We do often include affiliate links to earn us some pennies. See more here.
image
I got an interesting email from Michael A. Marks the technical director for Aspyr Media who built his own steambox and ran some OpenGL tests with interesting results.

Before you read too far into it just bear in mind that Michael knows what he is talking about given he's worked on some AAA titles like Call of Duty & Civilization in the porting process from DirectX to OpenGL.

Benchmarks:
image
image

Not noted in his article is that these tests where done on Ubuntu 14.04 LTS with Nvidia 331.79 and AMD 14.6 beta drivers which I managed to get after talking with Michael for a bit. So, it's a recent Linux distribution with recent drivers.

The benchmark used is this tool on github for reference in-case anyone else wanted to give it a go if they have different cards laying around. I would be very interested in seeing results from others.

That's a MASSIVE gap in performance between two providers, considering the AMD card is much more expensive too I didn't think it would be that big a gap. AMD only wins once and although it's a big gap on that one, it utterly fails at everything else. It's not like the AMD card is new either, so it cannot be blamed on drivers playing catch-up with hardware as it's ~8 months old.

Michael A. MarksThis breaks my heart. You present a talk in March of 2013 telling me about all the cool new features, present a bunch of performance graphs, and in reality: ‘it’s completely broken on AMD’

You have to assume AMD knew the talk didn’t apply to Linux in March. The alternative that they weren’t aware is even more heart-breaking. Full results are included in another post on OpenGL.


It's a sad state to be in and I really hope AMD pull it out and get their heads down on pushing OpenGL performance, but while they are so busy with Mantle, do they care? I wouldn't think so, they might care if SteamOS takes off in the way that a lot of people hope it does otherwise people will steer clear of them.

I am optimistic for the future of Linux gaming given how fast games are being pumped out for us, but I just like everyone else want hardware vendors to look at us just as serious as game developers are starting to. I know a lot of people are quick to come to AMD's defence for their open source driver effort, but results are results.

I seriously look forward to the day I can write a headline along the lines of "AMD kicks Nvidia's arse". Bring it on AMD the ball is in your court.

Source article "OpenGL Stop Breaking My Heart" Article taken from GamingOnLinux.com.
Tags: Misc
0 Likes
About the author -
author picture
I am the owner of GamingOnLinux. After discovering Linux back in the days of Mandrake in 2003, I constantly came back to check on the progress of Linux until Ubuntu appeared on the scene and it helped me to really love it. You can reach me easily by emailing GamingOnLinux directly. Find me on Mastodon.
See more from me
The comments on this article are closed.
38 comments
Page: «2/4»
  Go to:

no name May 31, 2014
  • AMD should start to be consistent on linux: one open GNU GPL protected driver and drop that horrible thing which is catalyst. Their closed source driver is illegal anyway. Linus T. is just letting them being naughty... for the moment...

  • They should start to code properly in linux! Sometimes I wonder if their code is not straight from crosoft shi., because it's sooooooo broken and bad. They should use common sense too: it's far better to have more code with simple code paths than less code but with brain damaged code paths (Over code factorization). But I'm bad mouthing: it seems they are *extremely* slowly moving towards that equilibrium

  • Opengl is a joke: from nvidia software engineers (!!), very little of the opengl API should be used in order to get the best of modern GPUs (see the steam dev days).

  • benchmarks are most of the time quite bias: that benchmark was optimized for that GPUs, the driver was optimized for this benchmark... or the benchmark code paths happen to favor this driver and not the other ones, etc. Devil hides in the details.

  • hardware design is easily worse than software design... then the driver won't be nice

  • As far as I'm concern, dota2/portal2/team fortress2 do run faster with the radeon open source driver (mesa/radeon southern island).

commodore256 May 31, 2014
Quoting: Anonymousamd sucks. I hope they go out of business. I really do.

