Slashdot Mirror
AMD, NVIDIA, and Developers Weigh In On GameWorks Controversy
Dputiger writes: "Since NVIDIA debuted its GameWorks libraries there's been allegations that they unfairly disadvantaged AMD users or prevented developers from optimizing code. We've taken these questions to developers themselves and asked them to weigh in on how games get optimized, why NVIDIA built this program, and whether its an attempt to harm AMD customers. 'The first thing to understand about [developer/GPU manufacturer] relations is that the process of game optimization is nuanced and complex. The reason AMD and NVIDIA are taking different positions on this topic isn't because one of them is lying, it’s because AMD genuinely tends to focus more on helping developers optimize their own engines, while NVIDIA puts more effort into performing tasks in-driver. This is a difference of degree — AMD absolutely can perform its own driver-side optimization and NVIDIA's Tony Tamasi acknowledged on the phone that there are some bugs that can only be fixed by looking at the source. ... Some of this difference in approach is cultural but much of it is driven by necessity. In 2012 (the last year before AMD's graphics revenue was rolled into the console business), AMD made about $1.4 billion off the Radeon division. For the same period, NVIDIA made more than $4.2 billion. Some of that was Tegra-related and it's a testament to AMD's hardware engineering that it competes effectively with Nvidia with a much smaller revenue share, but it also means that Team Green has far more money to spend on optimizing every aspect of the driver stack.'"
This article is very confusing to me
Bottom line, if a game runs poorly on a graphic device AMD and NVIDIA directly get blamed. This program is merely NVIDIA's tack towards improving user perception. They know if you have a problem running software on one of their cards you will probably go buy a Radeon. The computing hardware in each card is far beyond the privy of any single developer to understand at this point. You need a glue layer and technical resources to properly expose the interfaces. The problem is when one vendor is specifically excluded from the glue layer. Both of these vendors have been cheating benchmarks by analyzing what game is attempting to access the features and then dumbing them down selectively in barely perceivable ways to artificially pump benchmark results. The problem I have with NVIDIA doing this is mostly that they typically have their own black box code (that is closed) and you have no idea how that is interacting. If it interacts poorly with your application you are just screwed. There is nothing to fix you must patch around it. Ergo, the state of the current NVIDIA drivers in Linux. =)
When opengl 1.0-1.3 (and dx5/6/7) was king, gpus were fixed function rasterizers with a short list of togglable features. These days the pixel and vertex shader extensions have become the default way to program gpus, making the rest of the api obsolete. It's time for the principal vendors to rebuild the list of assumptions of what gpus can and should be doing, design an api around that, and build hardware specific drivers accordingly.
The last thing I want is another glide vs speedy3D...err I mean amd mantle vs nvidia gameworks.
So it's absolutely understandable that nvidia chooses not to open the driver code.
...try telling that to Stallmen et al!
Seriously though, maybe it makes me a Bad Linux User, but I'm absolutely ok with the state of nVidia drivers: installation is a piece of cake, 2D and 3D performance is great (I think 3D performance is on-par with Windows [OpenGL, obviously]).
I don't have any experience with new ATI cards under Linux, but I've had hit-or-miss luck the times I've used slightly older cards (interestingly, I've had much better luck with 3D performance than 2D...horrible tearing/update problems in 2D, but Nexuiz/OpenArena work fine...).
Frankly, it's time to stop blaming NVIDIA and start blaming ATi, yes everyone likes the underdog. But in this case seriously? They had 20 years to get OpenGL correct. Noone has been blocking them from writing their own drivers for Linux/Mac/Windows. Frankly I think that ATi has made a huge engineering mistake by only focusing on Win32 and by not supporting Unix from day one as a first class citizen, they've shot themselves in the foot, now they expect the industry to clean up the mess by conforming to ATi. I don't recall NVIDIA anywhere holding a shotgun to our heads and demanding we use OpenGL or else. They just made OpenGL available and importantly WORKING. OpenGL wasn't even NVIDIA's project originally, it's inherited from SGI. They've had approximately 20-25 years to implement an open spec, and they've failed to do so at every step. I've been watching the last 13 years as NVIDIA grew from a buggy hard to compile mess on Linux to the stable, fully featured driver it is now. ATi has never pulled off a competent GL implementation in all those years. Now people want to bring in conspiracy theories about NVIDIA blocking ATi from developing software? What a joke.
...try telling that to Stallmen et al!
I'm quite sure he would understand it, I'm quite sure he wouldn't find it acceptable though and that's his prerogative.
