FineWine===>FineVinegar FuryX Aging Horrible in 2017

[Vario]

Get a small enough monitor with a high enough resolution and you don't need AA, thats my strategy With most games once you get into the high action you don't notice if its on or off.

 

[cdawall]

4K and AA maybe redundant, 1440P and AA is a must.

I run AA on 4K and you can tell the difference even on a 23.8" 4K, but I also have 3 1080Ti's so I don't play this "what can I reduce game"

 

[MuhammedAbdo]

I guess being old and having poor eyesight helps, AA'ing at 1440p seems kinda conter-intuitive for me. For each it's own, though.

It's not AA that has the Fury tripping over, it's usually textures and Tessellation.

 

[Fouquin]

I kind of want to play with them some more now, how are you measuring power consumption?

I have a kill-a-watt on the wall and I did some baseline readings without GPUs while configuring the base system. As a secondary I used HWiNFO to get power readings within Windows when I was dialing in the undervolt.

 

[eidairaman1]

MuhammedAbdo, you have been a Member Since 2009, where are your system specs???

 

[Vayra86]

Not bologna. 48FPS is 20.833ms between displayed frames, 40FPS is 25ms. In a blind test I doubt anyone would guess the difference between those two.

Really? We run 120 fps/hz because of how it feels lightyears smoother than 60 fps.

And below 60 fps, you very much notice variation in FPS. If there was any game where you'd remain steady at 40 fps AND 100% GPU usage all the time, then your point makes any sense. Realistically, there is no single game that does this.

Give me a double blind test where I have to guess which setup is shifting between 40-48 fps and which is not, and I'll guarantee 10/10

Except Kepler hasn't aged poorly and the people proposing that idea were proven wrong multiple times. Also, what is the time constraint on considering hardware "high end"? 2 years, 5, 10? A lot of publications retire GPUs from testing at 3 years, and Fiji is not far from that mark. Do we expect GPUs to be topping performance metrics in applications designed to utilize exceptionally more resources?



If you're mining ZCash the Fury X is actually quite good. 550 Sol/s at 145W is better perf/watt than a GTX 1070.

Didn't you get the memo yet, its about aging poorly as in: the *relative* performance between the equal cards at about equal release dates of the past, has shifted in favor of one card. Ironically, but comparison, Kepler did age poorly to GCN's 7970 for example - mostly due to the latter gaining more over the years, while Kepler gained most of its performance early on, which was needed for it to catch up while AMD rebranded to 280x and Nvidia had moved on. And not unimportant - 7970 had 3GB

 

[Xzibit]

Really? We run 120 fps/hz because of how it feels lightyears smoother than 60 fps.

And below 60 fps, you very much notice variation in FPS. If there was any game where you'd remain steady at 40 fps AND 100% GPU usage all the time, then your point makes any sense. Realistically, there is no single game that does this.

Give me a double blind test where I have to guess which setup is shifting between 40-48 fps and which is not, and I'll guarantee 10/10

I believe he was talking about a constant where you introduced Variation.

 

[Vayra86]

I believe he was talking about a constant where you introduced Variation.

Right, but the topic at hand is in game performance, not a theoretical FPS cap.

 

[MuhammedAbdo]

Didn't you get the memo yet, its about aging poorly as in: the *relative* performance between the equal cards at about equal release dates of the past, has shifted in favor of one card. Ironically, but comparison, Kepler did age poorly to GCN's 7970 for example - mostly due to the latter gaining more over the years, while Kepler gained most of its performance early on, which was needed for it to catch up while AMD rebranded to 280x and Nvidia had moved on. And not unimportant - 7970 had 3GB

Exactly, Kepler underwent this exact transformation 2 years after it's introduction, most Kepler cards had VRAM limitations compared to their AMD counterparts (ie 680 2GB vs 7970 3GB or 780Ti 3GB vs 290X 4GB). Also Kepler was quite bad at compute. The combination of both factors lead to 780Ti being often behind 290X for example. And by quite a distance. Same thing is happening to the FuryX in fall 2017 now. (a combination of limited VRAM and bad Tessellation performance).

 

[cucker tarlson]

http://www.pcgameshardware.de/Star-...s/Star-Wars-Battlefront-2-Benchmarks-1243688/

980Ti SJS is

1.30x faster at 1080p
1.27x faster at 1440p
1.18x faster at 3440x1440
2.09x faster at 4K

In comparison, Fury X is only 6% faster than 1060 at 1080p and 13% faster at 1440p and loses severely at 4K, a resolution many people bought their Fury X for as AMD marketed it as a 4K card

It cost $650 in June 2015, 1060 was $250 (38% of Fury X price) a year later. Yes, it ages sooo well....

TPU overall performance charts are whatever they wanna report, based on inconsistent results. They still have Fury X beating 1070/980Ti in DOOM, and use DX12 in BF1/Mankind Divided.

 

[Vya Domus]

In 2015 4K gaming was largely still a joke. Neither the Fury X or the 980ti or Titan X could actually run games at 4K with playable frame rates without major compromises. And now you're talking about a game released just now and also punting it into the context of newer graphics cards ?

I wonder just how ridiculous this discussion will get.

 

[cucker tarlson]

Don't tell me that. Tell those who bought it.

