At least the Phoronix review is positive, if we can actually get these i’m seriously considering this.
I think stuff will get a bit better when the drivers have a moment to mature.
I would also like to see some vulkan tests.
I think some of the reviewers forget that traditionally… AMD parts go on sale more often than others do. (making for a better buy IMO)
And paul failed to point out the performance difference between dx11 and dx12. I’m not a Linus fan but he did manage to point it out.
Any idea if amd cards also generally do better then nvidia on dxvk?
Cant say i remember anyone doing performance testing with dxvk b/t amd & nvidia…
Would be cool if @wendell has a VII and could do it for us.
i would love that, but if anyone reading this also has some dxvk nvidia vs amd numbers, il take that aswell
well, i just watched pauls video about the radeon 7, and it looks like ill be in the “waiting for navi” line. i was hoping to get 2 of these cards, and crossfire them, but according to him, they dont support crossfire. which would make getting 2 of them pointless. oh well, i suppose im alright waiting a bit longer for AMDs absolute next gen.
As with all things when not equal. Read the reviews and measure to your work loads. The Radeon Vii kills opencl. Loses mostly to the 2080 and has none of the stupid nvidia shit no games has YET.
I would say AMD botched the drivers. No one is overclocking stable and some games plain are troublesome.
So its again a waiting game.
P.S If a lot of tech reviewers are leaving out 4K Benchmarks. There shilling like villains. Its kind of important for a 16G card with the fastest vram in history ? Don’t you think ? Win or lose.
I hope @wendell might get his hands on one and then see how it behaves with QEMU/KVM and if the reset bug is still there.
ty for the linux link, i would not have found that on my own. It pretty much seals the deal for me if this thing has sr-iov.
Seems like a good buy either way, However i do not belive i need such a powerful card. So if i end up with two cards i would buy two cheaper cards. And then just upgrade in a few years.
Or a used card :<
I have noticed this a lot recently, never tied to one resolution or metric but defi itely missing tests that should be included, I know testing takes a lot of time but, puts on the krinkly hat, it is looking like reviewers are picking and choosing their benchmarks.
Sometimes resolutions, present for one fame not for others in the same suite.
Some times renderer, DX11, 12, OpenGL, Vulkan rarely a direct comparison game to game and often just one used showing all sort of different performance when comparing games on different renderers.
Sometimes in game settings, some tested with AA, others with out. Some with the likes of vendor specific things like FXAA when there are other potentially fairer options, not to mention better looking and performing.
It is a shame.
i would not make such a call before we know if it does sr-iov. i doubt it but omg i want it to
When they will fix overclocking I will be really curious how much more will we get from water cooling. Der8auer got his up to 2150Mhz on dry ice:
Hey! My response was clever dammit
As we have this thread now: