Ti.
From my own experience I have to agree with you. I've been using a MSI 2070 since 2020 and I have kept up on the updates and I have seen no noticeable drop in performance since new. I also take my system apart approx. twice a year and clean it out. I do a fresh install on my system at least once a year also. There is still no noticeable difference from when I first built this system and the way it runs now.
As a mechanical engineer. I've quit working for companies before as i was explicitly told to design a mechanism to fail after 2 years when it was legitimately cheaper to produce a mechanism that woild last 5+ years. In most companies "planned obsolescence" is a matter of cost optimizing a product to last for x timeframe. But unfortunately many companies do design products to fail after a certain timeframe specifically to increase sales. Software companies doing the same thing would not surprise me at all.
It's never been a thing in computer hardware, the tech moves so fast that time does the trick just fine. Tho we're most likely hitting the top of the S-curve when it comes to hardware requirements for gaming, so it might be a good time to start looking for planned obsolescence as it may be introduced into the equation sooner than later.
Oh, but it is a thing. For example, when NVidia released 30 series, driver target temps went from 73-75C up to 85C, more than enough to push well cared for 10 series cards junction temps into vicinity of the danger zone of 95C, Enough to literally burn out the die if it doesn't just trigger thermal cutoff every time you run a full screen application or intensive task. And 40 series comes with it's own short fuse planned obsolescence device called the 12vhpwr plug, trying to run MORE power over LOWER transference contacts, a plug that is literally trying to push 3 8-pin connectors worth of power through less contact area with higher ohms than a 4 pin Molex connector.
@@HFRG-zq1qm Counter argument: Was that the driver, or was that just GPU's thermal conductors growing of age (like if you don't repaste a CPU it burns up), newer games that were tried; newer background operations. There are tons of factors. Annnd what does 40 series running more power have to do with anything, in particular about planned obsolesence?
@@thewoodenlego5379 As to your counter argument, the driver literally changed the "Optimal target temp" from 75 to 85, and any older card that has not been repasted is very likely to burn out from the increasing ambient to junction delta, being as the minimum delta is around 7-8C (out of box or freshly repasted), and with that, 85 puts even a freshly repasted GPU dangerously close to thermal cutoff (TC is 95, 85+7 is 92, a slight spike or trough in power can trigger as much as a 5 degree C difference either way, , and 5 degree increase at 3 degrees shy of death is past the point of death). And as to 40 series, the built in obsolescence fuse IS the higher power draw, done over a connector that isn't able to handle it, a built in self destruct mechanism. A 4080 draws the same power as a 3090, but the 3090 doews so without burning up over it's 3x PCIe 8pin connecters, the 4080 does it over half the total contacts at half the size each, meaning it has 2 times the natural resistance to transference and thus half the capacity, and that's before you even figure in the reduced contact surface area, and energy which has it's transference impeded puts off resisted energy as heat.
@q1qm No clue wtf you're talking about mate, 10 series have always had their temp targets at 82 degrees, and that target is just where the card throttles itself, it means absolutely nothing unless your cooler is hot garbage because it's the fan curve that determines whether or not your card reaches that temp on a cooler suited for the wattage of the card, and that curve is set by the manufacturer by default. It's also customizable in any OC software like afterburner or nvidiainspector, where you can clearly see that it's 82 degrees target by default. Dont make shit up to argue strangers on the interwebs mate, that's fkin pathetic. 12vh is a failed solution to a very real problem: a GPU shouldn't be needing 4 cables from the PSU.
@@rekt4guud134 82C has always been the point of thermal throttling, yes, but the target temp varied between 73 and 75 before the launch of 30 series cards depending on the driver. And once 30 series launched, the driver target temp went up to 85, not 82. And the target temp will be hit on an air cooled card regardless of cooling, because if it isn't running at the target temp it will increase power to bring it up to that temp until it can't draw any more or reduce power to bring it down. The fan curve just affects how much power it takes to hit that target, and at default fan curve it will hit 85C ambient die temp at about 90% of it's target power, but if you roll back the target temp to 75 and bump the fans up a bit, it will run 110% target power without throttling. It will also run far more stable, with far fewer crashes from a 2-3C spike in ambient temps pushing it into the burnout danger zone. And as to the 12vhpwr plug, no card inexistence today takes 4 power cables. Even with the adapters. The 4090 is 3x8pin to 1x12 pin. It's taking 24 full contact pins and running them down to 12 half size minimal contact pins. It was destined to fail, it was never a solution to anything but an attempt to increase sales by self destructing 4080 and 4090 cards. You can safely run more power through 2 8 pin PCIe connectors than you can a single 12vpwr plug. Consider the fact that AMD NEVER even considered using it from the start, nor did intel. Transient spikes will still cause the cards using the 12vhp connector to fail, and either go blank screen or restart the system, and it's only an issue on NVidia 30 and 40 series, regardless of TDP. EVGA stopped making graphics cards because they knew the fact that the 12vhpwr plug was dangerous and NVidia insisted they use it anyways. It's not about the plug being properly seated, it is the plug itself. And rather than fix the issue, they just find new ways to make it "acceptable" until that turns into a new way to burn up a $1k plus piece of hardware, in hopes tat their fanbois will keep buying the self destructing GPUS.
Hey Jay, if you do do that video comparing launch and current drivers, be sure to include games that were around when the card launched to, for a more "base line" type thingy
Home appliances market is filled to the brink of explosion with planned obsolescence products since the past 30 years or so. My mom had her own mother's oven, which lasted my grandma a good 40 years, and my mom another 15 years before she actually had to buy a new one because the components that did eventually break were simply not being produced anymore since at least 20 years or more. In other words, it couldn't be repaired simply because the product completely outlasted the market of components that were manufactured to _make them_ in the first place. THAT is a product 'built to last'. Same with old fridges, old radios, old toasters, even old televisions. There's plenty of examples of products made during the 60s and 70s that always worked very well for a "life time" (that usually means about 40 to 50 years or service before something finally does break, erode, or malfunctions at some point). Nowadays you buy a TV and you're lucky it doesn't fuck up within the next 3 to 5 years. Same with most microwaves or many other home appliances and kitchen products.
I think planned obsolescence started in light bulbs there is a light bulb in a Oregon fire department that has been on 24/7 since the early 1900s. The different companies that made them started an agreement that light bulbs would only work for so many hrs. Really took off in the 50s with appliances.
the revisions to the 551 drivers were made to adjust jitter in DLSS. 551.21 had terrible jitter while using DLSS, 551.52 was a hotfix to stop that but increased hardware usage.
