Fable Legends DX12 benchmark

Let's not lose perspective here folks.

In the end, we'll just have to play the DX 1-11 games we have in front of us now with the cards we haec available to us and then when these DX12 games come to be, we can all benefit from [H]ardOCP's no-nonsense alls-you-can-do-is-alls-you-can-do testing to see what's the best buy for our respective price points.

* I'm still *really* hoping that my GCN 1.0-based card, otherwise 5.5 year old system, finds extended life, and a latent curiosity of whether AMD was overly "forward-thinking" with GCN (and under-performed/missed-the-boat in the mean time).
 
arguing.jpg
 
You guys should really bring your 390s to the tropics some time and try gaming there, without air conditioning. Then try with a 970. You tell me which is more pleasant in a room with little to no airflow (no wind).

100W less heat doesn't just mean lower bills, it means more comfort as well. It also means working well in many many more situations out there (bad cases with very little airflow).

Either way, what I don't see people questioning enough is AMD's released results. Why do they differ so much from actual results? They say that it's the Ultra preset, but there's nobody that managed to pull those results off.
 
jwcalla said:
Why the hell would anybody publish AMD's benchmarks?
Click to expand...
ExtremeTech put a little paragraph in their article explaining why they posted them. It's because they only had a 980 Ti & Fury X in-house and wanted to show the other cards' results, too.
 
trandoanhung1991 said:
Either way, what I don't see people questioning enough is AMD's released results. Why do they differ so much from actual results? They say that it's the Ultra preset, but there's nobody that managed to pull those results off.
Click to expand...

What are you talking about, this is the first time in a while when AMD actually didn't sugar-coat their benchmark results. They used reference cards in their tests. Most of the sites used custom cards, thats why those results differ a lot.

AMD results:
AMD-Perf1.png


ExtremeTech used reference cards:
Fable-Retest.png


Here's what ExtremeTech wrote about those AMD 390 / 980 results:
Why include AMD results?
In our initial coverage for this article, we included a set of AMD-provided test results. This was mostly done for practical reasons — I don’t actually have an R9 390X, 390, or R9 380, and therefore couldn’t compare performance in the midrange graphics stack. Our decision to include this information “shocked” Nvidia’s PR team, which pointed out that no other reviewer had found the R9 390 winning past the GTX 980.

Implications of impropriety deserve to be taken seriously, as do charges that test results have misrepresented performance. So what’s the situation here? While we may have shown you chart data before, AMD’s reviewer guide contains the raw data values themselves. According to AMD, the GTX 980 scored 65.36 FPS in the 1080p Ultra benchmark using Nvidia’s 355.98 driver (the same we driver we tested). Our own results actually point to the GTX 980 being slightly slower — when we put the card through its paces for this section of our coverage, it landed at 63.51 FPS. Still, that’s just a 3% difference.

It’s absolutely true that Tech Report’s excellent coverage shows the GTX 980 beating past the R9 390 (TR was the only website to test an R9 390 in the first place). But that doesn’t mean AMD’s data is non-representative. Tech Report notes that it used a Gigabyte GTX 980, with a base clock of 1228MHz and a boost clock of 1329MHz. That’s 9% faster than the clocks on my own reference GTX 980 (1127MHz and 1216MHz respectively).

Multiply our 63.51 FPS by 1.09x, and you end up with 69 FPS — exactly what Tech Report reported for the GTX 980. And if you have an NV GTX 980 clocked at this speed, yes, you will outperform a stock-clocked R9 390. That, however, doesn’t mean that AMD lied in its test results. A quick trip to Newegg reveals that GTX 980s ship in a variety of clocks, from a low of 1126MHz to a high of 1304MHz. That, in turn, means that the highest-end GTX 980 is as much as 15% faster than the stock model. Buyers who tend to buy on price are much more likely to end up with cards at the base frequency, the cheapest EVGA GTX 980 is $459, compared to $484 for the 1266MHz version.

There’s no evidence that AMD lied or misconstrued the GTX 980’s performance. Neither did Tech Report. Frankly, we prefer testing retail hardware when such equipment is available, but since GPU vendors tend to charge a premium for higher-clocked GPUs, it’s difficult to select any single card and declare it representative.
Click to expand...

I'll actually give kudos to AMD for not sugar coating the results by running these without AF or some other bizarre settings like they did with Fury X :cool:
 
Last edited:
They have Fury X ahead of the 980 Ti, and the 390 ahead of the 980.
Can't find that replicated on AnandTech, PCPer, TechReport, or PCGH.

All 4 sites show the 980 Ti ahead of FX. TechReport shows the 980 ahead of the 390, the other sites didn't test those cards specifically.
 
Who cares about energy saving, not like my 980 ti uses less than 300w when OCed, but other than one outlier, other reivews have shown that 980ti beats fury X by a significant margin, give that they cost the same, it's kind of a non brainier.

And don't bring up the WC is extra $$$ BS, you can buy a non reference 980 ti, OC to get another extra 15%+ performance and still being quieter than a stock fury X, which can hardly be OCed at all.

BTW My last 4 cards are ATI/AMD, but unlike fanboys, I am a rational consumer who buy the best product for the money.
 
TaintedSquirrel said:
They have Fury X ahead of the 980 Ti, and the 390 ahead of the 980.
Can't find that replicated on AnandTech, PCPer, TechReport, or PCGH.

All 4 sites show the 980 Ti ahead of FX. TechReport shows the 980 ahead of the 390, the other sites didn't test those cards specifically.
Click to expand...

Anandtech used custom cards, TR used Asus Strix 980TI which boost, what almost 1.4ghz out of the box. PCPer most likely used custom cards. Reference 980Ti vs custom 980Ti makes a hell of a difference. Same with 980

Dum dum, PCGH is using custom cards for both vendors and custom 390X just pawns custom 980 (and beating even nano). Older cards like 280X / 7970 and 290 are also looking really good.
 
