Is Nvidia Screwed?

[Marinatall_Ironside]

For a game released in 2007, you'd expect something like a 960 being able to run it. Its had years for the drivers to mature, and I'm pretty sure some patches to the game as well from Crytek themselves if we use GPU's today compared to even the likes of a 580. Was it optimized? Debatable. And yeah, still, whichever setting that was turned on or off, it still doesn't make sense a 960 beats a 780 or a 780 Ti. What a way to screw with customers and push them to pay for Maxwell. I'm pretty sure not everyone who owns a Kepler card is willing to do that.

The bugs and glitches pretty much count for lack of optimization. Optimization isn't and shouldn't be only synonymous to FPS and/or smoothness or stuttering. I mean, look at Warframe. It has stable FPS. But for goodness sakes, its still filled with bugs and glitches. If anything, optimization should also mean a game should be refined to the point that bugs and/or glitches like randomly falling off of a map and stuck into a falling animation, getting stuck in a corner as if you're trapped by an invisible cage, or even your field of view is drastically affected (I could list more, but that game is buggy and glitchy to this day... its just that there's much less of it today, but still significantly buggy) should happen every 1/100000000 times, not 1/100 times (in Warframe, sometimes a bug happens 1/100 or even everytime.).

The reason why a 960 would beat a 780/Ti is because resources have been moved away from Kepler and put into Maxwell. Kepler really didn't mature all that well.

 

Did you also know that GPU performance almost stagnated because there were no games before Crysis that were really pushing the GPUs hard enough to warrant an upgrade on performance. So taxing and demanding games still has a place (And that's why Crysis is inspirational to me on the graphics technology side) Crysis, in its time, was far ahead in graphics technology, which is why we have very powerful consumer GPUs today. If we compare Crysis to today's games, we will see just how well Crysis has held up in terms of graphics. So ultimately we have to give kudos to Crytek for pushing ATI/AMD and Nvidia to make more powerful GPUs.

 

So the way I look at it, you can still create a well-optimized game that is very demanding and can run smoothly (Not having FPS going all over the place). It's possible, just that developers aren't willing to put in the time to make that happen.

 

 

There won't be any major development in the GPU area until the next generation of consoles, the way I see it. PC games aren't pushing the boundaries like they used to, even Crysis 2 has very reasonable specs. If there aren't any mainstream ways to put new hardware to the test, AMD and Nvidia will keep on doing what they're doing until the consoles catch up with PC hardware.

Either this is still happening, or I can be wrong and it has happened before Crysis.

 

okay, here how it goes: 

- a tasking intensive game should run heavy on any system but you will notice a significant difference in FPS count between high end and low end GPUs, like you see here in Crysis benchmark.

- an ill optimized game will run bad at any system, and even if you used a high end GPU you will not see a noticeable increase in performance.  meaning poorly optimized games don't know how to utilize the power. you can see this in something like DayZ Standalone (it's an alpha and still not ready for optimization of course) or Batman Arkham Knight (the first release) where people tried the game on different PC builds and noticed almost no changes, the game basically ran like shit.

--From Mesh Rash - But Can It Run Crysis - Modern Hardware Edition--

This is how demanding games work.

[WynLore]

Replies in blue

The reason why a 960 would beat a 780/Ti is because resources have been moved away from Kepler and put into Maxwell. Kepler really didn't mature all that well.
Thus further proves my point on number 2. Kepler can mature pretty well, but it seems Nvidia has shifted a significant chunk of their allocated man power on the ones incharge of Kepler drivers to Maxwell.

 

Did you also know that GPU performance almost stagnated because there were no games before Crysis that were really pushing the GPUs hard enough to warrant an upgrade on performance. So taxing and demanding games still has a place (And that's why Crysis is inspirational to me on the graphics technology side) Crysis, in its time, was far ahead in graphics technology, which is why we have very powerful consumer GPUs today. If we compare Crysis to today's games, we will see just how well Crysis has held up in terms of graphics. So ultimately we have to give kudos to Crytek for pushing ATI/AMD and Nvidia to make more powerful GPUs.
Sure, good on Crytek on that one. But looking how things are going today, I think the industry is going backwards. By backwards, I mean even consoles are affected negatively as much as PC, if not slightly more, when developers aren't given enough time by their higher-ups to perform an actual alpha test (where the game/s they made are tested by hired game testers or they play the game themselves and see what they overlooked) or if they want to at all.

 

So the way I look at it, you can still create a well-optimized game that is very demanding and can run smoothly (Not having FPS going all over the place). It's possible, just that developers aren't willing to put in the time to make that happen.