...but nvidia offers far better drivers and some extra features like physx
It's more than that. NVIDIAs drivers aren't even that good. It's just that ATI's (AMDs) are just so terrible that they look good in comparison. Who the hell decided the catalyst control center was a good idea? It reminds me of some glitchy 1990s spam ladened chat program. What a joke. The drivers are so sketchy almost every game I'd play would have "STICKY: For ATI users check here first!" at the top of their support forums. Trying to get hardware acceleration to work on my linux media PC was almost impossible until I switch to NVIDIA. Stop creating new cards I can cook and egg on and fix your damned drivers. I have enough fried eggs I just want to watch a movie without spending 30min dinking around with arcane driver settings while my wife keeps asking me why we canceled cable.
It's time for the principal vendors to rebuild the list of assumptions of what gpus can and should be doing, design an api around that, and build hardware specific drivers accordingly.
For the most part, they've done that. In OpenGL 3.0, all the fixed-function stuff was deprecated. In 3.1, it was removed. That was a long, long time ago.
In recent times, while AMD introduced the Mantle API and Microsoft announces vague plans for DX12, both with goals of reducing CPU overhead as much as possible, OpenGL already has significant low-overhead support.
Stallman would prefer if AMD and nVidia would only compete to make better hardware. Which may or may not be realistic. I'm not sure what such a world would look like.
AMD supports openGL just fine, but they aren't gracefully failing sloppy programming. The Nvidia driver tends to try and make "something you probably sort of meant anyway" out of your illegal openGL instruction and AMD fails you hard with an error message. That's no reason to blame the manufacturer. The game developers deserve blame for sloppy coding and sloppy testing.
I was promised a flying car. Where is my flying car?
Heh... You'd like to think that, but you'd be mostly wrong.
The problem with "broken" drivers and "quality" is less an optimization and overall fit-and-finish deal with AMD versus NVidia- and more of something most wouldn't get unless they'd been digging in either company's proprietary codebases at some point combined with being IN the Games Industry and really understand the story properly.
AMD's drivers tend to explicitly follow the OpenGL standards. To a fault.
NVidia's compensates for many inappropriate mis-uses thereof (But not all, mind...) but oftentimes doesn't get it fully "right"- but it's something that the game studios code towards...the noncompliant driver behavior.
Take a long, wild guess what devs at the Studios TYPICALLY do? They implement code that's "clever" and omits a shader parameter here, an Alpha there- and largely works with NVidia's drivers at speed because NVidia's code makes assumptions when you're missing stuff. They implement code that does things like recycling VBOs intraframe instead of the sensible interframe use (Because if you'd have READ the spec, you'd know that there's a good chance you'll stall the pipeline doing this stupid thing...) and dragging a top end card from a hundred or so FPS to seconds per frame.
Correct the engine code doing the rendering so it's not doing stupid crap and the game FLIES on both company's products and maxed out 100+ fps and does it stably...on both GPU targets.
AMD would love to get in the path with the studios like NVidia is right now, so they can work with them and help them make performant code that doesn't take shortcuts. In many cases, their games don't run as well as they could on both NVidia and AMD because of the stuff that NVidia's propping them up on. If you did it "right", the games would actually tend to run faster- and it's not even remotely hard to get it "right". Hell, they don't even get it "right" on D3D in the same senses, so it's not that OpenGL's difficult to code for. It isn't. It's just that nobody ever stops to check to see if they've pooched themselves on an API edge by...funny this...reading the damn standards and documentation for a change. AMD doesn't want "NVidia's 'optimizations'" because if they did, they'd have been reverse engineering them all along and getting damned close to them. Thing is, it's mostly a waste of time trying to play "catch-up" on things there- so, they work on the fast "correct" path per the specs and try for larger marketshares to get where they want to be.
(And a hint for you: The reason NVidia's not opened up is more of a brown-paper-bag-over-the-head reason than the one you allude to- much of their stuff isn't all that great...)
And Apple has followed that with its own "Metal".
"Hardware-wise AMD currently has better bang for a euro than nvidia,"
I would say this is a matter of opinion. The 750 Ti is a MONSTER at its price point, especially with ShadowPlay.
