Nvidia conspiracy ?

https://www.reddit.com/r/darksouls3/comments/4efemc/if_you_are_using_a_500_series_nvidia_card/
http://steamcommunity.com/app/374320/discussions/0/365163686081153774/

This is insane. I am not expert in drivers but how come a 2013 driver nets a 15-20fps increase for some people over the latest drivers...

Would love to get some input from the pros here. If this is true, I might give up completely on Nvidia and go the AMD route just by principle.
This is completely disgusting . I know they're not the only ones (yes Apple, I am looking at you) but a 20fps increase is not a minor increase, it's frickin HUGE...

 

Right before every new Gpu model come on the market start Nvidia to crippling their drivers for previous graphics. They want that you buy the newest they release. They haven't competition in the mobile market. Almost the same happens with desktop cards also. Amd is the opposite. They increase the performance with their drivers. So it is.

 

You've nailed it. It is worse than you think, @nick81 and yes... it's criminal and I hope there is a class action. They deserved to be spanked hard for their dishonesty, manipulation, and digital genocide. It's easy to demonstrate the next generation GPU is massively more powerful when they artificially cripple yesterday's best with cancer drivers in a systematic and premeditated manner.

I'm not sure going to AMD as a matter of principle is any different than cutting off your nose to spite your face, since all they seem to be interested in making in an MXM is weak trash.

I have spoken to an NVIDIA employee about it. They claim they don't know what's wrong and cannot reproduce the problems I and other have experienced. Well... I might have been born at night, but it wasn't last night.

The last good driver release for Kepler was February of last year. Everything else since then is cancerous. Here we are a year later, and Kepler customers have excellent GPUs that work flawlessly with old drivers, but malfunction and are effectively castrated by newer cancer drivers.

All I can tell Maxwell owners is get ready... buckle up... grab some vaseline. Pascal is coming and that "to die for" 980M that some of you may still be paying for, with interest, is probably going to get crippled exactly the same way. In fact, it is already happening. Maxwell behaves very differently with 359.00 and older drivers in my experience. With newer drivers the performance decreases little-by-little. And, that's not to mention the black screens and instability problems.

Bottom line: They do not care about us. They are not customer-centric. They are dishonest crooks that are willing to kill their own products, that customers paid their hard earned money for, to enhance sales of their new crap. It doesn't get any more dishonest than this. Maybe the person(s) responsible for NVIDIA's business to be run this way can go to jail. That would be sweet, and well deserved.

I've documented their driver shenanigans with a few videos on my YouTube channel.

Released April 28, 2015
[parsehtml]<iframe width="853" height="480" src="https://www.youtube.com/embed/aPxSge4NMlQ?rel=0" frameborder="0" allowfullscreen></iframe>[/parsehtml] Released June 1, 2015
[parsehtml]<iframe width="853" height="480" src="https://www.youtube.com/embed/zkqOfN0AQ7k?rel=0" frameborder="0" allowfullscreen></iframe>[/parsehtml] Released July 10, 2015
[parsehtml]<iframe width="853" height="480" src="https://www.youtube.com/embed/mpd2ZzAGrfo?rel=0" frameborder="0" allowfullscreen></iframe>[/parsehtml]

 

You can't really compare AMD and Nvidia on the driver front. AMD hasn't come out with a truly new architecture since the original GCN while Nvidia is constantly reworking their architecture which is why GCN cards get more and more optimized and Nvidia falls short. Obviously they aren't going to be concerned with performance for older cards, that's how it is. I'm almost willing to bet that when Polaris and Vega come out the same thing will happen to GCN cards with AMD. The bottom line is that both brands have their faults and both brands have their merits. Unfortunately Nvidia has a stranglehold on the laptop market and that's unlikely to change any time soon.

 

This also happens in the Doom beta. Upgrading from 3 years old drivers net my friend a 20fps LOSS on the same settings (minimum). I told him to go back to 314.22 as that helps out Dark Souls 3.

I still honestly feel it's more that their drivers suck absolute dog peepee rather than that they're taking their time to gimp Fermi, but either way, their newest drivers are not good.

 

I agree... cannot compare AMD to NVIDIA, or vice versa. NVIDIA should be compared to Satan and AMD the village retard. Both have severe flaws, but I can at least feel a bit of sympathy for AMD because it has to be painful to be a victim of your own incompetence. NVIDIA is just dishonest, and I do believe it is calculated manipulation on their part. So, I cannot give them a pass on the basis of incompetence.

 

I feel they tried to gain a lot more control through their driver over the past year which in the end made them loose the same completely...