Thus why number 1 is even more valid. I didn't mean to generalize all developers when I made that, but I did mean that even the likes of Bethesda, Ubisoft, the three developers working on CoD (Infinity Ward, or whats left of them, Sledgehammer, and Treyarch), and Dice have released their games with crap optimization (again, not only in frame rates, I mean things like a horse is in the air as if its standing on an invisible floor, swimming through solid concrete and can see everything on top of you, a never ending loop of falling down into something and you never saw it coming, an NPC thats stuck on a wall, some weird animation that was meant to look like you just stabbed a dude in the chest but in reality looks like you killed someone by giving them a handjob despite that its their chest thats bleeding).and we've had to wait atleast within a week that they'll patch parts of it or if they can patch fix the entirety of it. IMHO, just because you can patch a game, it doesn't mean you can and you should be able to get away with not giving it enough optimisation from the start before releasing it.

 

Either this is still happening, or I can be wrong and it has happened before Crysis.
Strangely enough, PC hardware has been a victim of being pushed in the wrong way due to lack of optimisation upon release, making patch fixes common and can or could take months. Whether or not it affects one's gaming experience, I'm pretty sure it doesn't hurt to delay a product prior to release to get it right. Plus, Crysis 2 pretty much uses an excessive amount of tesselation on things that really, really, really don't need it (like a rock or 5 square foot's worth of a piece of stone)

 

This is how demanding games work.
The closest we have to a demanding game right now is Crysis 3 (which I'd argue that Crysis 3 is the most well optimized Crysis game around... though I'm not sure if it still uses an unnecessarily excessive amount of tesselation on a rock or 5 square foot's worth of a piece of stone), followed by Witcher 3 despite things like this:

Wouldn't hurt if Witcher 3 was released in 2016.

[Marinatall_Ironside]

Replies in blue

With the lack of optimization, most of the time, we have to blame their publishers for setting release schedules that takes away the opportunity for a proper, actual alpha/beta test (meaning that they have one or two years to complete the game rather than three to five years). Blizzard Entertainment is literally one of only a small handful of companies in my opinion that truly cares about the quality of their games. They will only release their games "when they're done", which is why Blizzard Entertainment is known for not talking about release dates while they're midway through a development cycle. They will then set a release date when their game is nearing total completion (Performance optimization done, squashing out bugs/glitches, and finalizing game mechanics). Also they publish their own games.

 

I agree with the validity of your statements about Kepler and how their respective industries are heading backwards. However there is still a chance for those to move forward. Just give the developers more time to properly test their games like Blizzard and then we would see them move forward. It's not that hard at all if you look at the bottom line.

 

And I agree with The Witcher 3. Could have been released one or two years later to remove those glitches, then it would have been fine.

 

With Crysis 3, yes it's the closest thing we have to a well-optimized game that's demanding. However with the sporadic amounts of tessellation for rocks and stone, be glad they're not 100% texture maps (Like a log in Call of Duty: Ghosts). If you're trying to aim for realism and believability, games will end up demanding and with models with lots and lots of tessellation for organic shapes. If I was to make an eSports title, I would make that game really easy to run, whereas my AAA games are going to be very demanding. Both will be well-optimized for performance consistency and the removal of bugs and glitches.

 

If you haven't realized something about me, aside from being a PC enthusiast, I am actually majoring in video game design with an emphasis in 3D Animation and Modeling, Video Game Design, and Writing, with a side of voice acting on top of that. And I want to work with Blizzard Entertainment for the mentioned above. Being a perfectionist is not easy whatsoever, and I draw about half of my influence from filmmakers such as James Cameron, Peter Jackson, and George Lucas. The only kind of game influence I can draw from are from Chris Metzen, Crytek, Bohemia Interactive, John Romero, and Shigeru Mimayato. The games themselves I find to be the most influential to me are DOOM (everything about it), ARMA 3 (the sheer difficulty), Diablo II and TERA (The roleplaying mechanics of Diablo II and the combat mechanics of TERA for any of my ARPG games) StarCraft (the characters and stories), Crysis (graphics technology. My favorite game engine is CryEngine), and Tomb Raider (the character design of Lara Croft from the reboot series and some game mechanics).

[MMKing]

I will care about DX12 performance when:

 

1. More than a single test comes around. I don't look at a single test for DX11, nor will i look at a single test for DX12.

 

2. Polaris and Pascal comes around. I won't be upgrading sooner either way.

 

As far as i'm concerned, Nvidia still have time before DX12 becomes mainstream sometime late 16, early 2017.

