ASRock RX 590 Phantom Gaming X 8 GB

[cucker tarlson]

https://tpucdn.com/reviews/ASRock/RX_590_Phantom_Gaming_X/images/hellblade_1920-1080.png
https://www.techpowerup.com/reviews/Performance_Analysis/Hellblade_Senuas_Sacrifice/4.html

RX580 got 4 fps slower than 1,5 years ago?

not uncommon for the amd driver to improve one thing and gimp two in process
https://www.tomshardware.com/reviews/amd-nvidia-driver-updates-performance-tested,5707.html

This is the 590 i would get, if i was going to buy one. Then i get matching GPU and Mobo, but interestingly Asrock + Sapphire is also special as my Motherboard POST logo says ASROCK|SAPPHIRE when i have my 570- installed and i get a special "Sapphire boost "button XD havent pushed it yet...

Anyway the 1660ti is flat out a better buy, this needs to be $199 to make sense imo.

neither buying pascal nor polaris makes sense now,unless heavily discounted.this card at $250 is more like a wtf moment when you consider it pulls more than 1080Ti and you've got 1660Ti at $280.That's 20% more performance and half the power draw for $30.But yeah,it's got a game bundle they nailed it with the non-RT turing dies,should've done the same with bigger ones. The architecture is delivering a big performance improvement over pascal,but the xx80+ cards don't look as impressive as rtx2060 and down,12nm is holding them back a lot imo.Die shrink is badly needed for the upper mainstream and high end cards.

 

[ArbitraryAffection]

not uncommon for the amd driver to improve one thing and gimp two in process
https://www.tomshardware.com/reviews/amd-nvidia-driver-updates-performance-tested,5707.html


neither buying pascal nor polaris makes sense now,unless heavily discounted.this card at $250 is more like a wtf moment when you consider it pulls more than 1080Ti and you've got 1660Ti at $280.That's 20% more performance and half the power draw for $30.But yeah,it's got a game bundle

Yep true, I mean I would only buy if really cheap like 150 quid. Even then only to play with 12nm Polaris, test efficiency at lower voltages etc. Just from curiosity^^

 

[EatingDirt]

it'll be replaced by another title in the next round of rebenches

All we need is from gamedevs to stop using those engines, so that new games don't use them, so that I don't have to use them in my benchmarks.

Don't get me wrong, UE4 engine games should absolutely be included in the benchmarks because of it's popularity. My problem with DQXI is that it has always been an anomaly in terms of performance, even for a UE4 game.

 

[Ruru]

IMO this RX 480 rev3 is just so pointless. A rebranded card which is already rebranded, just with higher clocks. 1060 was already a better choice, and now when there's 1660 Ti (and 1660 coming), this is just a pointless card.

 

[cucker tarlson]

IMO this RX 480 rev3 is just so pointless. A rebranded card which is already rebranded, just with higher clocks. 1060 was already a better choice, and now when there's 1660 Ti (and 1660 coming), this is just a pointless card.

I think it depends on the choice of games,but overall it's true.
I just did a quick research,I chose games from the review that I haven't played but i'd be interested in.I obviously picked the from the list before I could see how they performed,even though I certainly had some idea of how 2-3 of them would do in terms of amd/nvidia optimization.I knew bf5 and strange brigade would favor amd (that turned out to have a huge impact on the end result),and I knew hellblade would favor nvidia. Despite there's big swings in optimization for both camps,the cards turned out dead equal in the end,both delivering the avg. of 64 fps in 8 games.There was only one instance of one of them delivering a non--playable experience at 1080p Ultra,and that was rx580 in hitman 2 at 38 fps avg.
That said,I can see how people can swing it both in favor of amd and nvidia based on the choice of games.Anyone claming rx580 is the king of midrange is a fanboy,so is the person claiming 1060 is a hands down better choice.TBH I see difference between the cards when it comes to performance numbers,but neither is really that much better than the other .1060 overclocks better and is less of hassle when it comes to noise,rx580 has the advantage of 8 gigs of vram but the card is gonna be noisy and hot unless you pick one with a quality cooler, on a 1060 the basic aib version will deliver more than decent results.
At this point I'd pick up neither of them,as both deliver sub-60 framerate at 1080p.In case of rx580 that was in 4 games out of 8 (!) I picked,the avg. result is due to rx580 delivering great framerate in 2 games where it pulled far ahead of the 1060 (the one I mentioned at the beginning).It's definitely the more inconsistent of the two and that's why I'd avoid it given how many various games I buy myself.1060 still delivered sub-60 avg. framerate in 2 out of 8,which is still a problem given how in order to stay above 60 at all times you probably need to see an avg. of +70 fps in a benchmark.
I'd still say 1060 wins this round,but that doesn't make it a definite winner.

 

[hat]

IMO this RX 480 rev3 is just so pointless. A rebranded card which is already rebranded, just with higher clocks. 1060 was already a better choice, and now when there's 1660 Ti (and 1660 coming), this is just a pointless card.

Overclocked rebrand of an overclocked rebrand...

Well, that's fine, but I think it needs to be a bit cheaper. The 1660Ti is 20% faster on average, though it does cost $40 more. I'd probably pay the difference and give up the games, though. The "free games" argument holds no water, for me. I don't buy a new graphics card to perform poorly in games!

 

[SIGSEGV]

All we need is from gamedevs to stop using those engines, so that new games don't use them, so that I don't have to use them in my benchmarks.

sad but true
Nvidia has an absolute power to claim all the market segment easily compared with AMD