[HardOCP] 390x too close for comfort to R9 Fury

[Arachnotronic]

Not surprised in the least they won't test Fury @4K because that is where the card comes into its own.

Another black mark against H.

How popular is 4K gaming today? I would say performance at 2560x1440 is far more relevant to gamers than 4K.

 

[AnandThenMan]

Fury is not a volume card it is high end, if you want to talk about popular then bench at 1680x1050 or 1920x1080 those are the most common resolutions by far. [H] is painting themselves into a corner at some point they will have to bench at 4K and when they do they'll look like hypocrites.

...plus 4K is the future and a card that does well at 4K bodes will for future games which will push the hardware.

 

[SK10H]

It is an outstanding deal but sadly not in Canadian dollars only US.

I was debating this card vs the MSI 390 and decide to get the 2x 390 instead. This should last until the watercool 14nm HBM2 with more than 4GB, as the Fury is 2x as expensive with too many unknown, and it's better to blow the difference on the next node and wait for Cdn to recover, if ever.

256sp more vs under-utilized 8GB, I pick the better warranty MSI in Canada eventually, may also be useful for loading all the high res texture mod if needed and better resale value as no one should believe any 390 has been used as a mining card.

 

[x3sphere]

Fury is not a volume card it is high end, if you want to talk about popular then bench at 1680x1050 or 1920x1080 those are the most common resolutions by far. [H] is painting themselves into a corner at some point they will have to bench at 4K and when they do they'll look like hypocrites.

...plus 4K is the future and a card that does well at 4K bodes will for future games which will push the hardware.

They are benching single GPU. 4K isn't doable on recent games with a single card, we already know this from Fury X, 980 Ti benches.

I'd imagine they will focus primarly on 4K in the Crossfire review.

 

[Majcric]

You beat me to it. I don't like to see 4k benches unless they are what is considered playable. And for the most part 4k still requires 2 cards.

 

[Silverforce11]

[H] is a useless metric for discussing overall GPU performance, they have few games, most are GameWorks.

They are intended to tell you how GPUs perform in a few select titles.

/thread

 

[tential]

[H] is a useless metric for discussing overall GPU performance, they have few games, most are GameWorks.

They are intended to tell you how GPUs perform in a few select titles.

/thread

I completely disagree. MANY people quoted the 290x vs GTX 980 article that came out earlier to say how the 290x really was closer to the GTX 980 than the GTX 970 and for far cheaper. It showed just how GREAT a deal the 290x was. They test differently, they do hit gameworks titles but it's not like they picked OBSCURE titles in this review? Besides dying light (I know nothing about this game.... didn't care gameworks wrecks the performance so I ignore the game), these are MASSIVE titles.

H is a great reference point for big games. Hitting 5 of the biggest releases I'm most likely to play this year is good for me. I do wish H would test more titles and more cards though. Not sure what their budget is that prevents them from doing more cards like other sites. But it's still a good site...

 

[Silverforce11]

I don't need to quote [H] to show R290X was really close to the 980. Actually if you read their own R290X vs 980 articles, it shows a gap of 25-50% due to GW titles. If you read elsewhere, like Guru3d, TPU etc, it's ~10%.

The only massive titles on their test: Witcher 3 and BF4.

The rest are actually obscure, as in few people even play them. Don't believe me? Just look for yourself!

http://steamcharts.com/top

So they can pick any other titles but nope, let's pile on GameWorks. They would include Batman: AK as well but it was canned.

ps. If any FPS needs to be included, its ARMA 3. Still very popular, still crushes GPUs.

 

[sandorski]

"300 series rebrands can't compete, AMD is so screwed.." was the chorus a week ago, can't wait to see how the haters spin this.

hehe, no kidding right.

 

[cen1]

The review reads as if the conclusion was written first and then they tried their best to justify their agenda throughout the article. [H] can't surprise me no more.

 

[soccerballtux]

The review reads as if the conclusion was written first and then they tried their best to justify their agenda throughout the article. [H] can't surprise me no more.

I thought that was normal for [H]?

 

[AnandThenMan]

The thing is what Kyle said makes no sense. They won't test Fury at 4K because they have determined it is too slow and they don't want to use resources to test. Well how do you know unless you test? His statement is contradictory and self defeating.

 

[Atomic Playboy]

The only massive titles on their test: Witcher 3 and BF4.

The rest are actually obscure, as in few people even play them. Don't believe me? Just look for yourself!

http://steamcharts.com/top

I'll grant you Dying Light is obscure, but are you seriously arguing that GTA V and Far Cry 4 are niche titles? Or are you of the mindset that they should be benchmarking using Minecraft, so we can see that Fury gets 500 FPS in a title you can max on an integrated GPU? I mean, seriously, look at the list of top Steam games; they're either not graphically intensive (like DOTA 2 or TF 2 or CS GO), or they're early access games like Ark whose lack of optimization you've already publicly lambasted elsewhere on this forum. Oh, and number 11 is that niche title GTA V that you've already dismissed as being "not massive" enough to warrant inclusion in HardOCP's test suite. So, pray tell, what games would you include for an unbiased representation of how a GPU does in the popular games available to the gaming public today?