[WynLore]

Replies in blue again since I'm likely to botch it less.

With the lack of optimization, most of the time, we have to blame their publishers for setting release schedules that takes away the opportunity for a proper, actual alpha/beta test (meaning that they have one or two years to complete the game rather than three to five years). Blizzard Entertainment is literally one of only a small handful of companies in my opinion that truly cares about the quality of their games. They will only release their games "when they're done", which is why Blizzard Entertainment is known for not talking about release dates while they're midway through a development cycle. They will then set a release date when their game is nearing total completion (Performance optimization done, squashing out bugs/glitches, and finalizing game mechanics). Also they publish their own games.

I personally believe that the necessary demons for video games to be published are, well, publishers. Sure, we have Kickstarter, GoFundMe, so on and so forth, but unless you're Tim Schaefer, the game you're "self-publishing" (for the lack of a better term) better atleast have a cult-following, even if its free from the get-go. I mean, a developer's gotta eat (and may also have a family to feed). I personally feel mixed towards game developers (a mix of love and sympathy, not hate... unless its Phil Fish. Dafuq does he think he is? Kojima? Kojima got more death threats on ending the Metal Gear franchise in one day than he got insults in his life, and has this complex that makes me sincerely wanna hit him. Sure, he made alot of other games prior to Fez, if I remember correctly, but just because you made a hit, it doesn't mean you can be a total asshole.) as I know some in person who love the craft and are dedicated and won't even leak their work unless they are sure as hell that its even in beta stage (tested pre-alpha and then a legit alpha-test, atleast thats how I understand when them), while there are cunts like Phil Fish. Regarding publishers? Atleast Blizzard is very reasonable.
 

I agree with the validity of your statements about Kepler and how their respective industries are heading backwards. However there is still a chance for those to move forward. Just give the developers more time to properly test their games like Blizzard and then we would see them move forward. It's not that hard at all if you look at the bottom line.

The questions would be:

-Are they allowed to? If they are, how much time do they have?
-Do they want to? Or can they? If they do want to, how much do they want to spend on it? If not want, then how much time can they allocate towards optimizing their game?

 

 

And I agree with The Witcher 3. Could have been released one or two years later to remove those glitches, then it would have been fine.

 

With Crysis 3, yes it's the closest thing we have to a well-optimized game that's demanding. However with the sporadic amounts of tessellation for rocks and stone, be glad they're not 100% texture maps (Like a log in Call of Duty: Ghosts). If you're trying to aim for realism and believability, games will end up demanding and with models with lots and lots of tessellation for organic shapes. If I was to make an eSports title, I would make that game really easy to run, whereas my AAA games are going to be very demanding. Both will be well-optimized for performance consistency and the removal of bugs and glitches.

Well, when I saw this video on Crysis 2 

I was really surprised on the amount of tessellation on things I'm not even sure is necessary. I mean, damn... really? Does it have to have that much tessellation? I don't really see why it needs that excessive amount of tessellation. I love the CryEngine, but come on. I'm pretty sure some of those details can get away with less polygons and still maintain the looks that it was intended to look like. I'm not sure if Crysis 3 shares the same.

And yeah, CoD: Ghosts is probably the most hated CoD around.
 

 

If you haven't realized something about me, aside from being a PC enthusiast, I am actually majoring in video game design with an emphasis in 3D Animation and Modeling, Video Game Design, and Writing, with a side of voice acting on top of that. And I want to work with Blizzard Entertainment for the mentioned above. Being a perfectionist is not easy whatsoever, and I draw about half of my influence from filmmakers such as James Cameron, Peter Jackson, and George Lucas. The only kind of game influence I can draw from are from Chris Metzen, Crytek, Bohemia Interactive, John Romero, and Shigeru Mimayato. The games themselves I find to be the most influential to me are DOOM (everything about it), ARMA 3 (the sheer difficulty), Diablo II and TERA (The roleplaying mechanics of Diablo II and the combat mechanics of TERA for any of my ARPG games) StarCraft (the characters and stories), Crysis (graphics technology. My favorite game engine is CryEngine), and Tomb Raider (the character design of Lara Croft from the reboot series and some game mechanics).

I must ask out of curiosity, though. Have you played Shadow of the Colossus? I mean, if you haven't, give it a try and it may help you with your concepts. If you did, what do you think of it? I mean, sure its an old game, but the story telling and the gameplay (despite that controls being somewhat finicky) are compelling.

