Tomb Raider Benchmarks On Linux With NVIDIA Graphics

[faldzip]

It is all about making the good action of retributing ports and showing that Linux is already a profitable platform.

If you look short vision then you are right, better play on Windows or Wine. So that in 2 years there will be no more ports. Voilà.

But most people here watch beyond that and bet on those ports so that they will improve in technology and increase in marketshare.
We bet we will have real native games more and more every day, maybe we are dumb, but for now it is the only way.

What you say is true only for only a part of Linux users which are less then 1% of steam gamers. And do you know what a typical gamer (which is a Windows gamer) says about the Steam Machines and SteamOS? That it's the biggest fail of the century and there is no one who would like to use this crap. Why? Because the performance is shitty compared to Windows running PC. Gamers do not care about the drivers if they are opensource or closed source or the porting team was small or big, they care about performance. You want them to treat a Linux as gaming platform - show them that games are performing better or at least same as on Windows. And what they see now? 50% worse performance and they just laugh at it.

Not quantity but quality is a key here. If you want to make a linux port - then make it really good, otherwise just don't do it as this will just have quite opposite effect on linux gaming then you would expect.

And once again - look at the Talos Principle - there is literally one guy taking care of the engine an he already made the Vulkan version competitive to windows DX11 version, while he still claims that this is in the very early stage. And I did not have to think twice about it - just bought this game as I think he is doing the job right.

[juno]

Wait a sec.

Here are our initial benchmarks of Feral's port of Tomb Raider on Ubuntu Linux with using NVIDIA graphics.

Was anti-aliasing on during the benchmarks? If you had 4xSSAA enabled, the results wouldn't be too bad at all. It's in the "advanced" tab, so I'm not sure...

[Michael]

Just heard back from Feral... They will not be supporting the CLI benchmarking for Tomb Raider. Ah well...

[wdb974]

Then sell it for the 20-50% of Windows price as this is what we really get.

So, how do you propose the game be sold then? Only Linux users should benefit from such a sale.

And (continuing your thought) as the team is much smaller - they can make profit for a much lower price as the game was profitable for the much bigger Windows team.

See above.

Oh, and remind me how big is the Croteam's Engine Team?

I wasn't aware Tomb Raider was using the Serious Engine.

As I guess it's... one guy. And I can have 80% of the Windows DX11 performance with Vulkan on the AMD GPU-PRO beta driver.

That's great news. What are you doing for Vulkan to be adopted by more game developers?

Now the linux gaming seems not profittable for me at all, as I bought my GPU for like 100$ and I would have to change it to a 500$ GPU to get the similar performance while still paying the full price for the game. On the other hand - for like 100-200$ I can buy Windows and HDD, put it into my PC and have muuuuuuch more games with muuuuuuch less problems and muuuuuch better performance.

Cool story, bro. Now, you'll have to excuse me, as I need to leave. I feel the urge to make an account on xbox.com to tell everyone how superior the PS4 is. Looks like a productive afternoon is about to start!


---

And once again - look at the Talos Principle - there is literally one guy taking care of the engine an he already made the Vulkan version competitive to windows DX11 version, while he still claims that this is in the very early stage.

If that's true, we can only hope that more and more devs use Vulkan.

[faldzip]

I know that TR2013 is not using Seriuos Engine, I've mentioned Croteam for comparison, to just show you that the size of the team does not matter that much.

Guys, I do really care about Linux, I use it for everything except gaming and I really wait for the day that I would not have to have the Windows partition for gaming - but I think that poor performing ports would not get us any closer to that day. Really, take a look at some PC gaming/enthusiasts sites (so Windows centric) - they see the numbers and just laugh at how bad is linux for gaming and with such poor ports - they are just right.

If that's true, we can only hope that more and more devs use Vulkan.

Any idea how to convince them to use Vulkan? I have one: do not buy ports with DX-to-OpenGL-wrapper crap and leave the comment on Steam that you would buy it if it was good Vulkan port.

[pinguinpc]

In short test game seems stable and runs very good, cpu and gpu use is very high

However tessellation is a very heavy feature (without tressfx) reduce around 15fps and gpu temp is higher

[Passso]

I know that TR2013 is not using Seriuos Engine, I've mentioned Croteam for comparison, to just show you that the size of the team does not matter that much.

Guys, I do really care about Linux, I use it for everything except gaming and I really wait for the day that I would not have to have the Windows partition for gaming - but I think that poor performing ports would not get us any closer to that day. Really, take a look at some PC gaming/enthusiasts sites (so Windows centric) - they see the numbers and just laugh at how bad is linux for gaming and with such poor ports - they are just right.

They see the numbers and start masturbate because they are the majority, the same goes for PS4.

[mike4]

Benchmark with identical Ultimate settings gives 37 fps on Ubuntu and 52 on Win10. No visual differences.
Kind of surprised. I was first under the impression more cpu cores would help, but as I'm on a i7 that seems not valid.
So either they do something very wrong or use some wrapper or wine alike code.
Feral should fix performance, probably by moving to Vulkan asap. Else many customers will move back to Windows.

[Michael]

Why not use windows version with wine-staging 1.9.5 csmt enabled or wine nine.

Haven't seen anyone contribute a Windows TR2013 test profile on OpenBenchmarking.org.

[justmy2cents]

As you wish, go and pay some 20$ for some shitty port and do not forget to buy the 980ti and take your friends too, Nvidia is going to be very happy.

care to enlighten me? i have 750Ti and i have 88fps average at 1080p where my tv is 60Hz. how in the world would i benefit anything if i played on any other OS? would my tv suddenly became 200Hz? how does that classify as shitty port? some features i have disabled and i would have them anyway because i can't stand them, i would do that even if i had 10 extra Titans strapped with tape just because all my available SLI slots were already taken over by them
you have to take it with a grain of salt, but last thing you want is following any game benchmarks on phoronix. Michael is not a gamer, uses vanilla OS, doesn't tweak it even the least and then numbers can stray like they did on SoM where his results were less than half of people who reported their own. he simply cares about getting some numbers, not numbers.

and TressFX in TR is simply not worth. what is even the point of it, most of the time game is dark and practically only has effect on pony tail? if it had some better implementation, i wouldn't say. problem is that TR uses is beyond mediocre without any real benefit unless losing framerate is achievement for you