 

[mindbomb]

For BF4, the shaders are specifically optimized for hawaii (and bonaire by proxy) since the game was actually bundled with cards, so the driver team felt obliged to do that.

I don't know why these optimizations were not ported to tonga and fiji, and why instead we are getting useless gimmicky features like VSR and frame rate control that are also inevitably have problems from being unmaintained.

 

[Silverforce11]

FC4 is obscure, how many people play that compared to Arma III?

I forgot about GTA V, its a great title. [H] once said they use games that are currently popular and demanding. A few of the titles they test don't match those criteria. But my original statement was regards their small test samples is not a good metric for overall GPU performance because its low sample size.

 

[Atomic Playboy]

The Far Cry series has sold well; FC 4 is no exception. So I don't see it as a niche title. I agree with you that they should expand their test suite beyond 5 games, but what would you add beyond Arma 3? Most of the big titles from the last year have been performance nightmares not because they're particularly graphically intensive, but because they're poorly optimized. Can't use Watch Dogs or Assassin's Creed Unity or Batman Arkham Knight or Project Cars. Sleeping Dogs? Tomb Raider? Who plays those anymore? Ryse? Wasn't ever that big. Shadow of Mordor, Bioshock Infinite, Alien Isolation? Not demanding enough. Maybe Crysis 3?

 

[Silverforce11]

I don't really care what titles [H] picks since I don't consider their results worthy of discussing overall performance. They simply lack the numbers of games. At best they are a cross-section view into a few cherry picked titles. Currently its mostly GameWorks, so they are the site you go to, if you want to find out how well AMD or NV performs in NV sponsored titles.

 

[n0x1ous]

FC4 is obscure, how many people play that compared to Arma III?

I forgot about GTA V, its a great title. [H] once said they use games that are currently popular and demanding. A few of the titles they test don't match those criteria. But my original statement was regards their small test samples is not a good metric for overall GPU performance because its low sample size.

Far Cry 4 is one of AMD's best showings on Fury - thought youd be happy with that inclusion

 

[Silverforce11]

Far Cry 4 is one of AMD's best showings on Fury - thought youd be happy with that inclusion

It's not about that, its lack of titles. Justification = popular titles & GPU demanding. A bunch of their games don't match criteria.

It's about the invalid nature of using [H] to compare overall GPU performance due to their low sample size and cherry picked nature. Get it?

If people want to discuss overall GPU performance, then use sites that tests a lot of games.

 

[n0x1ous]

If people want to discuss overall GPU performance, then use sites that tests a lot of games.

Testing a lot of games can't be the only criteria that makes a review valid for overall performance. Apoppin tests more games than anyone but you would be one of the first to say his review is worthless

 

[Silverforce11]

Testing a lot of games can't be the only criteria that makes a review valid for overall performance. Apoppin tests more games than anyone but you would be one of the first to say his review is worthless

The numbers/data would have to be close or match the trend from many other tests too ofc. Otherwise numbers way off one way or the other, cannot be trusted.

[H] is useful if you want to know performance in specific games. Other sites that test lots more, are the metric to use for overall performance.

 

[digitaldurandal]

Not surprised in the least they won't test Fury @4K because that is where the card comes into its own.

Another black mark against H.

What are you talking about.

Don't get me wrong I know H has a test suite that is gameworks heavy and sometimes they say things that I am unsure if they really believe like when they claimed 4 GB was holding back the Fury X but the 295x2 performed better with 4 GB per card.

However I believe you may be ignorant to a few facts here

1 - H tests up to the max playable settings
2 - at 1440p Fury was getting around 54 fps average in Far Cry 4, not going to do great at 4k
3 - the Fury is a different card than the Fury X and is not as powerful

 

[ocre]

Perhaps he assumes that H hasn't ever tried Fury at 4k even though the site preaches max playable settings.

 

[tential]

Over here in Australia, a Sapphire 390 is $509 while the 390X is $659. Is the X really worth that price difference? A Gigabyte 980 is $709.

R9 290 (r9 390 now) leads the price to performance pack for a reason my friend.

 

[tential]

Testing a lot of games can't be the only criteria that makes a review valid for overall performance. Apoppin tests more games than anyone but you would be one of the first to say his review is worthless

Sorry my friend but he earlier said:

I don't need to quote [H] to show R290X was really close to the 980. Actually if you read their own R290X vs 980 articles, it shows a gap of 25-50% due to GW titles. If you read elsewhere, like Guru3d, TPU etc, it's ~10%.

The only massive titles on their test: Witcher 3 and BF4.

The rest are actually obscure, as in few people even play them. Don't believe me? Just look for yourself!

http://steamcharts.com/top

So they can pick any other titles but nope, let's pile on GameWorks. They would include Batman: AK as well but it was canned.

ps. If any FPS needs to be included, its ARMA 3. Still very popular, still crushes GPUs.

50%? Ok.... ya that's the end of my conversation with him.

The review is here, if you're going to throw around widly inaccurate figures there is no point in talking with the person.
http://www.hardocp.com/article/2015/03/30/powercolor_pcs_r9_290x_video_card_review#.Vac2dflVhBc