I cant verify this but I've heard that there was a driver update when 40 series launched that massively improved Ampere cards, completely debunking this idea. Richard Leadbetter from Digital Foundry actually claimed that with these drivers there were 3060s beating 2080s in certain scenarios.
As much as I have zero doubt that some companies intentionally gimp their old products and design things to break as they age, NVIDIA isn't a company I would ever accuse of that. They make their things obsolete by making better hardware, and any driver issues can safely be assumed to just be mistakes.
It also stands to reason that they are going to be more concerned with the newest card functionality instead of a 10** series functionality so if they have to add something to the driver they are going to do it. That is why you can download any driver along the way though.
Oh hey, I've always wondered if older hardware can see an uplift in performance from a much newer driver. I'd love to see a launch vs 4 years later video.
My Asus Zenfone 8 started acting up the day after the warranty expired. Yesterday, a day after updating the system for "improved stability", the phone froze and died. To be perfectly honest, I expected an 800€ phone to last me for around 5 years, not barely more than two... Now I'm back to my old S7 Edge and looking forward to the launch of the Pixel 8a!
@@crashniels man my phone doesn't even turn on like the screen is black and even the notification led doesn't turn on. The weird thing is that the back area around the camera and cpu gets slightly warmer than the room temp after time so I guess it's somewhat alive? And it charges at 5.3V 0.4A but after a while it goes to 0.2A. I measured the power consumption and the phone literally ate 2 whole battery charges in just 24h?!
Thank you for this video, J2C 👍 I feel like for CPUs, Intel did something akin to this once or twice. But I can't remember exactly. I think Wendell mentioned it before unless I'm mistaken. Kindest regards, friends and neighbours.
Planned obsolescence is government mandated energy efficient refrigerators. The compressor is so weak that they burn out within a few years. I guess the government and manufacturers didn't think about landfills filling up with refrigerators as part of their green energy narrative. 551.52 actually fixed driver issues that I had with my 2080. Frames are as good as they were before I had driver issues.
3080 LHR 10G user here. I don't know if this was a driver issue, but I started to have texture bugs, fx in War Thunder where my plane would be completely black but also occlusion bugs with rainbow colored boxes fixated to the sky. It happened right when I updated my driver on Geforce Experience. Hasn't happened since, but haven't changed driver again.
It could also be perhaps he didn't DDU the drivers between .23 and .51. I've adopted the use of DDU to update my graphics drivers. Speaking of which, I should do that now. lol
I have micro-stutters in Assassin's Creed Mirage and Valhalla on all drivers after 537.58 (October). The frametimes jump from 12ms to 60ms, sometimes over 100. I'm on GTX 1070, it might be old, but it works smooth on 537.58 or older. Both games are on DirectX 12. So to me, these new drivers do feel like Planned Obsolescence. Other info if anyone's wondering: i7-6700, 16GB RAM, Windows 10. Other games work fine, Shadow of the TR on DX 12 included.
Hanlon's Razor. As absolute performance increases, applications get less efficient. The world couldn't have run on Electron apps 20 or even 10 years ago. Makes sense it applies to graphics (and every other component in the way) as well.
@JayzTwoCents, I d/l'd .61 yesterday and noticed immediately that horizontal panning in X-Plane 12 caused stuttering. Strangely enough, there were very few (no more than 3 fps) frames dropped during panning. Overall performance seems to be better, except for the stuttering. I'm considering going back to .52 to see if it's worth the downgrade, but first, I'll try .61 with MSFS 2020 and see how well it performs. Great video, btw. 😊
Sometimes new drivers aren't about making things faster, but improving stability. So, if you lose a frame or two with a driver update, if the stability has been improved without massive performance loss, then it's still a win.
how cleanly were the drivers uninstalled and cleaned on the test videos? i've had strange performance results in the past when i didn't 100% kill the old drivers from the system
I have found sometimes when you update your drivers if you do not do ddu the new driver is slower with the mix of the drivers. So if you are having this issue run ddu and then reinstall the new driver.
I'm pretty sure this group of people are the one that extrapolate everything, and remember their GPUs dropping like 100 fps, and come to you 4 years later and they says: "I was making 150fps 4 years ago, that's definitely planned obsolescence ! I lost 50 fps !!!" This ppl just remember having a card better than it really was, and this effect is called "nostalgia", the same ppls that tell you games on the PS2 where significantly better (which is true in most cases) but their is definitely a lot of nostalgia in most cases. Sometimes you launch an old game and you find it's actually not how you remembered, I think it's the same thing for graphics cards !
So if their benchmarks were separated by months, rather than back to back, dust would be a reasonable answer. Planned obsolescence is absolutely a thing, just not necessarily HERE. Most of what drives planned obsolescence in computers is increasing program demands/intensity and operating system bloat, or, you know, Apple. I've seen it most in appliances. They make parts that fail and stop stocking the replacements or charge so much for the replacements that people just buy new appliances.
I wanna see launch drivers VS now. That would be an eye opener for people. Yes some drivers can cause issues and lost performance, how ever the next update could fix all those problems and give you better performance. It's the gamble with updates and a million different hardware combinations. Years ago, i had an issue with drivers where some DX10 games wouldn't load, turned out the update i ignored fixed that issue :D
Can confirm that a poor reinstall or cluttered install of drivers can cause crashes or fps drops. Happened to me once and was solved after just DDU-ing and reinstalling the same driver.
Those of us still rolling with a 1080TI are like the '96 Toyota Camry owners of the GPU space - there are better options out there but we still stick with what we have since it still runs like a champ since it just wont die.
i had this kind of problem a few times in the past that some nvidia drivers just dont like my pc. no matter how often i install the version it performes horrible but a few earlier or later work fine. sometimes drivers are just luck with nvidia. Edit: the planned obsolecence for nvidia is the small vram on many cards.
If anything, not supporting dlss 3 on 30 series cards is comparable. FSR mods show that frame generation is possible on those. It may be a technical issue specific to dlss 3 but apparently there's ways around it and the "old" cards are generally powerful enough to make it work.
The example of planned failure was with light bulbs because the manufacturers were making a fortune, but it stopped because the bulbs were too good and lasted too long. So they meet and planned for failure to keep sales steady
I can't say I have ever visibly seen drivers for a supported card force obsolescence in the way people might think. I say this as someone who long ago ran dual 250's until I upgraded by choice to dual 460's, then to a 780 Ti until a 1080 and finally a 4080 over a long stretch of years. I can only count a few instances where a driver update borked something so bad it wasn't funny, but it usually ended up being a bug and was fixed by the next update. If anything, drivers usually mature in such a way you will actually see improved performance in most cases due to both changes in how the driver utilizes the hardware, but due to updates to the games themselves.