 

When a company dominates a market, they have very little incentive to do anything "properly." This is not limited to NVIDIA. Pharmaceutical companies do this all the time with medicine - alter the formula by a fraction, give it a new name, sell it as an entirely new product to make money having invested very little in its R&D essentially eliminating overhead and increasing net margins, giving them bigger bonuses year-end. It's a common business practice in a free market situation with limited government intervention. (No, I don't believe gov. intervention is good, just saying.) The only thing Amazon, NVIDIA, Walmart, etc. has to worry about is that they do not pass the 67% market capitalization mark to prevent them from legally becoming a monopoly, even though they look, feel, smell, and act as such already. If their market share is expected to increase beyond 67%, they re-brand their products and devote less resources to driver support, etc. to control consumer behavior and push them toward their "competitors." If you look at market share graphics over the past 10 years, you'll see NVIDIA goes up, up, up, and then dives down, and repeats. Coincidence? Not at all. It's profiteering at its finest and is an art in the eyes of the business world... It's the game.



NVIDIA letting AMD "beat" (or match) them this year with Polaris or whatever is actually an advantage long-term. Think about it... They can walk all over AMD in years to come.

 

Planned obsolescence works wonders in any market where the product isn't a strict necessity (like gaming capable computers).

In other news, latest NV Linux driver cooks my laptop again...

 

This is what I was thinking too. They quite obviously are breaking things over and over, yet still "adding things"... whether or not they work, or their previous tech works.

Though @J.Dre may well have a good point about planned obsolescence. I know full well that many other industries, especially US-based ones, do in fact do exactly whatever they want for profits, including non-new new products.

It's also the same thing I noticed since Maxwell took the same launch trend that Kepler did; midranged first top end later, with less-than-stellar (compared to their desktop parts) mobile variants of cards. I mean, people will sit there and tell me the 680M was a fine and amazing card, but the 780M improved on literally every point except heat output (naturally) and showed the 680M never needed to be so gimped (especially in its memory) in the first place. I had someone yell at me that as soon as engineers get something done it's launching before, claiming that it's impossible that they willingly delay their products for money-making schemes, but it just doesn't add up otherwise.

 

this reminds me of the Hyrdoxycut supplement, back in the days it had ephedra and was great. Then they were forced to remove it so it became less effective. Now every few months, they come up with new name with a fancy new packaging and give you the same useless crap just to sell more.

 

What makes NVIDIA evil is not this kind of thing. It's how they seemingly on-purpose cause harm to their existing products that is sinister and ought to be illegal. This is not something other businesses do. I understand the concept of holding back full performance so they can simply tweak the firmware to do something special on a "new" model a year later. Not my favorite approach, but it is what it is. Making older hardware that is not broken and still functions really well suddenly begin to malfunction by using cancer drivers is a really wicked and underhanded thing to do, and it needs to be dealt with in a harsh manner.

 

Is it possible for us to prove? Just curious, not doubting it or you. Is there anything we can see in drivers from 347.xx and before compared to the latest causing these issues? Would be interesting if we could find the "flaw" or whatever is causing such issues. The fact that NVIDIA hasn't addressed it after all this time is quite shocking to me... Shows they couldn't care less, or did so intentionally and are ignoring it.

Something happened. Somewhere something went wrong or someone really screwed up.

 

Install any r364 driver branch family drivers on a Fermi card.
Play Dark Souls 3 or Doom's beta.
Note performance.
Install 314.22 drivers from 3 years ago.
Note 20+fps added to performance.
Proven, hands down, irrevocably.

What I want to see is if someone tests 340.52 which I believe is the last pre-maxwell driver on the fermi cards and see if it works the same as 314.22.

 

Fermi was released in 2010. It is entirely unreasonable to expect driver support other than basic support for a product that has been out so long. It is definitely too early in my mind to kill Kepler too but with that said, the entire computing industry is moving towards a max 2-3 year support cycle. AMD has been just tweaking GCN this whole time so old cards just kept reaping the benefits as an unintended side effect. AMD doesn't have the resources to dump into older products. Nvidia does, however, that certainly wouldn't appease shareholders. The bottom line is that we are increasingly moving to the point where planned obsolescence should be expected and we can thank companies like Apple and Google for that. We all hate rebrands but look at it this way... as long as Nvidia is selling cards that use an older architecture, they have to keep supporting that architecture with drivers. Nvidia is going to be replacing all of their chips with Pascal over the next year... Guess where that leaves Maxwell?

 

Basic support only I completely understand, but you have to be doing something EXTREMELY WRONG, or on purpose, to manage to break older products' performance in NEW titles that supposedly require new driver support.