 

[ikeke]

Bought it, had fun, moved on.

http://www.guru3d.com/articles_page..._pc_graphics_analysis_benchmark_review,5.html

These results are more in line with what i've experienced first hand, wrt all the performance and positioning in performance lineup. (why would you run a DX12 native game on DX11? just to prove that nvidia has better DX11 driver/perf? dunno...seems funny)

 

[P4-630]

http://www.guru3d.com/articles_page..._pc_graphics_analysis_benchmark_review,5.html

Good to know my 1070 still does 79fps @ 1440p.....
I'm all good.

 

[Vayra86]

Its okay, when cards don't age well, you can just dial back a few sliders from ultra to very high and see a 30% FPS uptick.

 

[Vya Domus]

Its okay, when cards don't age well, you can just dial back a few sliders from ultra to very high and see a 30% FPS uptick.

Seems like not everyone agrees.

http://www.guru3d.com/articles_page..._pc_graphics_analysis_benchmark_review,5.html

These results are more in line with what i've experienced first hand, wrt all the performance and positioning in performance lineup. (why would you run a DX12 native game on DX11? just to prove that nvidia has better DX11 driver/perf? dunno...seems funny)

Can't do anything about it , fanboys will always cling to whatever they got.

 

[MuhammedAbdo]

TPU overall performance charts are whatever they wanna report, based on inconsistent results. They still have Fury X beating 1070/980Ti in DOOM, and use DX12 in BF1/Mankind Divided.

Yep, testing a lot of DX12 titles with worse DX12 performance than DX11: Battlefield 1 and Deus Ex Mankind Divided have horrendous DX12 support that harms fps on all GPUs. DX11 works better for all GPUs. Tomb Raider and Hitman perform badly on the 980Ti with DX12, but perform faster on DX11. They should test with the best API for each GPU methodology, a lot of sites do this like Anand, ComputerBase and PCGH.

TPU also test several games with less than max settings. Which sometimes skew the results in favor of a certain GPU.

 

[xkm1948]

Morale of the story: don't buy high end AMD GPU. And FineWine is now a lie.

 

[MuhammedAbdo]

http://www.guru3d.com/articles_page..._pc_graphics_analysis_benchmark_review,5.html

These results are more in line with what i've experienced first hand, wrt all the performance and positioning in performance lineup. (why would you run a DX12 native game on DX11? just to prove that nvidia has better DX11 driver/perf? dunno...seems funny)

There is something called "testing different areas in the game". Guru3d tested the opening sequence which relatively closed spaces, these run at insane fps because none of it is visually taxing. They also tested with DX12 which is proven to hampers fps on every GPU from AMD or NVIDIA.

https://www.computerbase.de/2017-11...nchmark/2/#abschnitt_directx_11_vs_directx_12

https://overclock3d.net/reviews/software/star_wars_battlefront_ii_trial_pc_performance_review/3

http://www.pcgameshardware.de/Star-...s/Star-Wars-Battlefront-2-Benchmarks-1243688/

PCGH tested a later level with wide open spaces and a lot of combat, these are the taxing areas, and these are the representative ones giving they are essentially the same as multiplayer maps. And these are the areas where the FuryX falls flat on it's face.

 

[Ithanul]

Seriously , don't lie to yourself that AMD burned you , in the end ya'll bought exactly what you wanted. No one deceived you , you're all pretty smart. Coming after close to 3 years and complaining that your (once) very expensive and powerful card doesn't keep up and eat alive all the new games is rather meaningless and pointless to say the least.

How about the first Titan ? Paying 1000$ for it and now a 1060 trashes it ? Should those be mad too ? Did Nvidia burned them ? No , you should buy something and set a realistic expectation of how it will fare over time.

If you think Nvidia will give you a card that will provide unrivaled performance millennia after you bought it then by all means , be our guest no one is stopping you. I don't think it will but hey.

As someone who owned an OG Titan and seeing all the others complain when the 9XXs and 1XXXs series came out.
One: dumb nuts for paying that price (I bought mine used under 700 [got to love individuals having buyer remorse])
Two: new arch, they should of know better to sell off the cards and flip over to the 9XX series. Hell, my GTX960 @ 1550MHz was pacing my OG Titan at folding. I knew it was time to dump the cards.

Pretty much, if you going to be into PCs, you better get use to something newer coming out and smacking your high end hardware that is no longer the top dog.

 

[Thefumigator]

I personally don't believe in "good aging" of hardware in general. But sometimes you have a winner in the purchase game and sometimes you don't.

 

[cdawall]

I personally don't believe in "good aging" of hardware in general. But sometimes you have a winner in the purchase game and sometimes you don't.

There are plenty of products that have aged well. First Gen i7's just became outdated nearly 10 years later, AMD 7950/7970 are still great cards for 1080P etc.

 

[xkm1948]

There are plenty of products that have aged well. First Gen i7's just became outdated nearly 10 years later, AMD 7950/7970 are still great cards for 1080P etc.

A good 2500K can still kick ass and take names today.

 

[eidairaman1]

A good 2500K can still kick ass and take names today.

The lowest Id Go for any machine today is Phenom 2/core i7 2k, phenom and 1st gen core i are finished...

 

[xkm1948]

That was just an example of good hardware aging.