Could this be due to new RTX features in the drivers? Like the RTX HDR and the RTX dynamic color thing ? Because when the diference is wider it shows more gpu usage on the lower fps drivers.
It always depends on the product and company but it's always safe to assume the worst because if they can get away with it, they'll do it and getting caught often costs them nothing. Generally speaking.
His CPU load (for the games) was lower with the new driver, for each game where there was a significant FPS drop. --> Maybe Windows was indexing the new files (unnecessarily) in the background and this way it took some performance from his CPU away. And without indexing, his CPU could spend more computing power for the game instead of for indexing files... But with a high performance CPU there was no difference. My idea: Install the drivers, let Windows do what it feels might be necessary (maybe it's indexing files...) - and after some hours of power-on time, the game would not be slower than before the driver upgrade. That's my guess.
i would like to clarify i still own a rx480 since 2016 and i have not lost performance over newer drivers and in some cases i have gained a few % in performance but that would just be margin of error.
I would do that video you're talking about but I would do more than two driver versions I would probably do the second or third driver update just so you can show a continued increase in usefulness because the first driver is not always great for any computer part.
Was dealing with a hell of a PC upgrade today. 7 hours of things going wrong and having PC hardware/accessories/everything just all over my tiny apartment. After all that, seeing Jay uploaded a video brings me joy.
Please do the launch Day vs four years later drivers. I’m interested and wana see the massive gap lol.
Yes, and do it with green, red and blue. Especially with Intel b/c it will be down right funny, some of those games have improved over 500%, we all know their day 1 was rough.
@@Dan-Simms only because blue drivers were so shit. Even early AMD ones weren't *that* bad.
@@Dan-Simms not just intel....amd also improves drivers over time which some pretty big jumps at times.
Use a drive a couple of versions in from the gpu launch, and compare with todays. These feels like Gameready drivers for new game launches, as it is 551.xx in both tests
I would say do day 1, year 1 and year 4
they don't do planned obsolescence with drivers, they don't need to. there are plenty of other ways the planned obsolescence kicks in, vram issues, longevity of the card itself, etc.
Or just collaborate with game developers to push the envelope on what's needed to run the game. I mean, CP2077... or Avatar. Avatar's got that "unobtainium mode" (or what it's called), which goes beyond ultra. When AMD and Nvidia release their next hardware, people are going to buy their most expensive GPUs just to have their game at the highest fidelity... and also bragging rights. ^^
"You got to buy the latest GPU from us because the following games just don't work as well with your older card" BS. Maybe if the game developers made GOOD code instead of slapping it all together and relying upon the GPU drivers and the bells and whistles that comes with the expensive cards to sort it all out, we wouldn't be having this discussion.
indeed. had a 570gtx for 7 years before the Vram was too little to open pretty much anything new. but the things it did run was smooth af. so plenty of power left.
Cancerous optimization on games is the main form of this...
@@silverwerewolf975 that's kind of a separate discussion because that's up to game devs, not the gpu manufacturers. Then there's going full tin foil and wondering if nvidia is supporting or even paying for poor optimization on older gpu's.
For some reason. VRAM usage is always higher on right one. Thats suspect in itself. Textures turned up?
You can see the lighting in CP2077 is also different. Looks like this person intentionally turned up settings to create a fake frame drop for clicks and attention.
@@zodwraith5745 I really hope it was a mistake but my intuition is the same
its always what these channels do to generate negative attention / clicks. wouldnt surprise me if they intentionally lowered the core clock & vram clock speed on the gpu too. or forget to restart the machine after driver update / re-apply the overclock for the gpu or just let a chrome browser open on a 2nd screen & u will get the 10-15 fps difference
i never had any driver that lowered performance except in 2 cases - 1st did lower my fps by around maybe 5% but got rid of the horrible microstutters and 2nd case was actually a driver issue they fixed after ~2 weeks that lowered performance in 2 titles i played. every driver since the gtx 9800 times gave me more performance, stability or nothing in specific titles i played but never willingly lower performance (cant speak about AMD though), some game updates however completely tanked it which got fixed mostly too.
I also thought that it was odd that Jay never mentioned VRAM when he mentioned every other stat.
@@squidwardo7074 One game, maybe. CP2077 is notorious for resetting your settings, but _every_ game? No way. That HAS to be intentional for either a narrative or agenda. I'm guessing that agenda is clickbait or just straight up Nvidia bashing.
What people often forget to do is double check the settings in the control panel. When downgrading the driver version, it also installs an older version of the control panel. If the version they had installed before added new settings or changed anything, it might cause that setting to change or go back to default. I even had it reset all changes made to the settings. That alone can produce either higher or lower performance, depending on what setting(s) changed. If you do not verify this, you will not know whether your results (old vs new) will even be comparable, regardless of game or test scenario.
This is a good point! I always do a clean install to zero out any issues like this or corruptible installation files from previous to newer. I know my settings and can always change them back again.
I always double check settings like some OCD crazed man with things like the control panel but even ingame. The amount of times that shit resets after an update or even the slightest file change is infuriating tbh
@mikev8746 Dude for real. Games that don't automatically revert back to fullscreen after alt tabbing are why I have trust issues.
@@DarkNia64 This is literally me 😂 It all started with my system and how often I had to re-check shit to be in its proper place. Now I fucking check everything twice in life. It ain't even fun anymore..
The lighting in the first video @5:17 for cyberpunk2077 is vastly different from the left to the right. That's not just a driver version change. There are other changes as well.
Good catch. I could see one title being an honest mistake if they didn't notice settings changing as CP2077 is notorious for this, but _every_ title? I smell some agenda tampering here.
Saw this too. I thought for a sec that it could've been atmospheric difference... Would be pretty extreme and make the benchmark totally inconsistent. Something is definitely up
Even in Forza you can see clear texture differences.
Dwindle FPS did a heap of testing on Cyberpunk and found Screen Space Reflections, Mirror Quality and Volumetrics affected the performance the most. If you look at that right side the ceiling light is shining through a thicker volumetric fog. That setting alone has around a 30% performance impact.
I bet he let Gforce experience in auto mode and it changed the settings!
You and multiple others have done enough testing to show that driver updates don't bork older cards, and i appreciate yall continuing to keep an eye on the topic.
You do see this a lot when people still have much older cards that nviida just dropped support for. So the nvidia 300 series stopped getting updates in 2016 so no new driver for them. The 500 series stopped getting updates in 2018 everything after the 500 series is still getting updates.
planned obsolescence is when you build a product that will fail or be incapable of working as intended like putting 8gb on a 70 class graphics card. this will cause you to buy a new one.
force obsolescnce is when you update a product, so it will run slower and causing you to consider buying a new one.
