Return to

The VEGA 56 / 64 Cards Thread! General Discussion



Doesn’t matter.


Depends on use of CUDA. Also then the FE Vega or one of the other pro cards is probably a better idea

Sorry, Vega 56 makes more sense but I don’t think anyone should buy a Vega 64.


It actually does, for games on linux the Vega cards are faster than anything nvidia has to offer.

For production it depends on the software, yes. That would have been the follow up.
Also @Leon_the_Loner wasn’t specifically naming Vega64, sooo …


Would really like to see benches to back that up but considering the linux gaming population is so tiny anyway it is an irrelevant factor for me. Would also have to be a hell of a lot faster to make the extra power consumption, heat and noise worth it.

In another thread he did. Sorry you wouldn’t have known that tho


That’s not a thing


Then why do I have 300 of those?


that being said im not ever going to buy vega as it stands currently.
i might even look at getting the 1070 Ti when that is a thing lol. If i can snag one before they inevitably sell out.


It’s a trend. You are very much the minority.

But ummm back to this…

So where is it faster exactly? I’m trying to find some more benches but having trouble? Any examples. Because AMD’s OpenGL performance is pretty bad generally so… Yeah really need to see some examples of where a Vega 64 is faster than a 1080 Ti. Hell even faster than a 1080…

Yes yes we know you can technically game on linux and people do. You also know that isn’t his point at all. That list is laughable and the amount of people actually gaming on linux is tiny. It is extremely niche. I guess because we are on a more enthusiast focused forum people don’t realize that but even here very few game on Linux

But that really isn’t the point of any of this. My point was Vega 64 doesn’t make sense to buy even if Linux somehow did make a significant difference (which it doesn’t).


1080 (or maybe ti) here I come… well in a couple of months anyways


Subpar trolling attempt. Go buy some DLC or somethin’ …

You might want to look at the other charts on phoronix, you know, where you got those from yourself.

Correct, the point was that I asked someone else a question.

Doesn’t mean I’m alone. Which is why I asked @Leon_the_Loner for his usage. (As in not you.)


I’m really anal about being able to run opensource as much as possible. That’s why I’m running proxmox instead of ESXi. Nvidia is king, but the opensource drivers on the radeon side are so much better than Nvidia has. Hell, they’re better than proprietary Radeon.

Also freesync is a fuck ton cheaper.

I know exactly what I’m walking into, huge power draw, hot, and loud. but that’s no different then my 2 290s, and I’m probably drawing more power than that now. But it does allow me the performance to play with freesync at the resolutions I want. With open source drivers and a cheaper over all rig, all be it not as powerful. But if I can hit 120hz on 1440p with adaptive sync on my competitive titles, give me the best open source driver support with any real performance on linux and still chug out 75hz on the big titles on windows I’m down.

Really done with Nvidia’s software too.

Edit snipe: AMD GPUs have always aged more gracefully as well. and AMD also has an opensource software suite for devs. So I’ll support opensource where ever possible.


I’m gonna mod my card with a morpheus II air cooler and will show the result here but that is still a week away or so. It is another 65,- bucks on top of the already expensive card for the cooler plus whatever fans you want to use. It won’t reduce the power hunger but the hot & loud part might be manageable that way. I bought the sapphire reference card which was a bit of a mistake, as far as I understand it today XFX are the only ones to still give you warranty even when changing coolers. So if this method works that might be the way for you.


I’ll probably do that too then. I gotta be able to post at sapphire that I bought their card for supporting the lan party I go too.


Yeah I did. There were 4 other charts I didn’t include. Why?

Well one was Bioshock Infinite an old game with relatively little players still that was weirdly tested on Medium settings. Vega was faster there. Seems like an outlier

The next was Dawn of War III. Vega lost that one too. To the 1080.

Next was Xonotic. A game I never heard of where the Vega 64 did win against the 1080 but lost against the 1080 Ti.

Last were a series of Unigine canned benches. The nvidia cards absolutely demolished Vega in every single one

There were also a few Vulkan tests where Vega also got completely destroyed so Vulkan isn’t saving it either.

EDIT: Made a mistake. AMD was Opensource. nVidia was tested with proprietary drivers. Still don’t think it changes the conclusion unless you reallly wanted to use open source drivers. However, even then, the performance difference would need to be considerable in order to justify the cost and power consumption.

No offense man but if you’re gonna make grandiose claims like “On Linux Vega is faster than any nVidia graphics card” the burden of proof is on you. Especially since, as far as I can tell, that is completely false…

If there are reasons you like it and you want to justify your purchase that’s cool but don’t make BS claims like that.


i dont really game on linux though tbh lol, just happen to have above mentioned games that work on linux… so there technically is gaming but really there is at best limited selection.
maybe im weird for instead of dual booting… having two systems, one that i use for most things and occassionally gaming (linux) and my other machine for gaming (windows)
in any case i find it amusing you picked that to comment on instead of the other thing i said but whatever … was agreeing with you lol

unless vega is actually at MSRP it doesnt make sense to buy any of them (even then, only kinda?)
well, FE vega makes sense but thats not for gaming really lol. It’s for workstations.


This almost certainly were not open source tests as nouveau is trash on current hardware. Can you link the specific article you were referencing?


Edit. I was wrong. I apologize.

AMD was running Open Source but nVidia was proprietary.

Will change that post. However, I still don’t think it changes the conclusion that, outside a few very specific cases, Linux doesn’t make Vega faster than any nVidia GPU…

I’m done debating this though. Do what y’all want.


I can.

“All of the NVIDIA hardware tested was using the 384.59 proprietary Linux driver atop the Linux 4.13 Git kernel.”

Look, if the closed nvidia stuff is in some cases outperformed already, that is only gonna get worse for green cards. Nvidia doesn’t give a shit about optimizing drivers on linux, AMD has now open source drivers which will become better and better.

Also, as I said, for me this was just about helping someone to not fall for a possibly very bad advice.

Maybe we can get around a mess like this next time?

One more thing, @wendell mentioned that Vega64 was faster on linux than a 1080Ti on a couple occasions. I simply took that as confirmation of what I had already read on phoronix. At some point we will hopefully get a complete Vega on linux video, then we know for sure.


I bought a Vega card because nVidia locks you down harder than a prison cell during a riot. Completely boring and un-fun, there’s nothing to do with the damn thing (last card was a 1070 so current gen).

Plus that awful software. Somehow the driver made all my Hangouts text bold, and nothing could undo it. As soon as I ran DDU, all back to normal.

Now, on with Blender HBCC testing before hitting the gym :grin:


I can lend you one… lol

You can tell them they are assholes for not warrantying a waterblocked card too!

For what I do, it is a great choice:

Hence why nvidia gave the Titan Xp the driver update which gave it up to 3 times performance gains in workstation applications.


but why did you not get FE then… and get the full workstation feature set support…