So you want Nvidia to be a Monopoly? Also, I had a 4870 and the open source drivers for that Card out performed the GTX 460 Linux Binary Drivers that I replaced it with a year ago. (even though that Nvidia card runs better in Windows than that 4870 ever could)
manny May 31, 2014
if amd focuses even more on the open source driver like intel does, I think that could be a game changer for the industry as a whole.

then nvidia would be the last left with closed drivers
Yulian May 31, 2014
It's worth mentioning that this tests are for 'cool new features in OpenGL 4.x', as mentioned in the source. He even says that most of them are not available in the Mac drivers, so I doubt there is any game on Linux currently that takes advantage of them. In my experience, the problem of the catalyst driver is not the performance, but the bugs that can even crash your system. Several devs mentioned that they had to work around those issues.
storma May 31, 2014
Quoting: Anonymousamd sucks. I hope they go out of business. I really do.

Agreed, The Nvidia open source drivers are heaps better.

Honestly, I think there is more to it than just Catalyst being crap. I would agree that they aren't as robust as the Nvidia drivers but from personal experience, they are capable and improving. Others through out the years have had similar experiences.

Currently my only long lasting niggle has been with the micro stutter in the heaven benchmark, which I don't experience elsewhere.

I don't have windows, so I can't compare gaming performance between the two. I haven't had the time recently to play to much but The witcher 2, for me, is playable with the latest beta drivers. All other games I throw at it are fine too.

Whether it's my use of Debian stable/testing or my AMD/AMD setups, I'm not sure what accounts for my mostly hassle free experience.

HTPC: 960/6670, stock Debian testing w/ OS drivers.
PC: 8350/7870, Debian testing w/ custom 3.14 kernel, latest Catalyst drivers.

Past rigs I have used a 2600 and a 4870, with similar AMD/AMD setups. I made the switch with the open sourcing news.

*Hopes he doesn't jinx himself. ;)
Deformal May 31, 2014
Hm. Technical director for Aspyr Media has interest in Steam OS. Steam db tell as about signs of Linux support of Civilization V. Civilization V was ported on Mac OS X by Aspyr Media.
So as I understand, we have another Mac developer, that starts to make Linux ports.
dyp handylinux.org May 31, 2014
hello, seems NVidia cool, but in fact, on aor forum, the most problematic issues are with NVidia.
Half-Shot May 31, 2014
Quoting: commodore256
Quoting: Anonymousamd sucks. I hope they go out of business. I really do.
So you want Nvidia to be a Monopoly?

This. Wake up people, this is why Microsoft have a bleeding control over desktops. Other operating systems went out of business due to problems of any kind and MS just took CONTROL of the general desktop market, its taken a long time for us to start competing again.

On the state of AMD. Their open drivers are what keep me sane these days. The catalyst drivers are a mess and the open ones are great for what they are. If AMD didn't support Open I would be deeply saddened but at the moment I'm just nervous for their future.

In all honesty, if they did a Intel and just fully open the lot then they would actually gain a lot of sales for likeminded people who support open over preformance.

And just a stab in the dark but Nvidia have been the most anti-competitive company in the GPU space since Microsofts numerous protocols that we have had to catchup on. The sheer number of prop extensions they have pushed developers to use has not helped AMD one bit.
Anonymous May 31, 2014
Unfortunately, results like these aren't so cut and dry. I'll look over the details of each test in a few days, but quite often there are little things, from buffer sizes to vertex formats, which greatly sway test results in one direction or another.
Anyone posting such numbers should be aware of this.
jarfil May 31, 2014
Ok, so the difference is obvious. Still, I'm kind of sick of people saying that 100 minus 647% equals anything but -547.
While you're here, please consider supporting GamingOnLinux on:

Reward Tiers: Patreon. Plain Donations: PayPal.

This ensures all of our main content remains totally free for everyone! Patreon supporters can also remove all adverts and sponsors! Supporting us helps bring good, fresh content. Without your continued support, we simply could not continue!

You can find even more ways to support us on this dedicated page any time. If you already are, thank you!
The comments on this article are closed.