Slashdot Mirror
NVIDIA Performance On Linux, Solaris, & Vista
AtomBOB suggests a Phoronix review comparing the performance of a Quadro graphics card on Windows Vista Ultimate, Solaris Express Developer, and Ubuntu Linux. The graphics card used was a NVIDIA Quadro FX 1700 mid-range workstation part. The cross-platform benchmark used was SPECViewPerf 9.0 from SPEC. Quoting Phoronix: "Using the Quadro FX1700 512MB and the latest display drivers, Windows Vista wasn't the decisive winner, but the loser... Ubuntu 8.04 Alpha 5 with the 169.12 driver had overall produced the fastest results within SPECViewPerf. In only three benchmarks had Solaris Express Developer 1/08 outpaced Ubuntu Linux, but with two of these tests the results were almost identical.""
I've wondered this a while. What is the difference between the gaming cards and the workstation cards from Nvidia and ATI? Do they just have better DACs? Certified driver support for business apps? Or is the GPU itself somehow?
Alex
I am surprised by this as I would have thought Nvidia would have put more effort into their Vista driver with Linux drivers being mostly on the back burner. I am assuming it is because their Linux driver is old code (which we all know contains less bugs then new code) whereas the Vista driver is written from scratch?
Either way I think this shows the awesomeness of Ubuntu and Linux. ^_^
This is serious question, I heard a while back that Vista had done something to make OpenGL slower.
Could Vista's bad performance be due to its nerfing of OpenGL on Vista in order to get developers to pick DX?
Vista has a new driver architecture and it is goiing to take some time for MS to improve the graphic subsystem performance. It will also take NVidia a while to optimise their code for Vista.
Even then, the Vista architecture might just have some inherent issues that are hard to code around.
Engineering is the art of compromise.
If you're a quattro user, your OS choice would surely be on software available for whatever particular professional application you are using the card for. As a sound designer, that would be for me, XP. I don't think many professionals are ready to jump to vista quite yet so I'm surprised that they have not included it. We are, after all, looking for stability.
I record my sleeptalking
Its an OpenGL test. The perf. difference between OpenGL and DirectX Nvidia implementations has always very large -- even in Windows XP.
What?! Windows did not have the best NVIDIA performance?!
This is a new one. No, really. Usually NVIDIA makes their Windows drivers their best drivers, and Linux is supported as an afterthought because they can make a few percentage points more in sales this way, and because it discourages reverse engineering their hardware, since those who would take the time and effort to do so won't on account of there being a working solution.
In other words, I am surprised that although Windows Vista has been such a mess in terms of compatibility and speed, that even the NVIDIA benchmarks put it last.
I read Slashdot every day, and until this moment I had never even heard of PCLinuxOS. I had to look it up.
Ubuntu, however... Ubuntu, my parents have heard of.
Don't know what metric Distrowatch uses, but it seems to be flawed.
Granted, I don't use Linux as a day-to-day OS, but I have some Linux apps I like which I run via Ubuntu in VMware Fusion. As a casual user, of the distros I've tried, Ubuntu wins hands-down. It's still too hard to set up for my parents, say, but not so hard that I don't just say "fsck it" and delete the partition, as I have done with all the others.
Uh... Dude? Did you even RTFA? (No, that'd presume too much- it is Slashdot, after all...)
Ubuntu PASTED Vista, and fared really good against Solaris, even when it was beaten by it.
Reality is, this largely has nothing to do with whatever Distro you care to favor- it's that an out
of the box Linux distribution pretty much pasted an out of the box Vista install.
Nothing more. Nothing less.
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
One more step towards Desktop Linux. But we need some real games to use these 3D capabilities!
Persian Project Management Software as a Service
Matrox had their lesson trying to sell their excellent cards to gamers. Everyone came to forums and whined about seeing "30 fps" while their friends have "120 fps". Some sane people tried to tell the specs of human eye but it didn't matter.
,his game got locked to 30 fps :)
I bet there are rich but non techie guys buying Quadro for gaming right now. I know a one bought ATI FireGL along with 15K RPM SCSI disk and couldn't sleep because of noise. Not just that
If you're running Maya, would should be running the drivers/distro that Autodesk blesses. Last I checked, that was 2-3 year old drivers on RHEL 4/SLES 9/Fedora Core 5. I run the blessed packages for a small animation studio and only have problems when people out of memory their system (8GB RAM should be enough for anybody). http://usa.autodesk.com/adsk/servlet/item?siteID=123112&id=9683256 has the list of blessed stuff.
There's definitely a different between say, 30 and 100 fps: http://100fps.com/how_many_frames_can_humans_see.htm
This space for rent.
I have it on good authority that the next Windows Driver Model will run Crysis on 3 SLI 8800GTs and render it in 8-bit color at 640x480 resolution at over 50 FPS! So take that you Linux/Unix hippy beatnik freaks!