If it's forced you still need to plan how to do it. So still planned.
You're gonna need to qualify that a bit more. 8 gb on a 770 would've been massively overkill for no good reason. Same with a 670.
And you're also missing the fact that a card that can run a specific game at a specific performance level will almost certainly be able to run that same game at that same performance level for the entire life of the card, unless updates massively change the game. You might as well complain that your AMD K5 can't run Cyberpunk 2077. Just because technology advances doesn't mean it's planned obsolescence.
i mean is not like 8gb is not longer usable only in like 2 title on ultra is having issue at 1440p
@@BobBobson I'm pretty sure he was talking about the RTX 3070 and RTX 3070 Ti, not 770 or 670. The first 70 class card with 8GB VRAM was the GTX 1070 from 2016 and the last one was the 3070 Ti from 2021, finally in 2023 the RTX 4070 (Ti) got 12GB
Yet most systems running a 3070 will be designed around it and therefore couldn't really make use of the newer games on the settings more ram would allow anyway.
Vram requirements have been shooting up for many years - long before the 3070 came out.
If you buy a card thats only good enough for average performance on todays games why would you think its going to do that on games years later?
A certain big builder who shall remain nameless used to pair them to i3 8100 and 9100's depending on application. I wouldn't expect those cpu's to run CP2077 on high everything either.
the biggest thing that happened in the 551 driver branch is the fixing of v-sync micro stutter and it's great for people who were in the 537.58 driver for a long while.
how is there still micro stutters with single card setups?
@@kevinerbs2778 I didn't experience it too much in my setup compared the others. I did see in every update pre-551 drivers in a lot of reddit posts that the micro stuttering with v-sync + g-sync/free sync was causing people to have issues while playing games (think of a jumpy fps graph in msi afterburner while v-sync is enabled in the driver type of issue).
I guess it depends on the setup. A lot of the people complaining had 4090's, 4080's, and some 30 series too. My 5600x/3090 setup wasn't entirely affected since I run at 4k60 capped at 58 fps with vsync + g-sync, but I still did notice bigger stutters pre-551 drivers that were fixed on 551.52.
got any documentation on this? id be curious. i can' seem to find anything on the topic myself
Without going into a bunch of detail, I have an RTX 3070ti, with the newer drivers I have been able to finally get rid of the micro stutters and get better framerates when aligning my monitor settings and v-sync. I have MSI monitors with adaptive sync and have battled micro stuttering forever. I can now finally eliminate most of the stuttering.
@@drmorgan6903 good to hear! glad that adaptive sync and v-sync has finally worked out for you with the recent driver release.
The only industry I know that doesn't really have any very obvious planned obsolescence is the enthusiast and professional camera market.
Camera gear is expected to last for ages. Cameras like the good old Nikon D850 are built like a tank. You could knock someone out with it if they try to mug you and just keep shooting.
With modern mirrorless cameras they are even removing the last couple of moving pieces (the mirror box and even the shutter curtains in some models) so the chances for a failure are even lower than before.
Many brands support you with firmware updates for years and it's not rare to get your warranty extended to 5 years. I've been a photograpger for about 14 years now and I have yet to see a single piece of my gear fail on me (except for one very cheap SD card but I got that one from a crappy manufacturer).
When Canon and Nikon introduced their new mirrorless systems they also sold you adapters (separately or as a bundle with a camera) so you could just keep using your old lenses. They made it clear that they would love it if you buy their new lenses because they are way better (which they actually are) but nothing is stopping you from just using your old gear while swapping the camera body only.
Musical instruments are the only other thing I could think of
Never thought about that, but you're absolutely right.
Nope, Cameras have "planned obsolescence" in the form of batteries. Now, I don't believe most products have planned obsolescence so much as they have planned no longer supporting the hardware/software/device. And if there is a market for it almost everything can have a 3rd party aftermarket parts. Unless you are malicious and locked out at the software level. Which most companies don't do. Well, most companies except Apple, John Deere, and car manufacturers.
yeah and they sell them a fortune for that exact reason
Also wired audio gear tbh
Phoebus Cartel, established in 1925.
Major lightbulb manufacturers joined together in agreement to not produce lightbulbs that lasted beyond the agreed allotment. Planned obsolescence.
was gonna say light bulbs! you beat me to it! 😎
That is in fact, false
Look up technology connection's video on it
This is not the slam dunk you think it is. The reality is that while it looks bad. There are good reasons to do what they did. And since no one can prove why they did it we have no way of knowing. The main ones are efficiency and heat. The only way to make an incandescent bulb last longer is to make it run cooler and because the way an incandescent makes light is black body radiation you get dimmer bulbs on the orange-red end of the spectrum that generate a lot of heat but not a lot of light. So you would need more bulbs to provide the same amount of illumination but you would be generating way more heat while doing that and increasing the chances of in-wall electrical fires. It could have been planned obsolescence. or it could have been to make the light more attractive to customers by making it brighter than candles. or it could have been a thing to reduce the energy required since the grid wasn't as large as it is now. We will never know.
this example cannot be proven true, there are legitimate reasons to limit the lifespan of an incandescent lightbulb
@@MaheerKibria👈 this guy have fallen for the capitalist propaganda 😨
Thanks for fact checking! This would be an interesting series to keep up - testing continuous/older generation cards and keeping a tally of improvements with modern drivers.
I'd love to see a driver test with older cards such as a 1060 or 1070 for example. Would be out for 8 years this summer (for the 1070 anyway) so maybe do a launch/2016 driver and a driver from every 2nd year up to the current driver at the time so 2016->2018->2020->2022->current and see what the diffrences would be with the drivers.
I had some major glitching, artifacting and general graphic bugs on 551.52, to the extent that I spent a few hours troubleshooting with NVIDIA customer support and ended up downgrading to a much earlier driver. Just updated to 551.61 today and the issues seem to be gone. Whatever the issue was, it has definitely been resolved for me.
The problem with software is, that the combination of lets say your OS patchlevel, different divers and versions, software, etc. might cause strange behaviour. So "downgrading" might be the solution for a very individual problem that looks like "planned obsolescence" but is just a coincidence.
@@hw2508Oh I'm not saying it was "planned obsolescence" or anything of the sort. I presumed that the driver files I downloaded for 551.52 were either corrupted in some way or there was something else going on with my system. I was actually implying what you said, that each system can produce different results based on a wide variety of potential variables.
@@AxR558 yes, it is quite hard to prove planned obsolescence in ordinary products. But in a software environment, so much more difficult.
Yes, please do what you proposed at the end. But with older hardware. To me 30 series isn't actually that old. I have a 1070 still. That would probably be more of a target for planned obsolescence, if it exists.