Good-bye
I'd also add that considering the NVIDIA Binary blob works on FreeBSD, Solaris, Mac OSX, and Linux as well as Windows, that it is well engineered. The AMD/ATi driver doesnt' even work correctly on Linux, and Apple had to write their own driver for Mac OSX. There is an officially available (from nvidia.com) driver for Mac OSX for their Quadro cards. It is pretty obvious that AMD/ATi has always favored Windows/Microsoft and has put minimal effort into supporting Unix based platforms. Now they're reaping what they've sown and everyone's trying to defend them? No. You don't get to abandon a platform for years, then try to claim victimhood status for your own poor business choices. That's doesn't fly. Intel is having their first crack at this graphics stuff within the i915-current drivers. They're solid, reliable, a little lacking in features. But overall it works. My code written for Linux works on both nvidia and Intel. Not sure why the fact it breaks on AMD doesn't shout to people that AMD is the problem, but there you go.
Last I checked, the 750 Ti had amazing performance per watt, but performance per dollar it fell behind AMD cards such as the 260X and 270X.
It reminds me of some glitchy 1990s spam ladened chat program.
Sounds to me like you are using a 1990's card too, AFAIK "catalyst" is no longer supported and it's certainly not bundled with recent cards. I updated my NVIDIA driver just the other day, sure the driver is enourmous (250MB) but it installed flawlessly in the background without a reboot. I play WoT regularly at maximum detail on an i7 and have no issues other than the 200ms round trip from Oz to the US but it stays playable until that hits ~350ms. I've also been mucking around with CUDA for a few months, the developer resources are excellent and free for non-commercial use. If you really want to squeeze every last flop out of the card NVIDIA provide free resources such as the online book "CPU Gems" and the white papers that accompany some of the demo source, such as the optimised n-body example.
Stop creating new cards I can cook and egg on
My $150 GE Force 750 maxes out at just over a terraflop, significantly faster than ANY super computer that existed pre-Y2K. It uses less wattage than an old fashined light bulb, sure it can fry an egg, even scramble it with the on-card fan, but why is that a problem if it remains within its operating specs, is it that difficult to keep your eggs away from the video card. If your chickens are attracted to the heat then move the chickens outside where they belong.
In my professional opinion, I think the problem on your system is the chair to keyborad interface, it has nothing to do with NVIDIA or AMD's trully amazing technology since there is absolutely no need to fddle with default driver settings just to watch a movie. Listen to your wife and save yourself two headaches, forget about the PC drivers and just pay for the damned cable.
And did you exchange a walk on part in the war for a lead role in a cage? - Pink Floyd.
to me it sounds like again like the beginning of Internet Explorer vs. Firefox compliance to HTML standards.
Down to the detail of how it pans out:
- one company being the popular one (Microsoft, Nvidia), so everybody code to their platform (IE, drivers) and end up unknowingly produce bad that code that happen to rely on the peculiarities of this platform (the non standard assumption of Nvidia's drivers, the weird re-interpretation of HTML done by IE's engine). When there are problem, they tend to hack their own code.
- the other company being the underdog (Mozilla, AMD) making a platform (Firefox, Catalyst) that tries to follow the open standard to the letter (HTML5, OpenGL), but in the end other person's code (websites, code) behaves poorly, because it breaks standard and relies on quirks that aren't present in that platform. The users complain of problem (broken HTML rendering worse under Firefox than IE, non-compliant openGL code's performance being more degraded on AMD then Nvidia hardware).
Funnily, if past history is any indicator, on the long run AMD's approach is better and either them or one of their successor is bound to manage to bring opengl-compliance more important than driver tricks.
(the fact that AMD is dominating the current iteration of consoles, might help bring more power to them)
Interestingly the embed world might one also end up helping just like it did the browser wars (Internet Explorer was far less prevalent in embed machine like PDA/Smartphone/Tablet than on desktop and the problems with broken HTML became much more apparent, and compliance with HTML5 [sure to run on as much platforms as possible] was determinant. Also the embed eco-system mostly centered around compliant engine (like Webkit)) due to the same factors (extremely heterogeneous ecosystem hardware-wise, where Nvidia is just one player among tons of others with their Tegra platform. compliance with OpenGL ES is what is going to be determinant as the embed platforms are going to need a lingua franca to insure that porting an engine is as smooth as possible and works easily on all smartphones/tablets, no matter if they boast PowerVRs, Vivante, Lima, Adreno, etc.)
Maybe we might need something along Acid test and w3c conformancy test to exercise drivers and test game code for standard non-compliance.
(That partly exist as "piglit" - the test suite that freedesktop.org uses to test opensource mesa and gallium drivers).
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
AMD's perspective is that Mantle is less problematic:
- Mantle's spec are open.