That's interesting. Considering that I am a developer for the CHUD Tool (no quotes) and I do performance analysis and benchmarking for a living, I don't think they did anything wrong. Things that aren't running on a system rarely affect run-time performance. Going from a distribution like Ubuntu to Debian just removes a bunch of things from disk, but those things have zero impact on the metric being measured. For Vista, it might make a difference if the version used was shown to have less idle activity, but in practice, you want to compare what a typical user would be running. So, since the OSs chosen reflect typical users, the data is perfectly valid for a comparison between them. If you want absolute performance numbers, then you need to start tuning the OSs before you run the tests. Things like disabling daemons or services and unplugging network cables can cause measurable differences in some benchmarks.
As for the CHUD Tools, they are completely inert unless you happen to be running one of the tools and even then, it isn't likely to cause any significant difference. The kernel extensions used by the CHUD Tools are designed to do absolutely nothing until they are asked to. If you are running a Time Profile in Shark, it will have some impact, but it will be limited to 1-2%.
kc8apf
No, you've just accidentally switched to Firefox on Slashdot.
Wait a few moments, the cognitive dissonance will pass and you'll be up and trolling like a champ again.
"I've got more toys than Teruhisa Kitahara."
You think? I remember and have followed a similar kind of scenario, it started over 30 years ago when I was younger. It went something like this: GM was the Biggest of the Big. Had a market share greater than all other automobile manufactures combined. Had revenues higher than the GNP of 90% of the worlds nations. etc. etc. etc. They developed this Business Model called "Planned Product Obsolescence". (Your vehicle was planned to be scrap in about 10 years or before.)
There was also another little automobile manufacturer called Toyota with a very small market share, they made crappy little vehicles, used to be called "piss pots". They had a Business Model called "Continuous Improvement". There was a historic event in 2007 that went quietly unnoticed, Toyota surpassed GM in world market share and revenues.
back on topic...nvidia pdf from september of 2003 explaining the differences. Yeah, old, but it's the only document on nvidia's website that I could find that would explain the differences.
You argument is flawed. You're arguing because windows is better known it will always
have the largest market share. The same could have been said about IBM pcs, or lotus 1-2-3,
Borland's compiler suite, or wordstar word processor.
The fact of the matter that next winner has to start out small because it gets to grab
marketshare. Google is an excellent counterexample to your argument. They were just 2-3 people
in 1998 working on a master's thesis project when Yahoo and AOL were the big thing. And where
is AOL now? How much marketshare does Yahoo have for search engines?
Personally I think that Dell selling preinstalled Linux boxes in the U.S. was the first toll
of the death bell for Microsoft. Then walmart selling out the green PCs was the next tolling of
the bell, and now that Asus is selling Eepc laptops I think is the first nail in the coffin for Microsoft.
Will Microsoft die overnight? No. Will they go out with a bang? No. I think they will go out with
a whimper within the next 5 years unless they somehow manage to reverse their course like they
did in 1995 and embrace the fact that GPL software is here to stay and start using it.
I have to call BS on that. If I have to choose between the latest versions of Open Office and Microsoft Office, I will take M$'s closed solution hands down. The interface on 2007 is vastly improved over other office offerings out there. Making something free and open source does not make it good. I can think of many free applications that don't make the grade in cleanness and usability when compared to commercial offerings.
These are the same people who when asked what kind of computer they have answer with "black". Also, not many people can associate the maker of the softare they use with the actual software application. You ask them which browser they use and they will say "I don't know. I just click on the blue 'e'." despite the fact that the title bar says "Internet Explorer" 100% of the time the application is open. So I hope you don't expect them to know Microsoft created it if they don't even know its name.
As far as marketing capabilities, I hardly ever see a Microsoft commercial. When I do they don't ever specify any particular product in the commercial. How does that really sell Windows or Office? All the marketing seems to happen behind the scenes from the point of view of the end consumer using deals that happen between OEMs and Microsoft salespeople.
this nation, under God, shall have a new birth of freedom. -- Lincoln, Gettysburg Address
I'm as much for open source as the next guy, but for the love of all that is holy, what are you talking about? If marketing can be open sourced, how will it work? "Someone will innovate and figure that problem out"?
Square pegs don't fit every type of hole. No matter how much we sit around and think about it, no "innovation" will make it fit. We can make some sort of hack and call it a square peg fitting in a round hole, but it isn't really.
The difference between programming and marketing is that marketing isn't about standing on the shoulders of others. Giving away your previous work isn't going to help your successor market to any significant effect.
They have invented "open-source" marketing in the sense of hacks, like viral marketing, that aren't really open source but sort of a vague gesture in that direction, but don't expect traditional marketing to be going anywhere.
*I think I just hemorrhage about 5 mod points indirectly with this post at a poor attempt at humor
Once you start despising the jerks, you become one.
An Education is the Font of All Liberty
man woman Story of my life
It is difficult to get a man to understand something when his job depends on not understanding it.
This is an openGL test. Nvidia's linux drivers for openGL have been really fast for a long time now. In fact they've confirmed that they use the same driver code for windows and linux, just with a different API exposed.
What you're talking about is that the video acceleration APIs are not exposed for linux (purevideo). This is still the case, and annoying.
