NVIDIA R565 vs. Linux 6.13 + Mesa 25.0 Git AMD / Intel Graphics For Linux Gaming

[jojo7887]

There's something odd on the GTA V test, I mean a 7800XT shouldn't get only a mere 35 fps at 1440p.

[skeevy420]

Michael
I'd love if you could expand your lineup of steam/proton games. I like seeing the benchmarks about gaming on linux. It would be nice if you could integrate additional titles over time.

For some reason it seems like a lot of those are older, but in reality I'd only count yQuake2, GTAV and maybe Xonotic and Unvanquished(Engines) older. The rest are fairly recent. Maybe because We see them often to compare distro/kernel improvements.

So would a lot of us. Not all games have a benchmark mode to make it automated and, like Michael said in different words, it can be a pain in the ass to automate all the inputs necessary to create our own benchmarks. Even if we managed to do that we can't always ensure that in-game lighting, NPC locations, etc will be the same from run to run.

I don't think any of that really matters that much. More titles and more in-depth testing would only make Phoronix's Linux game benchmarking a fuller time job than it already is. It's a lot for one person, logistically. A 5 minute benchmark becomes 55 minutes of raw testing due to the 11 GPUs being tested. It adds up fast. Take a nap and if a run finishes your tests are behind that many hours because someone has to physically swap GPUs. As much as I'd have liked to see medium, low, and 1080p tests, I understand why they're not there. Medium and Low add two scenarios per test and 1080p increases all the testing by a factor. That's a lot of extra work.

[Dukenukemx]

Looks like Intel is off the table.

[pyrex777]

This test will not tell us that because it was only 4 games so you can nearly only cherry pick AMD won at least in 2 titles of 4 or 5 with a 470 Euro in Germany card against a 530 Euro Nvidia card often even against the Super that is over 600 Euro.

Also it seems that there is used the stupid native mode, F1 22 is also available as Windows version so obviously there is a huge bug, the Nvidia hardware is not like 3x as fast as the other games show, so like most Linux versions of games (like Ark Survival Evolved) is horrible buggy.

So why would you limit yourself to use the garbage native version of the game instead of the 99% sure much better running proton version?

And there are not worse result for AMD because of the bad Hardware but because Phoronix compared entry level AMD cards with mid level Nvidia cards... the cheapest 4070 ti super costs here 820 Euro a 7900 XT only costs 670 Euro So obvious if you compare more expensive and therefor highend cards against cheaper therefor mid / low end cards of amd amd looses...

What matters is frames per buck...

you and Leopard, are you being seious right now? just because Michael didnt have a 7900 xtx at hand doesnt make this test unfair. all you have to do is literally compare gpus with the same price. 7800 xt, vs 4070 (which, vram aside, is still faster while consuming 50W less. maybe 4070 super. why are you looking at the 4070 ti super? you're trying to pass off as objective when your bias seems to be clear as water. This is not a marketing slide from Nvidia, this is a test from Michael for linux users who in theory should know better than the average consumer. he probably never even thought this would be used as an argument against the test.

RDNA3 has been pretty bad, there is no way around it. AMD on the software side is stuck in 2017 while Nvidia has lossless magic that adds +20-30% for free and RT x-times faster than AMD.
Even Intel in the span of 1 generation has caught up to DLSS with XeSS 2 and XeLL. If you've bought RDNA3 because of better Linux compat/a good deal, that's a valid reason, but strictly gaming speaking it's not close. plus, they are solid cards for casual compute work thanks to cuda.

[pyrex777]

I very much regret buying an Arc A770. I thought Intel's Linux drivers would mature, but still many years afterwards it has not been the case.

I should have gone with AMD in retrospect; or even Nvidia considering they have started open-sourcing their drivers.

Yeah unfortunately some issues just can't be fixed with drivers. eg, the high idle power consumption which required a BIOS setting that sometimes didnt even work. but Battlemage is looking good.

[Quackdoc]

This is more or less what I would expect for these titles, the Alchemist cards are physically good cards, but well, they are still a first try gaming card with a lot of driver issues. There are other titles which do have a lot better performance, but I know none that are easily automatable for benching.

you could use screen recording based automation, where "if selected area has match, perform x input" that would at least get you past loading screens and menus.

[ET3D]

MichaelI'd love if you could expand your lineup of steam/proton games.

I agree that the benchmarks feel rather pointless without this. My bet is that anyone who wants to seriously game on Linux will use Proton, and not limit themselves to a few benchmarking-friendly native Linux titles. I don't really expect Michael to make Phoronix into a gaming benchmark site, but as is I feel that the results aren't that meaningful.

I very much regret buying an Arc A770. I thought Intel's Linux drivers would mature, but still many years afterwards it has not been the case.

The A770 was released 2 years ago, which is as few years as "years" can refer to.

[PapagaioPB]

Great benchmark, but I would like to see the RX 6600 and RTX 3060 included, because Intel's cards actually compete with these. Alternatively, when the B580 arrives, we can see how it performs against the RTX 4060 and RX 7600.

One thing to take into account is that in more recent games, the 8GB on these two cards is an extreme limiting factor for gaming; some games even crash with 8GB or less.

Looks like Intel drivers are only good for benchmark engines. Actual gaming tests show a very bad situation for those interested in gaming on Intel GPUs.

However, an interesting point is that in performance for workloads outside of gaming, these cards are actually top-of-the-line.

[Quackdoc]

I agree that the benchmarks feel rather pointless without this. My bet is that anyone who wants to seriously game on Linux will use Proton, and not limit themselves to a few benchmarking-friendly native Linux titles. I don't really expect Michael to make Phoronix into a gaming benchmark site, but as is I feel that the results aren't that meaningful.

the large issue is how bad modern games are for automated benchmarking. even if you get mangohud or dxvk hud or whatever to dump frametimes, many games you cannot just automate with inputs because of

A) The games are Non deterministic (RNG is a bitch for automated benchmarking)
B) Have a myriad of loadtime issues
C) Input emulation can be hard as it may need uinput for a generic solution

etc.

Some programs may be "TAS friendly" and those could be benchmarked, but finding them is hard.

[ET3D]

the large issue is how bad modern games are for automated benchmarking.

I understood this, and I think it was clear from my message. However, this doesn't invalidate the point that these benchmarks aren't really useful for gamers. The question is whether there's any real point in running these benchmarks. In some sense, sure, always nice to see some numbers. But I can't tell if they reflect actual gaming scenarios in any way, which makes them academic at best.