Upgrade to Radeon Vega 56 or RX 580

[cruuds3027]

Hi all,

I was just looking for some info or advice on my next GPU I would buy. I have done some research and have come down to either the Vega 56 or the RX 580. However I am trying to pay as least as possible, but if my choice would be better to go with the Vega then I will. I currently have a Zotac GTX 1060 6GB AMP Edition, so I guess my question is, what would be a better option from what I have now? Obviously the Vega is a substantially better graphics card than the 580 and my current one. But if I were to get the 580, would the changes be significant enough where I will be more satisfied compared to what I have now, or would I be better off spending the extra money and get the Vega?

Thanks for any help!

 

[2sidedpolygon]

Why would you get AMD cards over Nvidia ones? Especially the 580, that performs worse than the 1060.

 

[4745454b]

I wouldn't say worse. They perform very similar however. If I had the 1060 already I wouldn't bother sidegrading to the 580. Worse, the power consumption for both cards is more than the 1060. To be honest if those are your choices I'd keep the 1060.

 

[2sidedpolygon]

I wouldn't say worse. They perform very similar however. If I had the 1060 already I wouldn't bother sidegrading to the 580. Worse, the power consumption for both cards is more than the 1060. To be honest if those are your choices I'd keep the 1060.

The 1060 performs slightly better in most games, but yeah, they are equivalent for the most part.

 

[cryoburner]

Why would you get AMD cards over Nvidia ones? Especially the 580, that performs worse than the 1060.

The 1060 performs slightly better in most games, but yeah, they are equivalent for the most part.

That's not actually true. In practically all reviews I've seen, it's the RX 580 that on average comes out slightly on top, particularly in newer games. Perhaps you're thinking of the RX 480. See this recent review, for example, where they average the performance of 22 recent games with both a stock and factory overclocked RX 580...

https://www.techpowerup.com/reviews/MSI/RX_580_Mech_2/31.html

However, the differences are so minor, that it would definitely be a sidegrade, and any framerate differences would be imperceptible. Plus, the RX 580 is more power hungry, which would result in more heat output, so the "upgrade" would not be worthwhile.

Vega 56 would actually be a decent amount faster. However, Nvidia has a new generation of graphics cards that will be rolling out over the coming months, and AMD is rumored to have new cards coming soon as well, so it might be worth waiting to see what those have to offer within this price range.

 

[2sidedpolygon]

Why would you get AMD cards over Nvidia ones? Especially the 580, that performs worse than the 1060.

The 1060 performs slightly better in most games, but yeah, they are equivalent for the most part.

That's not actually true. In practically all reviews I've seen, it's the RX 580 that on average comes out slightly on top, particularly in newer games. Perhaps you're thinking of the RX 480. See this recent review, for example, where they average the performance of 22 recent games with both a stock and factory overclocked RX 580...

https://www.techpowerup.com/reviews/MSI/RX_580_Mech_2/31.html

However, the differences are so minor, that it would definitely be a sidegrade, and any framerate differences would be imperceptible. Plus, the RX 580 is more power hungry, which would result in more heat output, so the "upgrade" would not be worthwhile.

Vega 56 would actually be a decent amount faster. However, Nvidia has a new generation of graphics cards that will be rolling out over the coming months, and AMD is rumored to have new cards coming soon as well, so it might be worth waiting to see what those have to offer within this price range.

I'm not thinking of the 480. I got that from Hardware Unboxed's 27 game benchmark of the cards. Though, for the most part, it still depends on what games you're playing, it still doesn't make sense to buy an entirely new card.

As for Vega being faster, yes, it would be faster than the 1060, but not another Nvidia card like the 1070ti, as shown in this benchmark. As for the newer generation of Nvidia cards, 20-series is a wash in my opinion. There's literally no reason to get one unless you really care about ray tracing at sub-60 1080p with narrow support or AI antialiasing with little-to-no actual implementation in games. Or if you just want the best of the best with the 2080ti. Way too pricey to be near worth it, in any case.

Side note: AMD cards also have significantly higher power consumption

 

[4745454b]

I don't have your Nvidia goggles on.

https://www.patreon.com/posts/radeon-rx-580-vs-19635434

This is a graph that I got from your link. I'm not listening to someone ramble about stuff and make it hard to see. For every game the 580 loses in like Assassin's Creed: Origins, there is another game like BF1 where it ties or Dirt 4 where it wins. Even the first comment on that page reads in part, "these cards really are neck and neck."

What I hope all this shows the OP is that if you have the 1060, the 580 isn't an upgrade. You'll get freesync if that interests you but that's about it. Vega would be faster, but not by tons. Unless you are going to pony up a lot more money for the card you already have a good one.

 

[2sidedpolygon]

I don't have your Nvidia goggles on.

https://www.patreon.com/posts/radeon-rx-580-vs-19635434

This is a graph that I got from your link. I'm not listening to someone ramble about stuff and make it hard to see. For every game the 580 loses in like Assassin's Creed: Origins, there is another game like BF1 where it ties or Dirt 4 where it wins. Even the first comment on that page reads in part, "these cards really are neck and neck."

What I hope all this shows the OP is that if you have the 1060, the 580 isn't an upgrade. You'll get freesync if that interests you but that's about it. Vega would be faster, but not by tons. Unless you are going to pony up a lot more money for the card you already have a good one.

Yes, but if you did actually watch the video, you would find that in the conclusion he says that the 1060 is faster on average (If only by a 3% margin). They are neck and neck, but the 1060 is slightly better and it has superior power consumption to the 580 by a significant margin. I'm not wearing Nvidia glasses, these are benchmarks. Not to mention, like you said, even if the 580 was the 3% better card, upgrading it would be a bad idea. It's simply not worth it just to get Freesync, as nice as it is.

 

[4745454b]

3%? "If I actually watched the video"? You are going to argue over 3%? That's something you will never notice or feel. I'm moving on. 3%....

 

[cryoburner]

Yes, but if you did actually watch the video, you would find that in the conclusion he says that the 1060 is faster on average (If only by a 3% margin). They are neck and neck, but the 1060 is slightly better and it has superior power consumption to the 580 by a significant margin. I'm not wearing Nvidia glasses, these are benchmarks. Not to mention, like you said, even if the 580 was the 3% better card, upgrading it would be a bad idea. It's simply not worth it just to get Freesync, as nice as it is.

Having watched the video, they noted that the RX 580 actually performed 3% better than the GTX 1060 in their previous roundup, which included a mostly-different selection of games. So, a lot of this comes down to which particular games they decided to test each time. And definitely worth noting, they highlighted at the end that one particular game (Frostpunk) was an outlier that threw off the averages, and without that game included in the results, there was only a 1% average performance difference across the remaining titles, which is definitely within the margin of error. Some games may perform 10% or 20% faster on one card or another due to the different architectures, but overall, the two cards off very similar levels of performance. In the video, they actually said that it's "just too close to call" and to "get whichever one is cheapest".

Another thing to note is that they manually overclocked each of those cards beyond the manufacturer's default clocks, and when it comes to overclocking, there's some luck of the draw involved, as some individual cards will tend to overclock better than others, which could have also potentially played a role, albeit a minor one.

In any case, yep, it definitely wouldn't be worthwhile to switch from a GTX 1060 6GB to an RX 580, or vice verse, as they are just too similar in performance.