There has been hardware rescaling to TV modes on their cards for a few years so you'll find the cheapest models with TV-out do a good job. Other features have improved a lot in the linux and other drivers - look at the README on the nvidia download site for the long list and how to turn some on or off.
Microsoft has been reporting 15% growth in revenues the U.S., 20 to 30% growth abroad each quarter. [emphasis added]
Amazingly enough, Microsoft has been known to lie about some things. I suggest you review the fine print on those "reports", and then ponder why, if Microsoft's growth is really as reported, their stock hasn't been doing as well as it historically used to. Their cash reserves are also shrinking. Then there are the legal battles they're fighting.
That is the picture of a company on the way down.
-- Alastair
Comment removed based on user account deletion
You might not be able to tell the difference between 30fps and 120 fps. I can (and I'm not superhuman). I probably can't tell the difference between 85 and 100, but I've been playing games long enough to know there's a perceptible difference between 30fps and 60fps. 30fps is just "playable", >= 60fps = "good.
Just find a game, play it at a low FPS and then compare it at a high FPS. I used to play Doom and Doom 2, and believe me in many cases low res high FPS was better than high res low FPS. Plenty of other games allow you to lock the fps, and unlock the fps etc. If you can't tell the difference, well good for you and bad for you, good = you don't need to spend so much on graphics cards, bad = your eyes are probably below average in that.
33ms is a substantial amount of time for games. Even though the picture gets "smeared" in the time domain due to the way eyes work, there's a difference if you are seeing some stuff 25ms later, because it's just "not time to show the frame yet".
The Quadro boards allow OpenGL stereoscopic images to be displayed in a window, and the non-Quadro boards do not. If you want really good 3D, you need a Quadro.
I use them for my stereoscopic video stuff with either a pair of shutter glasses or 3D HMD goggles, and can do a live, 3D viewfinder to compose the scene, align cameras, etc.
Maybe you've heard of a little game studio called Id Software? Or Epic Games? I'm not even going to mention what works on Wine.
Oh, I don't know, Maya? That's off the top of my head -- I don't do 3D professionally.
But while we're at it, why did you bring up games in what is clearly an article about professional graphic design hardware? Or do you actually buy Quadro cards and wonder why your games run like shit?
Like what? Closest I can think of is blender, which is under the GPLv2. Is that what you're not talking about?
Yeah, because that was totally unique to the Titanic. Except it wasn't -- it actually stands for "Royal Mail Ship".
I'm doing nicely, thank you.
Never suspected it was the Windows fanbois living in their mother's basements all along, though. Thanks for that, you just made my day.
Don't thank God, thank a doctor!
I lol'ed the first time I need to check my security log, and typed `less secure'.
Acts 17:28, "For in Him we live, and move, and have our being."
Not really, OpenGL and DirectX have always been more than competitive.
Also OpenGL technically benefits MORE from the new WDDM in Vista because of the RAM allocation system and GPU scheduling as the OS handles all these details for OpenGL and OpenGL applications.
The ICD still has to be optimized to pass through and work with the new Vista WDDM model, so as Vista was first released to now, just like with DirectX - OpenGL on current drivers is considerably faster than the horrid RTM drivers from both NVidia and ATI for Vista.
Right now in MOST circumstances, even games running in an Aero Window, they are running faster under Vista than XP, no matter if they are OpenGL or DirectX.
http://www.opengl.org/pipeline/article/vol003_9/
One game a technician here plays is City of Heroes, that is a hybrid DirectX/OpenGL application (see NCSoft for more details), the tech runs the game inside a Window with Aero active, as it is 10% faster than running it full screen, which turns off Aero's composer. Why the improved performance with Aero is unknown, but measurable and a testament to the speed of how Aero is implemented with the shared device context and texture methods it uses instead of dual memory or double buffering like you find with Linux or OS X.
I personally have more regard for DirectX because of being involved with SGI and the 90s OpenGL specifications, where MS couldn't force the OpenGL participants to move to 3d Hardware gaming type constructs, even after writing a few test specifications for OpenGL. If OpenGL would have had a better view of the future, there would have NEVER been DirectX as MS wanted to be a big OpenGL proponent.
I think OpenGL shot themselves and a lot of users in the head with their closed minded moves, and if it hadn't been for the gaming movement of Linux when 3D acceleration was becoming a normal aspect of computing, OpenGL to this day might have disappeared or remained a 'high brow' 3D specification that didn't want to dirty their hands with more direct 3D hardware support or features condusive to gaming.
Anyway, check out the link, there are several posted about OpenGL performance on Vista in the past few months comparing both it and DirectX to various situations and XP, showing that the rewriting and optimizataion of the Vista drivers fro NVidia and ATI for the past few months are finally as mature as the XP drivers. (Which isn't too bad considering they were written from scratch late 2006, with no real world performance or game profiling optimizations that the XP drivers had built on for years.)
Here is another thread a tech here has been following and forwarded to me this morning, since I was reading it right before I flipped to SlashDot, I thought I might as well include it as well, not a concrete study or test, but more of what users are experiencing to their surprise after all the negative Vista vs XP press:
http://futuremark.yougamers.com/forum/showthread.php?t=72298