- Also it's just a very thin layer above the bare hardware. Actual problems will mostly be confined in the actual game engine.
- Game engine code is still completely at the hand of the developer and any bug or short coming is fixable.
Whereas, regarding GameWorks:
- It's a closed-source blackbox
- It's a huge midleware, i.e.: part of the engine itself.
- The part of the engine that is GameWorks is closed and if there are any problems (like not following standard and stalling the pipeline) no way that a developer will notice and be able to fix, even as AMD are willing to help. Whereas Nivida could be fixing this by patching around the problem in the driver (as usual), because they control the stack.
So from their point of view and given their philosophies, GameWorks is really destructive, both to them and to the whole market in general (gameworks is as much problematic to ATI, as it is to Intel [even if it is a smaller player] and to the huge diverse ecosystem of 3D chips in smartphone and tablets).
Now, shift the perspective to Nvidia.
First they are the dominant player (AMD is much smaller, even if they are the only other one worth considering).
So most of the people are going to heavily optimise game to their hardware, and then maybe provide an alternate "also ran" back-end for mantle. (Just like in the old days of Glide / OpenGL / DX backends).
What does Mantle bring to the table? Better driver performance? Well... Nvidia has been into the driver optimisation business *FOR AGES*, and they are already very good at it. What is the more likely, that in case of performance problems developers are going to jump on mass to a newer API that is only available from one non-dominant PC player, and a few consoles, and completely missing on any other platform? Or that Nidia will patch around the per problem by hacking their own platform, and dev will continue to use the ?
In Nvidia's perspective and way to work, Mantle is completely irrelevant, barely registering a "blip" on the marketing-radar.
that's why there's some outcry against GameWorks, whereas the most Mantle has managed to attract is a "meh". (and will mostly be considered as yet another wanabe-API that's going to die in the mid- to long-term)
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Developer see NVIDIA has better experience in creating great hardware than AMD, I assumed.
Its feature set, in my opinion blows away those cards. Better drivers, cheaper to run, cooler, no power connections needed. Cheapest doesnt mean 'best value'.
Good-bye
First, I don't think AMD (or any other company) execs would recognize driver optimization if it hit them in the face.
Second, do you think nVidia is hiring from a different talent pool than AMD? Neither company has any special secret magic sauce driver optimizations that a well trained monkey at the other company cannot come up with. If you look into nouveau and radeon open source kernel and Mesa drivers you will be able to see how much easier nVidia hardware is to work with, that may be one of the reasons nVidia drivers are less of a mess. But not by much. I had a fair share of driver hangs and memory corruptions on my laptops with nVidia cards (GTX 560M on Asus G53SX in particular).
AMD's drivers tend to explicitly follow the OpenGL standards. To a fault.
That is a popular excuse, especially for the open source drivers that frequently have problems with newer commercial games, but having more complete support for what is in the standard and being more permissive to what is not are not mutually exclusive. For example, see this page for some actual conformance testing results: http://www.g-truc.net/post-0655.html#menu As you can see, the Nvidia binary driver clearly passes a higher percentage of the tests than any of the others, and it is the only driver to pass all samples from OpenGL 3.3 to 4.4.
From a consumer's point of view, it is also a poor attitude from Mesa developers to interpret "implementation defined behavior" as "license to break anything as we see fit" (GCC developers tend to do the same, by the way, even though the compiler has its own set of non-standard extensions as well). They are free to add a configuration option that lets the user choose between strict conformance (mainly for developers testing their code) and maximum compatibility, but the casual consumers will not care why the game they paid for fails to work, if it keeps happening, they will ignore the excuses and just delete Linux and go back to Windows/Direct3D.
...Stop creating new cards I can cook and egg on...
I think I've found your problem. What you are looking for is called a skillet, and it does not go in the computer.
Nvidia PAYS for removal of features that work better on AMD
http://www.bit-tech.net/news/h...
Nvidia pays for insertion of USELESS features that work faster on their hardware
http://techreport.com/review/2...
Nvidia cripples their own middleware to disadwantage competitors
http://arstechnica.com/gaming/...
Intel did the same, but FTC put a stop to it
http://www.osnews.com/story/22...
so how exactly is that not Nvidias doing??
Nvidia is evil and plays dirty. They dont want your games to be good, they want them to be fast on Nvidia, any means necessary. They use "means to be played" program to lure developers in, pay them off and hijack their games to further nvidias goal.