The test that was done makes no sense in my mind. It's basically the same driver with fairly modern hardware.
That gpu is almost 10 years old now, at this point any performance drop would be from thermal paste or something like that being old from use. Would you count that as planned obsolescence?
@@MSusername69 You're just assuming they haven't changed that old paste? I do that with my older cards, hell I did it to my 3070ti not too long ago and it dropped 5c. Such a silly condecending response to someone asking for legitmate content, be on your way.
Newer drivers just don't support old hardware, but they do continue to provide access to the last drivers that supported the old hardware. At least the 10 series is still supported by the current drivers. My previous card to my old 1080 (the great 780Ti) is 474.82 [my current card is the 3080 10gb]
@@ihateeveryone8161 His argument is pointless anyway, because Jay would test different drivers on the same card and the card will not magically age inbetween tests, so paste or no paste makes no difference.
@@ihateeveryone8161 and the same to you. I was just pointing out something that could happen to older cards. I could also point out that some of the new software might not even work on the old cards, so the tests wouldn't even prove anything.
Unless you call companies stopping support for older hardware planned obsolescence. What about the silicon degrading overtime? Is that the same thing?
Congrats on the 4M subscribers!
Thanks. I love this stuff. I'm super interested in a launch driver comparison. I've done my own testing, but it was very inconsistent.
Of course Planned Obsolescence is a thing, Going to Electronics Engineering College 40 years ago, Mechanical engineering one of my first Engineering professors was GM Engineer in Automotive, he told he left after he designed a great Transaxle that would last a miilion miles at 99%, they told him to redesign it for 200k at 90% because they make $20k for the Car BUT 50k on Car Parts..
Planned obsolescence is a thing for decades.
Phoebus Cartel
Yeah, came here to say this. As a car enthusiast, I'm sure Jay is aware
Light bulbs are the definition of planned obsolescence.
It is legal in your country?
Agree. However, at least for me personally I’ve always upgraded video cards when I find them obsolete. Which is usually about every other generation in my case. Usually because the games I’m playing are more demanding/larger/more players and I’m not getting the performance I want.
Planned Obsolescence is ABSOLUTELY a thing in many industries. Cell Phones (and anything with a non consumer replaceable battery) are the perfect example
Thanks for this test and demonstration.
With fake benchmarks going rampant on youtube, can you vet this Gaming Uploads channel is legit, and not just a ad revenue farm?
There are so many of these channels that claim what ever they want to claim with zero proof of the actual hardware being used. It's all intentional bait to get views.
How smooth brained are you? The point of the video is to show the amount of variables in any give test setting, doesn't mean they're fake (they could be but you also can't say for sure which just makes it conjecture) more so that some people don't put in all the forethought to make test samples as constant as possible.
@@solarix I think you're giving people too much credit user error is a very common thing, while some people sure could be nefarious they are far more dumb people out there.
@@Collector_Of_Incel_Tears I polish it to mirror shine, thank you for noticing!
I do similar testing on my hardware. Also I have talks with Gaming Uploads, I am not sure about hardware but he does seem legit person.
Nvidia is def guilty of this. 8 Gigabytes on more than half of the 30 series and the 20 series, they also slowed the 1080Ti down
Love your videos been watching for years!
A friend of mine gave me his old computer, a Ryzen 9 3900X with a 2080Super, the computer had launch drivers, and I could swear I saw a huge performance upgrade when I updated to the latest drivers both on GPU and CPU.
See if you’re an AMD gamer like me, you would know that if a new driver has issues, it’s probably “AMD being AMD and launched a buggy driver again”
It’s easy to jump to conclusion with these things, remember CZcams slowing down with Adblock that turned out to be an Adblock issue? If Intel’s story is an indication, drivers are difficult things and buggy stuff happens. There’s also BIOS updates, OS updates, and even overall how good the hardware is maintained, there’s so many factors where it’s hard to find proof for this unless if it’s really blatant
Shh. We're trying to blame corporations for everything bad over here! Get outta here with your "facts" and your "logic".
Please Jay make that video with launch vs current drivers to see the performance difference. Much appreciated of all the work you do
If you're going to do the launch day driver comparison to recent one, I think including the specific games game ready driver would be interesting as well, although I don't think the difference would be significant regardless.
Game ready drivers are very nice, and sometimes might have huge performance uplift.
I doubt they're purposely _slowing_ down older cards, but they absolutely stop optimizing for them long before EOL. It's ridiculous that AMD EOL'd Vega when they literally still sell laptops with5000 series Vega APUs in them *_TODAY._* That should be a class action level offense.
Obsolescence is certainly something some companies try to do, take phone batteries for example, they made it much harder to replace your battery, the first thing that starts making your phone much more annoying to use after a couple of years. They could keep the same great looking backing, but have multiple tiny holes, like the one you use to get the sim card out, around the device, when you push them all in the back comes off. My phone I have now I considered getting a new one because the battery isn't as good as it used to be, but then I ended up buying a new battery instead... The battery is still sitting here because it's a huge hassle to change it.
It's time Jay take a relaxing North Ridge Fixs shop tour trip.😎😎👍👍
I remember when i had the 980ti classified when the 20 series came out and omg idk if it was just me but every other update would cause major issues and i would have to roll back sometimes 3 updates for it to work
Every company nowadays has some planned obsolescence and if you cant notice that you'd be crazy
Refrigerators, washing machines, motor vehicles, you name it.
Not so much planned obsolescence and end of life with no future support. Eventually, the time to maintain the code base or manufacture parts is not worth the money it takes to make it. And if there is money to be made by supporting these older products you will get aftermarket solutions. My d5000 no longer has OEM batteries. But I can still get batteries for it from after-market manufacturers. same with my car and most things. Almost everything today can still be fixed if you have the know-how unless it's made by Apple or John Deere and maybe some car manufacturers. Apple really likes to use serial numbers for everything to keep you from fixing your own stuff.
Billion percent. People are brain dead if they can't understand this. It started with light bulbs in the 50's. They've been pulling this shit for half a century.
No non have, it just that a product is made for X years of life
LOL! The Over, and Over, and Over, and Over moment, LMFAO!
Planned obsolescence does exist in the GPU market, it's just in the long game. When they give people small bus width's on cards knowing that not everyone has made the 1440p or 4K jump yet, or when they offer VRAM capacity that can be an issue later down the line. Or when they purposefully nerf a platform for future releases. Things like that constitute planned obsolescence in my opinion. Driver/software side I heavily disagree that AMD or Nvidia have any other interest but that of improving performance or keeping it up to par at worst. Considering their are a lot of times they without improvements on the driver side because of new generation GPU releases that they want to have exclusive features on, etc. What makes it planned obsolescence and not just evolving product lines, is that there is intent in nerfing specs for that very purpose. Intent matters.
