NVIDIA GeForce RTX 2070 Linux Gaming Benchmarks

[pcxmac]

To be fair, even the 3rd gen GCN is nearing 2 years old. Meanwhile, Vega is actually pretty power efficient in the APUs and for certain compute tasks. That being said, Vega seems to scale up poorly, which is why the 56 was an overall better choice than the 64, but neither are especially efficient in general.

considering AMD has been using GloFo's low power 14nm node, maybe not so bad. Should be interesting to see what AMD can do with 7nm and a node that is more optimized high power applications. If they can hit over 2Ghz, they should be able to eat in to NVIDIA's margins like they started with Ryzen, 1st edition. With more investment they might be able to catch up to NVIDIA if they can break their GPUs in to small parts and scale out, and even find a way to use RAM and multiple card's RAM better. 5/3 nm seems to be the next barrier, finding ways to make smaller dies and scale out is a great way to find performance in the mean time. Imagine a GPU with like 6 or 7 small chiplets and data pathways direct to other cards and other cards memory.

[theriddick]

Once the Mini-ITX version is out, mail me.

Won't take long, Zotac are normally on top of Mini-ITX variants for nvidia cards, its a real shame they can't do the same magic for AMD cards!

Maybe Navi will change things around for AMD, we will see,

[bug77]

Benchmarking games on Linux will always be pointless. You never know if it's the hardware pulling you back or the countless layers of buggy/incomplete software that runs the whole thing.

True dat. Instead of showing you clearly what is holding back your system, benchmarking games on Linux only shows how well you can actually play them. Why does Michael even other?

[anarki2]

Benchmarking games on Linux will always be pointless. You never know if it's the hardware pulling you back or the countless layers of buggy/incomplete software that runs the whole thing.

[dimko]

With current prices, there is no justification for me to upgrade from 780TI, which cost me almost 400 euro to something significantly better, as its gonna be somewhat better, but not THAT significantly better. And that card is like 3-4 years of age card. Fuck this bullshit when comes to Video card prices. I hope more game developers user Vulkan so we can squeeze GPU power.

[ruthan]

It's not missing. NVIDIA is no longer manufacturing 9xx parts so you can't get them new/retail unless buying used / refurbished / maybe random third parties perhaps that might still have stock.

I know, but its stop producing them is Nvidia thing, not consumer thing, reviews should be done from consumer view. Some shops still have them, or its not hard to use / calculate refurbished price.

[Michael]

In Performance per dollar are missing Geforce 9xx cards, otherwise nice test - good cards set, games set, there is performance per dollar chart.

It's not missing. NVIDIA is no longer manufacturing 9xx parts so you can't get them new/retail unless buying used / refurbished / maybe random third parties perhaps that might still have stock.

[ruthan]

In Performance per dollar are missing Geforce 9xx cards, otherwise nice test - good cards set, games set, there is performance per dollar chart.

[artivision]

Once the Mini-ITX version is out, mail me.

[Azpegath]

I'm quite uneducated in the performance test department, at least with regards to Windows testing sites, but is AMD Vega XX under-performing on Linux, or are they just not very competetive at all?
I checked the prices in Sweden and I can get the cheapest 2070 for about $100 USD less than the cheapest Vega 64, but the card is performing better in every of these tests.

So is it the hardware, the driver or the games, or all of the above?