I'm fully fine with, and understanding of, keeping a current-gen-plus-last-gen product cycle of updating, as far as support and new benefits come; inclusive of new games' support. The problem is that while every game up until the new consoles' launch basically said "8800 GTS for min spec", they, in the course of two years, changed from 8000 series to Fermi for min spec to Kepler for min spec to Maxwell for min spec. I've seen 750Ti cards as min spec be touted before, and I expect it to continue. The problem is that min spec is better than a Xbox 1, which runs almost every single game at higher than "low" (and "medium" too) presets on PC... 30fps or not. Why was Fermi and Kepler skipped over in a matter of months? Besides, planned obsolescence is a stupid thing. Old GPUs generally don't have the grunt to do things on newer titles; unoptimization or not. Nobody today is going to seriously convince anyone that they have a high end PC with a 560Ti or even a 580. Either there isn't enough vRAM, isn't enough core number-crunching grunt, or not enough technology support (like their relatively low tessellation capability, or inability to perform async compute at all for DX12 titles) to remain truly relevant. But having drivers break performance for already aging cards is something else entirely. And I don't like the 2-3 year support deal either, even though card lifetimes are generally two years long, and thus by the fifth year, someone would have a decidedly obsolete product anyway.

I know AMD doesn't. People like to say that their products get better with time (and that's true) but it doesn't make sense if an older product gets better enough to be relevant after 4 years... shortly after that it's going to phase out. And their drivers already don't age very well.

If we go with this trend of discarding even the very last generation of cards, we're going to end up in an even darker place for hardware. As if Maxwell wasn't bad enough.

 

This is a very deep pocket dirty strategy, as obvious as it sounds but think about the engg part, to put it simply How much work does it take to remove the good working profile and add a borked new feature that's still amateur. They should optimize each and every feature then release it to public else make the feature sets modular that should fix If not interested then slam in a warning atleast and leave the option to user why just play the orwellian crap. The mafia deals / contracts / corporates are ruining the PC market and the consumer satisfaction factor as a whole...embracing the $$$ side with a overload of evil nasty side of human intellect put to work, wish it was the other way around..

Hope maxwell doesn't get gimped like this...very unfortunate.

EDIT : Only way to get this fixed is fight back, remember that pre 347.xx mobile OC block fiacso.

 

It's not a conspiracy if there is factual evidence. Ask fermi and kepler users (hint: They gimped the cards when Maxwell came out)

There is no evidence to support that nVidia will change this when Pascal comes out. Forced obsolescence.

 

To be frank, I wonder how much performance gain made by maxwell came from power switching tricks and chopping FP64 instead of architectural gains.

We all know how much power a GM204 actually consumes compares to how much its suppose to consume.

 

Almost none. Power switching tricks actually makes consumption WORSE on GM200 cards; I've heard users with 980Tis claim that their power draw rating on their UPSes have shown a marked decrease after using custom vBIOSes for constant voltage in the P0 states. And FP64 was dead with Kepler; only the Titan and Titan Black had FP64 calculations... and even then, you were required to enable it in NCP and it had a warning that gaming performance would decrease as long as it was on.

Maxwell's gaming benefits are basically all architecture.

 

I was more talking about FP64 costing die space on GK110 compared with GM200. The biggest turn off for me on Titan X was exactly that. Titan was meant be a bridging product between quadro and geforce. A no frill quadro for those who wants to game at times as well.

I am not sure how at a constant load, power switching can results in higher power consumption.

 

GK110's 780Ti/780 had no FP64, and didn't have any kind of benefits, is what I meant. If Titan and Titan X were the only GK110, that'd makes sense, but they weren't. Titan Black was by and large a better card than the 780Ti, even with the FP64 block. Unless the addition of the FP64 block on GK110 gimped potential cores (meaning GK110 could have had more cores in general), but that means that it'd have drawn well over its 250W TDP rating and been even hotter (and GK110 was hot.. though not as bad as Fermi or Tesla).

If power switching ends up with higher voltage more often than lower, it can. Remember how Adaptive Voltage on intel CPUs often go well above the voltage you set it to in some stress tests? The same thing could happen for the GPUs (just more rapidly to notice that the voltage is higher). Setting it to a constant makes sure it never deviates. When you have high power, high end cards running overclocked and/or full bore, the adaptive voltage can be a downside. It's not the first time I heard someone claim it happen either, so I don't think it's them pulling something out of their butts.

 

GK110 wasn't THAT hot... the annoying thing about GK110 was how unstable it got when it passed ~85C. That chip hates heat...

 

Maybe depending on your cooler and fan setup and room I guess. But it was a hot chip. People in my country that use maxwell over Kepler always remark how it's amazing how cool it is (80c or less) compared to Kepler cards (90c was common). On the other hand, they remember Fermi lived at nearly 100c here, so...

 

Nice documentary, right on topic.))

 

Forgive me for being noobish in this, but I have a question. I need to run the newer drivers for proper SLI support on Black Ops 3. Is it possible to export a game profile from the newer driver and import it into a older driver, and have it actually work? I may have done it wrong, but I tried doing that previously with nvidia inspector and it seemed like it didn't work.