Hey Jay! Thank for another, as always, great video!
I just also want to ask how you´re doing.
Thank you again for all your content and videos.
BR //M
Hi Jay. I still remember what you said years ago about installing new drivers. To Delete the old driver before you install the new one.
I'm not ready for an admiral TwoCents
yes please do a Launch Drivers vs. 4 Years Later Drivers would make a badass video
Good video...I would love to see launch day of the 20/30/40 series vs today as well...
Thanks for this video sounds really interesting. And thanks for the info about how different drivers work with some games in different way,s yep. 🤔😉😀
the WOW ad always makes me laugh i swear jay is the "i like turtles" kid grown up
Got my retro mat(x2) today, much bigger than i thought but well used now for my gaming room
Definitely want to see the video on launch drivers vs todays drivers!
14:15
Yes please. Would be great to see it for all three brands.
Planned obsolescence is a thing...but not with video card drivers, usually the opposite in fact, giving us much better performance over time.
I'm old enough to remember when Nvidia had issues with older generation cards getting worse and worse over time. I'm also old enough to remember when they sold a product advertised as having 4GB of VRAM and oopsy they shipped it with 3.5GB.
@@prima164 The iPhone 4 last iOS version was 7.1.2 in 2014. Another reason it wouldnt stay connected to a cell network is because 3G was being shut down, which is what the 4 primarily used. Also, phones tend to slow down with time is just lack of optimization from app developers/the manufacturer. Apple, or any other company that does software development is going to spend tons of man hours to resolve issues for a device/product that an ultra small population of their users end up using. Does it suck? Yeah, totally. But for any business, they can't support devices in perpetuity. There comes a point where you have to rip off the bandaid, even if that means upsetting some users.
@@RAZGR1Z Old enough, come on, it's not like it was in the 80s xD I had that card and it was only 10 years ago. It was GTX 970. They didn't ship it with half GB less. One memory chip was defective and was operating at only 10% speed so later on when the bug was discovered they had to turn that module off via drivers hence the 3.5 GB total memory. Never had a problem with that card and memory was never the issue btw. In fact it lasted me 7 years, swapped to 3060ti like 3 years ago.
Only had to roll back a driver once and that was when my frame time tanked in Horizon 5 on my 4090 and tripled. Rolled back and all was good, since then I've kept the driver up to date and using DDU each time and all has been good. Lately seen an FPS boost in Horizon 5 at 4K.
How about doing 1 -4 year drivers so you can see the progression.
this is a clear test that these driver update did not brick a recent (last gen) graphics card. but as other people are saying, what about an actual older card? like a 1060 or 1070 with current drivers. also what about last year drivers or something like that. i understand it would take a lot of effort to test that but it would make it for an actually comparable planned obsolescence test
The channel thegamrone has a video on their channel testing a 1080 Ti in November 2022 comparing its 378.78 release driver to the then current 526.98 driver across a variety of games, the current driver actually showed in an increase in performance in some titles while others performed basically the same. This lines up with my personal experience with my 1080 Ti I've had for over 6 years now where performance feels very similar, if not a bit better, with current drivers (well technically my drivers are 6 months old being version 537.13, I only like to update when I feel I need to, but still relatively recent) compared to the old ones.
I have also found that as time goes on the performance of old NV GPUs improves. I ran a 280GTX system right up until I replaced it with a 980GTX and did not come across any performance drops in older titles that came out around the time of the card. The same when I had my 980GTX running until I managed to get my 3080Ti.
From my own experience I have to agree with you.
I've been using a MSI 2070 since 2020 and I have kept up on the updates and I have seen no noticeable drop in performance since new. I also take my system apart approx. twice a year and clean it out. I do a fresh install on my system at least once a year also. There is still no noticeable difference from when I first built this system and the way it runs now.
As a mechanical engineer. I've quit working for companies before as i was explicitly told to design a mechanism to fail after 2 years when it was legitimately cheaper to produce a mechanism that woild last 5+ years.
In most companies "planned obsolescence" is a matter of cost optimizing a product to last for x timeframe. But unfortunately many companies do design products to fail after a certain timeframe specifically to increase sales.
Software companies doing the same thing would not surprise me at all.
"Should i buy a new GPU? Or just roll back to a faster driver?"
It's never been a thing in computer hardware, the tech moves so fast that time does the trick just fine. Tho we're most likely hitting the top of the S-curve when it comes to hardware requirements for gaming, so it might be a good time to start looking for planned obsolescence as it may be introduced into the equation sooner than later.
Oh, but it is a thing. For example, when NVidia released 30 series, driver target temps went from 73-75C up to 85C, more than enough to push well cared for 10 series cards junction temps into vicinity of the danger zone of 95C, Enough to literally burn out the die if it doesn't just trigger thermal cutoff every time you run a full screen application or intensive task. And 40 series comes with it's own short fuse planned obsolescence device called the 12vhpwr plug, trying to run MORE power over LOWER transference contacts, a plug that is literally trying to push 3 8-pin connectors worth of power through less contact area with higher ohms than a 4 pin Molex connector.
@@HFRG-zq1qm Counter argument: Was that the driver, or was that just GPU's thermal conductors growing of age (like if you don't repaste a CPU it burns up), newer games that were tried; newer background operations. There are tons of factors. Annnd what does 40 series running more power have to do with anything, in particular about planned obsolesence?
@@thewoodenlego5379 As to your counter argument, the driver literally changed the "Optimal target temp" from 75 to 85, and any older card that has not been repasted is very likely to burn out from the increasing ambient to junction delta, being as the minimum delta is around 7-8C (out of box or freshly repasted), and with that, 85 puts even a freshly repasted GPU dangerously close to thermal cutoff (TC is 95, 85+7 is 92, a slight spike or trough in power can trigger as much as a 5 degree C difference either way, , and 5 degree increase at 3 degrees shy of death is past the point of death). And as to 40 series, the built in obsolescence fuse IS the higher power draw, done over a connector that isn't able to handle it, a built in self destruct mechanism. A 4080 draws the same power as a 3090, but the 3090 doews so without burning up over it's 3x PCIe 8pin connecters, the 4080 does it over half the total contacts at half the size each, meaning it has 2 times the natural resistance to transference and thus half the capacity, and that's before you even figure in the reduced contact surface area, and energy which has it's transference impeded puts off resisted energy as heat.
