Radeon Windows 10 vs. Linux RadeonSI/RADV Gaming Performance

[DanT]

PTS only ran a few of the tests in the article since all the others had to be done manually due to limitations of the games.

Yes, I understand. However, the orientation is sometimes difficult even when comparing linux-only drivers (sometimes, there is "renderer: ..." in the header, but sometimes not). Anyway, I don't know the PTS codebase and I have no idea how difficult it would be to add such an info. I just thought, it would be nice to have it :-)

[Solid State Brain]

It's about always using the same hardware, even the same hard drive, throughout all of the tests. And so no, I don't have an extra 950 PRO NVMe SSD... I wish I did. And last I knew, Windows wasn't so friendly when swapping out an installation between systems and booting it up on a completely different system, has anything changed with Windows since?

I recall too that past Windows editions had problems when transferring an installation to a radically different system, but I've just tried cloning my current Windows 10 installation (Using Macrium Reflect free for Windows) from my system (Intel i5 3550, H77 mainboard, 16GB DDR3, 3 SSD, RX480) to an older Intel Core 2 Duo/G45-based system and after "getting devices ready" for a while during the first boot it seemed to operate just fine. So I guess that long-standing problems on this regard have been solved or strongly mitigated. I don't have unusual configurations (NVMe storage, RAID, etc) to test, however.

[baka0815]

Thanks Michael for the tests and thanks to all amdgpu/radeonsi contributors, you're doing a great job!

[duby229]

so you are saying that because there is no still no performant open source vulkan driver for AMD that game developers cannot target vulkan? I guess AMDGPU pro is not a viable alternative as the distro and kernel support is so limited, although it is way faster than RadV.

If you are right and AMD's open sourcing of vulkan driver is not gonna happen anytime soon then maybe that's why Valve is so eager to improve RadV.

That's not what I said. In fact RadV is performing nicely and its development pace has been impressive. But regardless of that it's not in a shape where games can be deployed commercially. Also please try to understand that AMD recommends to gamers that they use the latest OSS stack available to them, which means that RadV will need to be completed in time for Vulkan game releases.

It's because of these circumstances that Valve has a vested interest in contributing, and AMD has a vested interest in letting them.

[kirgahn]

You've said this before.
Is your reasoning that you expect new subscribers to pay for new articles? If so, that suggests no consideration is given to readers once they've subscribed.

Yo liam, chill dude. I think Michael is just disappointed and being honest about it. I concede he should've been more cautious with wording, but it really is not a matter of "not caring about existing subsribers" rather than a venting a little. The guy runs a small data center at home and buys most of the hardware himself.
Plus he works his ass off 24/7. Give him a break.

[kirgahn]

It's about always using the same hardware, even the same hard drive, throughout all of the tests. And so no, I don't have an extra 950 PRO NVMe SSD... I wish I did. And last I knew, Windows wasn't so friendly when swapping out an installation between systems and booting it up on a completely different system, has anything changed with Windows since?

well, I'm not up-to-date with windows 10 but sysprep was the tool to go to when moving an existing windows installation to other hardware. You would need to re-activate it, but that shouldn't be an issue.

[rycodge]

RX 480 owner here on Ubuntu 16.04, kernel 4.8 using the Padoka PPA. Rocket League is performing amazingly well after they implemented the fixes. No stuttering, butter smooth menus. The framerate was always rock solid on amdgpu but now everything just works as expected.

With the recent fixes I need to test Rocket League on my HTPC with an old Core 2 Quad 6600 and 7870 graphics. I'm sure the 7870 will pull through, its just the slower system was rendered almost unusable.

[Herem]

Michael as a premium subscriber these are exactly the sort of tests I subsided for, great work.

In future rather than running such a broad selection of 10+ games in a single article, have you considered limiting each article to a single title presenting both AMD / Nvidia results on Linux and Windows? If you were to provide the results with some nice graphs, you could spend a lot less time testing, use a little more time for analysis but overall significantly reduce your workload per article. The other advantage would be you'd have the scope for a lot more articles without repeating content too often.

Most other hardware sites tend to use this approach. They focus most benchmark articles on a single title, reserving mass analysis across an entire benchmark suite only for new GPU and CPU reviews.

[liam]

Yo liam, chill dude. I think Michael is just disappointed and being honest about it. I concede he should've been more cautious with wording, but it really is not a matter of "not caring about existing subsribers" rather than a venting a little. The guy runs a small data center at home and buys most of the hardware himself.
Plus he works his ass off 24/7. Give him a break.

I completely understand that Michael is frustrated. A few months ago this came up, and he was saying how burned out he was. Plenty of us made suggestions which might've helped (though, I'd imagine, none of the suggestions were new to Michael). So, please don't think that I'm unaware of why he's saying such things. I've been here awhile, after all
I only pointed this particular instance out because it seems to be happening more frequently (over the last couple of weeks, i believe), and his recent phrasing has really seemed like an fu to current subscribers.

[liam]

I think it's the other way round... writing articles aimed at existing subscribers but hoping they will attract new subscribers... all the while keeping an eye on what actually happens.

Well, i had a fairly lengthy reply but decided to abandon it since i didn't think it adequately (and accurately) expressed my thoughts.
i understand that language can be ambiguous, but, unfortunately, I'm not always able to dispassionately consume what i perceive to be its meaning (even when understanding that it's only one possible interpretation).
So, yes, what you say it's absolutely possible.