[Marinatall_Ironside]

Replies in blue again since I'm likely to botch it less.

To answer some questions you posted:

 

-Are they allowed to? If they are, how much time do they have?

Depends on the publisher's set release date for a game. Most aren't that generous enough to give developers the time to properly test and optimize their games (Electronic Arts anyone?)

 

-Do they want to? Or can they? If they do want to, how much do they want to spend on it? If not want, then how much time can they allocate towards optimizing their game?

If they want to? Really depends on the developing team. Some teams want to test their games while other teams are like "Fuck testing, we need to get this game out for the holidays" (Remember E.T.: The Extraterrestrial?). For the amount they want to spend on it, if the budget did cover the testing phase, then yes they should. But if not, that's something to debate with the upper-ranked personnel of whatever company you happen to be working with. Some are reluctant while others are not so. And lastly, if the development cycle does afford them the time (which also depends on the scale of the game. Smaller ones tend to have better times in a 2-4 year development cycle), then they can allocate the necessary time to properly test and optimize the game. With my games, the development cycles are going to be anywhere from five to seven years to allow for the full proper development, testing, and optimizing of my games. StarCraft II: Wings of Liberty had a really long development cycle after Brood War was released. The former was released back in 2010.

 

 

 

I was really surprised on the amount of tessellation on things I'm not even sure is necessary. I mean, damn... really? Does it have to have that much tessellation? I don't really see why it needs that excessive amount of tessellation. I love the CryEngine, but come on. I'm pretty sure some of those details can get away with less polygons and still maintain the looks that it was intended to look like. I'm not sure if Crysis 3 shares the same.

The shapes are organic. You have to remember that. For something really geometric, then tessellation isn't necessary. The amount of tessellation is really dependent on the design choices, not the engine. And I intend to also prove that CryEngine is also just as great for mainstream titles.

 

 

 

I must ask out of curiosity, though. Have you played Shadow of the Colossus? I mean, if you haven't, give it a try and it may help you with your concepts. If you did, what do you think of it? I mean, sure its an old game, but the story telling and the gameplay (despite that controls being somewhat finicky) are compelling.

I have heard of Shadow of the Colossus, but I haven't played it yet. Is it on Steam? I may give it a run. Looks fun to play

 

Also another fun fact about me: I hate instant gratification in games because games with that element aren't that fun to play in my opinion. It doesn't really challenge the player(s) to overcome something difficult to reap the rewards. Instead those rewards are easily accessible. If you hate instant gratification as much as I do, I am sure you will enjoy my games, because they will be difficult as balls.

[Majestic]

80%+ market share, yeah I'm sure they're really worried.

[vanished]

80%+ market share, yeah I'm sure they're really worried.

It didn't get that way for nothing though - given time, and enough right and wrong decisions by either company and this number could sway drastically just as it has done before

[Majestic]

It didn't get that way for nothing though - given time, and enough right and wrong decisions by either company and this number could sway drastically just as it has done before

 

AMD Athlon to now is a good example of that, indeed.

[Curufinwe_wins]

Just reminding everyone that after revisting some of the early indications of dx12 in AoS gm200 was well on par with fiji, and while Hawaii gets some beastly improvements it's clear gm204 isn't quite there from a hardware perspective.

[Andrew Sh]

http://www.youtube.com/watch?v=zsRlKm5Oq_8

May be an oldie.  But I'm rooting for team red.  Yes, I've been a green fan for the longest time.

No one fully supports DX12 yet so don't think Nvidia care much.

They've been dominating the market for quite some time.

Fortunately for them, reputation is more important to most gamers than what is currently on the benchmarks so people that have always went with Nvidia will stick to them even if they lose a bit of performance out of it.

 

Both companies are great but I don't see either side dieing off anytime soon

[lior1111]

Everytime i see those posts im suddenly feeling relieved about my new purchase of the r9 390 but then i read the comments and im getting inside a black hole that tells me how much nvidia is superior and i feel like i shouldve gone with them.. Not true though imo

[fred_mertz]

80%+ market share, yeah I'm sure they're really worried.

If anything, Nvidia fanboys should worry about a monopoly from team green.

Should that ever happen, you'll be paying $500 for a GTX 750 ti and dealing with nerfed performance of your gpus.

Oh wait, that's already happening.

[Majestic]

If anything, Nvidia fanboys should worry about a monopoly from team green.

Should that ever happen, you'll be paying $500 for a GTX 750 ti and dealing with nerfed performance of your gpus.

Oh wait, that's already happening.

 

You're projecting. 

 

Also, massive slippery slope argument. 

[vanished]

If anything, Nvidia fanboys should worry about a monopoly from team green.

Should that ever happen, you'll be paying $500 for a GTX 750 ti and dealing with nerfed performance of your gpus.

Oh wait, that's already happening.

It's already happening yeah but according to certain people here (this exact argument has come up for different products elsewhere at least twice before) it would not simply balloon out of control.  There's only so much total demand, and that is spread across the entire supply, so if nvidia has 80% market share now, prices could only go up another 20% (if AMD disappeared) before sales would start to drop off.

[App4that]

It's already happening yeah but according to certain people here (this exact argument has come up for different products elsewhere at least twice before) it would not simply balloon out of control.  There's only so much total demand, and that is spread across the entire supply, so if nvidia has 80% market share now, prices could only go up another 20% (if AMD disappeared) before sales would start to drop off.

BS, and you know it. Intel has already fired shots at both Nvidia and AMD that the Iris Pro is better than 80% of the dGPU's on the market. If AMD can't pull their head out of their ass and it costs them going out of business, Intel would keep Nvidia in check.

[vanished]

BS, and you know it. Intel has already fired shots at both Nvidia and AMD that the Iris Pro is better than 80% of the dGPU's on the market. If AMD can't pull their head out of their ass and it costs them going out of business, Intel would keep Nvidia in check.

I don't think we're actually presenting conflicting ideas here.  I'm saying that nvidia's prices could not get more than 20% higher even if they wanted to without causing them to sell fewer cards. Please correct me if I'm wrong but you are saying it would not even get to that stage because intel is also making GPUs and would keep nvidia pricing under control that way.  If so, yes, you are right, but that wouldn't stop me from also being right...

[App4that]

I don't think we're actually presenting conflicting ideas here.  I'm saying that nvidia's prices could not get more than 20% higher even if they wanted to without causing them to sell fewer cards. Please correct me if I'm wrong but you are saying it would not even get to that stage because intel is also making GPUs and would keep nvidia pricing under control that way.  If so, yes, you are right.

I would love to see Intel sell dedicated graphics cards, but I don't think they will. I think both AMD and Nvidia are on a shit or get off the pot position as iGPUs are going to start competing with the lower end dedicated cards. And since Intel has the keys to DX12, things could get really interesting, really quick. Pricing in my opinion will stay consistent through the next series of cards from both companies. So I completely agree that the market wouldn't support a price increase from anyone.

[CostcoSamples]

I personally insult and defend both companies as required. They're both after money and they both use misleading marketing.

You say they are after money as if it's a bad thing (and if not, I apologize for the rant).  Working for money is what we all do, and its the only reason modern technology advances so quickly.  There is nothing wrong wit it, in fact, it's pretty damn awesome.  Without money as a motivator, technology would be a hobby at best (though there'd be very little time for such things since we'd all be living a subsistence lifestyle just like humans did for thousands of years).

 

I do agree with your second point though - misleading marketing is not cool.  I mean, I understand stretching things to sound a little better, but there have been some real dick moves by Nvidia lately.  I'm definitely leaning away from them for my 2016 upgrade.

[CostcoSamples]

I would love to see Intel sell dedicated graphics cards, but I don't think they will. I think both AMD and Nvidia are on a shit or get off the pot position as iGPUs are going to start competing with the lower end dedicated cards. And since Intel has the keys to DX12, things could get really interesting, really quick. Pricing in my opinion will stay consistent through the next series of cards from both companies. So I completely agree that the market wouldn't support a price increase from anyone.

Great points.  It is important to remember that should AMD go bankrupt, Nvidia still has to invest heavily in R&D.  Other players may enter the market.  With there being plenty of competition in the chip manufacturing sector, even a relatively small startup with good funding could, in theory, design a new chip and contract out anything they don't have expertise to do themselves.  Same idea as Google did with the Nexus phones.  They do some design work and sub out the rest to Huawei.  And even if nobody enters the market right away, sales will fall if they can't offer new models that improve upon previous gen, gamers will simply not upgrade.  R&D is what drives their own sales, and to some degree it doesn't matter that much what competition they have.

 

I guess what I'm trying to say is, if AMD dies, GPU advancement might slow down for a time, but will not stop, and certainly other players will enter sooner or later.

[fred_mertz]

It would take years for a competitor to rise should AMD go under.

 

Nvidia wouldn't have to innovate to maintain sales.  Just look at the smart phone market.  Had Android never been created and the dozens of manufacturers created devices, the current iphone would be circa 2008.