@q1qm No clue wtf you're talking about mate, 10 series have always had their temp targets at 82 degrees, and that target is just where the card throttles itself, it means absolutely nothing unless your cooler is hot garbage because it's the fan curve that determines whether or not your card reaches that temp on a cooler suited for the wattage of the card, and that curve is set by the manufacturer by default. It's also customizable in any OC software like afterburner or nvidiainspector, where you can clearly see that it's 82 degrees target by default. Dont make shit up to argue strangers on the interwebs mate, that's fkin pathetic.
12vh is a failed solution to a very real problem: a GPU shouldn't be needing 4 cables from the PSU.
@@rekt4guud134 82C has always been the point of thermal throttling, yes, but the target temp varied between 73 and 75 before the launch of 30 series cards depending on the driver. And once 30 series launched, the driver target temp went up to 85, not 82. And the target temp will be hit on an air cooled card regardless of cooling, because if it isn't running at the target temp it will increase power to bring it up to that temp until it can't draw any more or reduce power to bring it down. The fan curve just affects how much power it takes to hit that target, and at default fan curve it will hit 85C ambient die temp at about 90% of it's target power, but if you roll back the target temp to 75 and bump the fans up a bit, it will run 110% target power without throttling. It will also run far more stable, with far fewer crashes from a 2-3C spike in ambient temps pushing it into the burnout danger zone.
And as to the 12vhpwr plug, no card inexistence today takes 4 power cables. Even with the adapters. The 4090 is 3x8pin to 1x12 pin. It's taking 24 full contact pins and running them down to 12 half size minimal contact pins. It was destined to fail, it was never a solution to anything but an attempt to increase sales by self destructing 4080 and 4090 cards. You can safely run more power through 2 8 pin PCIe connectors than you can a single 12vpwr plug. Consider the fact that AMD NEVER even considered using it from the start, nor did intel. Transient spikes will still cause the cards using the 12vhp connector to fail, and either go blank screen or restart the system, and it's only an issue on NVidia 30 and 40 series, regardless of TDP. EVGA stopped making graphics cards because they knew the fact that the 12vhpwr plug was dangerous and NVidia insisted they use it anyways. It's not about the plug being properly seated, it is the plug itself. And rather than fix the issue, they just find new ways to make it "acceptable" until that turns into a new way to burn up a $1k plus piece of hardware, in hopes tat their fanbois will keep buying the self destructing GPUS.
Hey Jay, if you do do that video comparing launch and current drivers, be sure to include games that were around when the card launched to, for a more "base line" type thingy
Home appliances market is filled to the brink of explosion with planned obsolescence products since the past 30 years or so. My mom had her own mother's oven, which lasted my grandma a good 40 years, and my mom another 15 years before she actually had to buy a new one because the components that did eventually break were simply not being produced anymore since at least 20 years or more. In other words, it couldn't be repaired simply because the product completely outlasted the market of components that were manufactured to _make them_ in the first place. THAT is a product 'built to last'. Same with old fridges, old radios, old toasters, even old televisions. There's plenty of examples of products made during the 60s and 70s that always worked very well for a "life time" (that usually means about 40 to 50 years or service before something finally does break, erode, or malfunctions at some point).
Nowadays you buy a TV and you're lucky it doesn't fuck up within the next 3 to 5 years. Same with most microwaves or many other home appliances and kitchen products.
Would love to see launch drivers compared to current ones.
13:38 Literally, every tech 'maniac' on CZcams dived into that hardware (plus drivers) and shared their opinions. We've got your point, Jay.😆
I think planned obsolescence started in light bulbs there is a light bulb in a Oregon fire department that has been on 24/7 since the early 1900s. The different companies that made them started an agreement that light bulbs would only work for so many hrs. Really took off in the 50s with appliances.
He ran his tests at 1080p and you ran yours at 1440p
the revisions to the 551 drivers were made to adjust jitter in DLSS. 551.21 had terrible jitter while using DLSS, 551.52 was a hotfix to stop that but increased hardware usage.
I cant verify this but I've heard that there was a driver update when 40 series launched that massively improved Ampere cards, completely debunking this idea. Richard Leadbetter from Digital Foundry actually claimed that with these drivers there were 3060s beating 2080s in certain scenarios.
As much as I have zero doubt that some companies intentionally gimp their old products and design things to break as they age, NVIDIA isn't a company I would ever accuse of that. They make their things obsolete by making better hardware, and any driver issues can safely be assumed to just be mistakes.
Apple was sued for this on iPhone and lost. That's why this doesn't really happen anymore. It definitely was a thing especially on phones for years
It also stands to reason that they are going to be more concerned with the newest card functionality instead of a 10** series functionality so if they have to add something to the driver they are going to do it. That is why you can download any driver along the way though.
Oh hey, I've always wondered if older hardware can see an uplift in performance from a much newer driver. I'd love to see a launch vs 4 years later video.
Id have to say i think they do, my 2070 is playing enshrouded on ultra settings 1080p and not falling below 58fps
My Asus Zenfone 8 started acting up the day after the warranty expired. Yesterday, a day after updating the system for "improved stability", the phone froze and died. To be perfectly honest, I expected an 800€ phone to last me for around 5 years, not barely more than two... Now I'm back to my old S7 Edge and looking forward to the launch of the Pixel 8a!
they tried blocking the bootloader unlocking method but managed to fuck over users since the update was broken
@@crashniels man my phone doesn't even turn on like the screen is black and even the notification led doesn't turn on. The weird thing is that the back area around the camera and cpu gets slightly warmer than the room temp after time so I guess it's somewhat alive? And it charges at 5.3V 0.4A but after a while it goes to 0.2A. I measured the power consumption and the phone literally ate 2 whole battery charges in just 24h?!
@@anzekranjc3312 does it show up as some qcom device when you plug it into your pc ?
@@crashniels dead. It doesn't show up at all
@@anzekranjc3312 is your phone made in April? If yes probably your flash storage died
Thank you for this video, J2C 👍
I feel like for CPUs, Intel did something akin to this once or twice. But I can't remember exactly. I think Wendell mentioned it before unless I'm mistaken.
Kindest regards, friends and neighbours.
When have you ever downloaded and installed CPU drivers?
@@zodwraith5745 Thank you 😂
@@zodwraith5745 When you ask?
Every time your mom bakes cherry pie;
Or never . . . whereas microcode patches are a thing 😎
Kindest regards, brother.
Planned obsolescence is government mandated energy efficient refrigerators. The compressor is so weak that they burn out within a few years. I guess the government and manufacturers didn't think about landfills filling up with refrigerators as part of their green energy narrative.
551.52 actually fixed driver issues that I had with my 2080. Frames are as good as they were before I had driver issues.