For example how come Watch Dogs, a console title build from the grounds up with AMD GPU/CPU optimizations to run good on both current gen consoles, is crippled on PC when played on AMD hardware? How does this shit happen?
This is something FTC should weight in just like in Intels case.
Who logs in to gdm? Not I, said the duck.
In my experience the AMD drivers&software have been more stable. Probably I just have bad luck, but just today I got a pop-up from the Nvidia Experience that there's an update. Clicked it and got 'cannot connect to Nvidia servers' or something similar. The last driver update failed to update one of the components. And the display adapter has crashed once (managed to recover though).
I'm currently using GTX 670 (got the Windforce model and it is really quiet) and I'm reasonably happy with it, but I had none of the update problems with my previous HD6850. Of course, my thinking may be coloured by the horrible nvidia chipset I once had. Or the lack of support for my previous laptop gpu. Compared to those, my current GPU problems are minimal at worst.
It is what it is.
what driver issues? i have yet to see gaming related drivers issues...
launch day games i've tried with no issues on my 5870 include BF3 (includling "paid beta"), Sup Com 2, Starcraft 2a, Portal 2, Crysis 2 trial, NFS: Shift, Metal Gear Rising, Borderlands 2, Skyrim*, Train Simulator 2014* (* = launch + 1 month)
only issues i've had were a few demoscene nvidia-favoring demos and bitcoin-related OpenCL driver combinations.
It's a well known reality that ATI had severe problems with driver quality from early stages, and these problems persist today long after AMD bought the company.
And by "from early stages" you mean from the beginning, I hope. I've been having ATI graphics blow up windows since 3.1 with the Mach32. Even RADIUS made more reliable video cards. I wish they'd stuck around and ATI was gone now.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
I'd say the fundamental problem is that the specifications themselves are a patchwork of code changes written in a natural language.
The original specification is written before the original driver code is modified, or derived from an existing driver for one hardware system, and then recoded for a new driver for another hardware system. With other device drivers (networking), each extension specification is actually specified in a high-level language which can be processed straight into device driver code.
Direct3D has the advantage that the hardware must match the software specification, while OpenGL is more extension applied over extension on different hardware. Since each vendor has different hardware and supported extensions, the implementation of one extension may or may not affect other extensions. For example, you could support FBO (framebuffer objects) using textures as a destination. But then if you implement compressed textures, then those textures can't be used with FBO's, and so additional code has to be added to prevent that use. Usually the reason that you can't use a particular combination of extensions is simply because the hardware logic hasn't been implemented yet.
Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
You should look at the latest OpenGL ES specification. This is OpenGL optimized for mobile devices and gets rid of most of the old API bits while still supporting vertex, fragment and compute shaders. Anything else is just implemented using shaders.
But mantle gives you access to the hardware registers (those descriptors) while avoiding the overhead of updating the OpenGL state, then determining what has changed and hasn't, then writing those values out to hardware.
Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
AMD made about $1.4 billion off the Radeon division. For the same period, NVIDIA made more than $4.2 billion. Some of that was Tegra-related and it's a testament to AMD's hardware engineering that it competes effectively with Nvidia with a much smaller revenue share, but it also means that Team Green has far more money to spend on optimizing every aspect of the driver stack.'"
While that's true for revenue, the difference in profits between AMD and NV are very close.
http://www.accountkiller.com/en/delete-slashdot-account Stop visiting Slashdot.
I don't use nvidia experience as I prefer to have manual control over most of the card's features. That said, I've had the update problems as well, and most of them were firewall-related.
Good for you. I've had more than my share of games that worked atrociously on release date when I was still playing on 4870 and driver settings that just plain refused to work.
I'm not alone in that experience either unfortunately. It's not like it has scared me off AMD, as I said I still use it. Just not for performance stuff where I need reliability and stability more than a few extra FPS.
Not sure why the AMD hate. I gave my friend my 3 year old AMD 5770 that I bought for 150 a couple years ago and he plays everything on high. I feel they have better Linux drivers by far and they have attempted to keep old hardware around longer. In my experience AMD has won bang 4 buck and lasts longer just stay away from xx and other cheap hardware brands. NIVIDIA is trying to push them out they are already trying to force nividia only hardware on monitors, the companies are saying no universal is the key will give you a slot.
Their OpenCL drivers are pretty bad. Completely arbitrary changes can have huge impacts on things like VGPR usage. Optimizing an OpenCL kernel on an AMD card is like black magic. Not that I'm praising NVIDIA here, they're still on OpenCL 1.1...