So how does AMD’s first example of RDNA stack up? For AMD and for consumers it’s much needed progress. To be sure, the Radeon RX 5700 series cards are not going to be Turing killers. But they are competitive in price, performance, and power consumption – the all-important trifecta that AMD has trailed NVIDIA at for too many years now.
AMD’s new graphics cards have already managed to do what has been sorely needed for a long time now: they’ve forced NVIDIA to lower prices and release new cards sooner than they otherwise would’ve.
This is why competition is so important.
That’s about the most shallow takeaway I can imagine…. at the very least it could have been now I can buy an AMD GPU that is actually delivers competitive FPS/watt and FPS/$ in the mid tiers, while also offering many new bells and whistles in the drivers.
Or, AMD finally delivers a fully competitive GPU, forcing NVidia to drop prices, but will consumers fall for that tactic or wise up and give the underdog a chance, while also benefiting from an enhanced driver platform, and a new enhanced architecture that will set the standard for the next 5-10 years.
Or AMD has done it once again, and revolutionized the graphics card architecture, leading with strong performance numbers, and delivering a heavy hit of open source support, the best in the industry. AMD delivers some of the most innovative and usable features we’ve seen in ages, meeting critical acclaim for their performant implementation of sharpening and upscaling, while Nvidia stumbles with often soft and blurry DLSS. AMD enables budget gamers to hit above their weight, by sharpening and upscaling subjectively as much as 20-30% with few discernible differences even more in face paced games like fortnight and overwatch. On the open source front AMD’s Navi is already showing it’s strength on Linux with performance numbers rivaling Nvidia’s best in games. On the enthusiast front AMD has left powerplay tables in tact ripe for the picking, and early results indicate the 5700xt can hit an entire class above when liquid cooled at around 20% overclock and 250w power consumption.
If you aren’t going to bother to at least write your own articles quite wasting our time with reposts of anandtech…. everyone has already seen it 4 times on reddit.
RDNA is a good start. It is competitive on power because of 7nm and competitive on performance because it is a good architecture. You can play any game at 1440P at 60FPS on a $350 card. Hopefully AMD will improve RDNA year by year just as they have done with Ryzen.
Bravo!!!
You are so right!
One of those GPUs are my next ones. Besides now being competitive in the gaming performance sector, AMD has been open to undeterred compute applications, anywhere from GPU compression speedups to easy access on non windows operating system via drivers, without being a proprietary addon blob.
I’m done with the nVidia premature abandonment of features as well, like physx, which is now being integrated into their proprietary Gameworks, or ending the physx secondary card benefits.
Poseidon,
Yeah, nvidia’s proprietary code combined with restrictive license terms are just terrible for compute. AMD is better for open source and priced more competitively, but unfortunately the performance per time & per watt lag nvidia. The review didn’t paint AMD’s compute readiness in a great light either.
From the article…
For gaming, i’d likely support AMD. For compute though AMD’s deficiencies are disappointing. I hate proprietary stuff, but nvidia still has a lead in compute, so it was hard for me to choose. Too bad you can’t have everything and we’re always forced to compromise on something or other.