If I could run 359.00 or earlier with the new profiles, that would be awesome.

 

Sure it works!
The one and only @j95 posts profiles regularly on Techinferno
https://www.techinferno.com/index.p...t-upgrades-driver-support-modded-inf/&page=47

As a new user you won't be able to download profiles and modded drivers but you can pay a minor fee to do so (which I did and wholeheartedly recommend considering how much the people there helped me with my 980m)

 

GK110 runs quite cool for the most part for me. I had an high ASIC 780 ti and it runs extremely cool. I can imagine for the first GK110 comes into consumer are horribly leaky monsters, but I imagine use it got to 780 TI, those chips are well binned.

I was more saying the unused FP64 acts as dark silicon and probably helped a bit with the heat. Where in GM200, its pretty much all FP32 on a massive die. Hence I dont think gaming GP100 running on FP32 with FP64 disabled would run much hotter than GM200 if the leakage is not absolutely horrible.

 

With proper pads and paste, I've never had any issues whatsoever with 780M SLI thermals in Alienware laptops. Any CPU or GPU(s) are going to run too hot if the those are not done correctly, or if the heat sinks are poorly designed or do not fit correctly.

 

_______________________________

To underscore your point here, this is the best example of why not having true competition and choice has hurt the consumer in America, and in particular, the PC gaming community.
With AMD not really being a true force to contend with now, it has only embolden Nvidia- and to a lesser extent Microshaft as well, to control and dictate the game, and it's not in our best interests.
We collectively spend billions each year buying the best products our quickly dwindling dollar can afford, but are quickly short-changed by the tactics of a company which pretty owns and controls the PC gaming business in America.
All of this is due largely because of lack of competition from AMD, which as we all know, would have kept Nvidia from taking advantage of us with releasing sub-standard drivers, and keeping us the consumer from experiencing the benefits of what we paid for.

 

Agree. The only thing that is not 100% correct is reference to "America" because the NVIDIA monopoly is a worldwide problem that affects any and all gamers the world over. Same can be said of the Redmond Mafia's monopoly.

 

I don't know about that... the dollar has been rising in comparison to ours as far as I know.

I don't blame lack of competition from AMD. I blame nVidia for being greedy, pathetic, unmentionables. Competition from AMD would indeed force their hand, but the fact that a hand needs forcing is the real problem we should focus on.

 

Right on... if they were truly good in every sense of the word they would do good without the force of competition. Unfortunately, many companies (and people) only rise to the level of their incompetence, and then only when they have to.

 

_________________________________

With respect to rising prices domestically, a stronger dollar hasn't really helped our economy. The cost of overall cost of product domestically has risen exponentially within the last 3 years, due to increased taxation, healthcare mandates, etc., which makes the USD weaker in terms of buying power- especially in the Middle Class, which more than likely comprises the biggest buying segment of the Nvidia-based products.
(Just check out the prices for a new car these days, and it becomes all too clear).
With Intel announcing big layoffs, along with shutdowns and closures of retail chains, we are seeing a much, much weaker US economy trending now, which really should prompt Nvidia- theoretically at least, to go the extra mile to please and appease us in the gaming community now more so than ever, one would hope.
Of course, as we all know, that ain't gonna happen, as they and Microshaft have had a stranglehold over the PC consumer market for many years now, and it's only getting worse- especially with Microshaft and Intel and post-Skylake CPUs.
Point being, Nvidia has little to no competition with anyone globally, and that puts an extra squeeze on all of us.
It's really a pity that we as consumers don't have a viable alternative or option from which to choose to teach Nvidia a lesson about screwing us all over.

 

To be truly honest with you, that's all preventable. Your cost has risen because people want to make more money. It's the same thing that happens here. If prices of a certain good raise, or taxes are raised for the richer, then they want to make more money, so everything's price raises, and then everybody needs more money that they're not getting, so everything else's price raises, and it's a neverending cycle of the rich trying to be richer because they're mostly the scum of the planet and their ridiculous luxury lifestyles must be maintained or improved.

There's so many mixed-market economies in the world and governments refuse to put price ceilings in place when they change taxation levels or certain laws where the rich people end up having to pay more. I understand that rich and poor *MUST* exist in the world for freedom to exist, but there needs to be a limit on ridiculousness. The US is NOT in any sort of weak economy... it's just cannibalizing itself and people believe the economy is weak, and then the self fulfilling prophecy begins. Taking measures to counter a coming problem causes the problem to exist in the first place.

As for BGAtel and nVidrosoft and Macro$haft, they're the same way. They want control, because control gets them money. Money is the end goal. Business without morality is one of the weightiest sins of the world, regardless of what religion you do or do not believe in. And this is how they roll. Hence the hatred of them. The only problem is, we need their products and they know it, so they can do what they want. *sigh* why don't people with some semblance of morality own businesses?