3080 LHR 10G user here. I don't know if this was a driver issue, but I started to have texture bugs, fx in War Thunder where my plane would be completely black but also occlusion bugs with rainbow colored boxes fixated to the sky. It happened right when I updated my driver on Geforce Experience. Hasn't happened since, but haven't changed driver again.
It could also be perhaps he didn't DDU the drivers between .23 and .51. I've adopted the use of DDU to update my graphics drivers. Speaking of which, I should do that now. lol
I have micro-stutters in Assassin's Creed Mirage and Valhalla on all drivers after 537.58 (October). The frametimes jump from 12ms to 60ms, sometimes over 100. I'm on GTX 1070, it might be old, but it works smooth on 537.58 or older. Both games are on DirectX 12. So to me, these new drivers do feel like Planned Obsolescence.
Other info if anyone's wondering: i7-6700, 16GB RAM, Windows 10. Other games work fine, Shadow of the TR on DX 12 included.
Hanlon's Razor. As absolute performance increases, applications get less efficient. The world couldn't have run on Electron apps 20 or even 10 years ago. Makes sense it applies to graphics (and every other component in the way) as well.
@JayzTwoCents, I d/l'd .61 yesterday and noticed immediately that horizontal panning in X-Plane 12 caused stuttering. Strangely enough, there were very few (no more than 3 fps) frames dropped during panning. Overall performance seems to be better, except for the stuttering. I'm considering going back to .52 to see if it's worth the downgrade, but first, I'll try .61 with MSFS 2020 and see how well it performs. Great video, btw. 😊
Sometimes new drivers aren't about making things faster, but improving stability. So, if you lose a frame or two with a driver update, if the stability has been improved without massive performance loss, then it's still a win.
how cleanly were the drivers uninstalled and cleaned on the test videos? i've had strange performance results in the past when i didn't 100% kill the old drivers from the system
Please do the video "launch driver vs 4 years later driver" 😊
I have found sometimes when you update your drivers if you do not do ddu the new driver is slower with the mix of the drivers. So if you are having this issue run ddu and then reinstall the new driver.
Please do a video comparing launch drivers to present day drivers!
I'm pretty sure this group of people are the one that extrapolate everything, and remember their GPUs dropping like 100 fps, and come to you 4 years later and they says:
"I was making 150fps 4 years ago, that's definitely planned obsolescence ! I lost 50 fps !!!"
This ppl just remember having a card better than it really was, and this effect is called "nostalgia", the same ppls that tell you games on the PS2 where significantly better (which is true in most cases) but their is definitely a lot of nostalgia in most cases. Sometimes you launch an old game and you find it's actually not how you remembered, I think it's the same thing for graphics cards !
I had screen flickering while gaming with 551.52 and 551.61, Had to rollback to an earlier driver to fix the issue
My very first thought at about 4 minutes seeing those side by side is "Someone needs to blow the dust out of the video card."
So if their benchmarks were separated by months, rather than back to back, dust would be a reasonable answer.
Planned obsolescence is absolutely a thing, just not necessarily HERE. Most of what drives planned obsolescence in computers is increasing program demands/intensity and operating system bloat, or, you know, Apple.
I've seen it most in appliances. They make parts that fail and stop stocking the replacements or charge so much for the replacements that people just buy new appliances.
I wanna see launch drivers VS now. That would be an eye opener for people. Yes some drivers can cause issues and lost performance, how ever the next update could fix all those problems and give you better performance. It's the gamble with updates and a million different hardware combinations. Years ago, i had an issue with drivers where some DX10 games wouldn't load, turned out the update i ignored fixed that issue :D
maybe 1080p is where they referred Jay
The guy in the ad really looks like Jay this is crazy 😂
Can confirm that a poor reinstall or cluttered install of drivers can cause crashes or fps drops. Happened to me once and was solved after just DDU-ing and reinstalling the same driver.
Those of us still rolling with a 1080TI are like the '96 Toyota Camry owners of the GPU space - there are better options out there but we still stick with what we have since it still runs like a champ since it just wont die.
i had this kind of problem a few times in the past that some nvidia drivers just dont like my pc. no matter how often i install the version it performes horrible but a few earlier or later work fine. sometimes drivers are just luck with nvidia. Edit: the planned obsolecence for nvidia is the small vram on many cards.
If anything, not supporting dlss 3 on 30 series cards is comparable. FSR mods show that frame generation is possible on those. It may be a technical issue specific to dlss 3 but apparently there's ways around it and the "old" cards are generally powerful enough to make it work.
The example of planned failure was with light bulbs because the manufacturers were making a fortune, but it stopped because the bulbs were too good and lasted too long. So they meet and planned for failure to keep sales steady
I can't say I have ever visibly seen drivers for a supported card force obsolescence in the way people might think. I say this as someone who long ago ran dual 250's until I upgraded by choice to dual 460's, then to a 780 Ti until a 1080 and finally a 4080 over a long stretch of years. I can only count a few instances where a driver update borked something so bad it wasn't funny, but it usually ended up being a bug and was fixed by the next update.
If anything, drivers usually mature in such a way you will actually see improved performance in most cases due to both changes in how the driver utilizes the hardware, but due to updates to the games themselves.
Could this be due to new RTX features in the drivers? Like the RTX HDR and the RTX dynamic color thing ? Because when the diference is wider it shows more gpu usage on the lower fps drivers.
This looks more like the changes to Low Latency and Defaults slightly changing
It always depends on the product and company but it's always safe to assume the worst because if they can get away with it, they'll do it and getting caught often costs them nothing.
Generally speaking.
If you notice the graphic detail in the new drivers is higher especially visible in the first few shown.
His CPU load (for the games) was lower with the new driver, for each game where there was a significant FPS drop.
--> Maybe Windows was indexing the new files (unnecessarily) in the background and this way it took some performance from his CPU away. And without indexing, his CPU could spend more computing power for the game instead of for indexing files...
But with a high performance CPU there was no difference.
My idea:
Install the drivers, let Windows do what it feels might be necessary (maybe it's indexing files...) - and after some hours of power-on time, the game would not be slower than before the driver upgrade.
That's my guess.
i would like to clarify i still own a rx480 since 2016 and i have not lost performance over newer drivers and in some cases i have gained a few % in performance but that would just be margin of error.
Do what you said at the end with the driver check. That would solve that debate and be interesting.
The 30 series had a huge performance improvement just before the 40 series came out in v522.25
I would do that video you're talking about but I would do more than two driver versions I would probably do the second or third driver update just so you can show a continued increase in usefulness because the first driver is not always great for any computer part.
Was dealing with a hell of a PC upgrade today. 7 hours of things going wrong and having PC hardware/accessories/everything just all over my tiny apartment. After all that, seeing Jay uploaded a video brings me joy.