Last edited:
trandoanhung1991 said:
You guys should really bring your 390s to the tropics some time and try gaming there, without air conditioning. Then try with a 970. You tell me which is more pleasant in a room with little to no airflow (no wind).

100W less heat doesn't just mean lower bills, it means more comfort as well. It also means working well in many many more situations out there (bad cases with very little airflow).

Either way, what I don't see people questioning enough is AMD's released results. Why do they differ so much from actual results? They say that it's the Ultra preset, but there's nobody that managed to pull those results off.
Click to expand...

Why would you leave the air conditioner off if you lived in the tropics?
 
jwcalla said:
Why the hell would anybody publish AMD's benchmarks?
Click to expand...

So they don't get blacklisted. AMD has made it clear if you don't give them a positive review, they won't give out a review sample.

Just another reason to not buy their crap.
 
Osjur said:
Anandtech used custom cards, TR used Asus Strix 980TI which boost, what almost 1.4ghz out of the box. PCPer most likely used custom cards. Reference 980Ti vs custom 980Ti makes a hell of a difference. Same with 980

Dum dum, PCGH is using custom cards for both vendors and custom 390X just pawns custom 980 (and beating even nano). Older cards like 280X / 7970 and 290 are also looking really good.
Click to expand...
Ah so that's what AMD did, they ignored Nvidia's dynamic boost. Probably locked all the Nvidia cards to their sheet clocks. Effectively, the absolute lowest any of those cards could possibly perform. ET went out of their way to make a point of that.

On the other hand, some of the other sites went too far in the opposite direction. I don't see any evidence that AnandTech or PCPer used custom cards so I'll throw that out the window for now. AT did say they used old drivers, though. Both AnandTech's and PCPer's numbers are low enough to be reference cards (compare the performance gap to TechReport's numbers)... TechReport totally screwed the pooch in their test.

So in short it seems AMD used the fairest way to test the cards but also the most detached from reality. :rolleyes: It's their way of saying "GPU Boost 2.0 is more effective than what we use, so we're going to ignore it".

It looks like TechReport is about to get blacklisted, though. lol
 
I don't really believe they locked Nvidia cards to sheet clocks as it would have been shown in ET test when compared to AMD results. Heck AMD's ref. 980 got better result than ET's ref. 980. Anandtech said they used Evga 970 but they didn't say if it was custom or not.

Why can't sites say if they are using custom cards or not ffs. Everybody and their mom knows that custom 9xx series boost like nothing else with custom cards. A lot of misconceptions would be solved when respectable sites would bring the facts to the table and not keep some info out.
 
Whenever i need to explain to someone what does Mental Gymnastics mean i just point them to these threads and the back n forward, goal post moving and the like.


Silly me thinking in page 4 that the results were gonna be discussed.


Let's try something else, i'm seeing a 550$+ card (390x) beating a 650$+ card (980) at ultra 1080p, not bad, it will draw more power so the yearly cost is gonna be around the same, but it also comes with more ram so there is that, seems that the "old tech" still has some life on it.

Still i wish we could see more cpu scalings done properly to see what could be going on with AMD and i also wish that we could see some actual gameplay-ish performance, again, with cpu scaling, testing the cpu overhead improvements of the dx12 api.
 
Osjur said:
I don't really believe they locked Nvidia cards to sheet clocks as it would have been shown in ET test when compared to AMD results. Heck AMD's ref. 980 got better result than ET's ref. 980. Anandtech said they used Evga 970 but they didn't say if it was custom or not.
Click to expand...
Both PCPer and AMD used the same X-axis in their graphs, so I scaled them to the same size and overlaid them.
In AMD's results, all of the red bars are slightly longer and green bars slightly shorter.

So yes there is something fishy. Either AMD is exaggerating their GPUs, or shrinking Nvidia's performance, or PCPer did the opposite.
Seems to be a discrepancy of about 5-10% at the most. So not exactly earth shattering.

OXtVm3c.png
 
Last edited:
Never believe in bar graphics, both companies are notorious for never starting them in 0, it exaggerates everything and thus makes themselves look better while diminishing their opponents.
 
Revdarian said:
Never believe in bar graphics, both companies are notorious for never starting them in 0, it exaggerates everything and thus makes themselves look better while diminishing their opponents.
Click to expand...
Hey bar graphs with poor normalization is a great way to get people to think your cards 10% increase over a competitors is a 100% increase.
 
TaintedSquirrel said:
Both PCPer and AMD used the same X-axis in their graphs, so I scaled them to the same size and overlaid them.
In AMD's results, all of the red bars are slightly longer and green bars slightly shorter.

So yes there is something fishy. Either AMD is exaggerating their GPUs, or shrinking Nvidia's performance, or PCPer did the opposite.

OXtVm3c.png
Click to expand...

A bar graph without numbers doesn't tell much for sure. More often than not the only thing accurate is the actual number. Now i only looked at the benchmark on YouTube, but it didn't look very impressive at all..at least not cutting edge
 
TaintedSquirrel said:
Both PCPer and AMD used the same X-axis in their graphs, so I scaled them to the same size and overlaid them.
In AMD's results, all of the red bars are slightly longer and green bars slightly shorter.

So yes there is something fishy. Either AMD is exaggerating their GPUs, or shrinking Nvidia's performance, or PCPer did the opposite.

OXtVm3c.png
Click to expand...

Or the explanation is simple as it gets: AMD used different test setup than PCPer and nvidia got a little boost with better CPU (Skylake), also not all 9xx series are equal, they boost differently from one unit to other and when benched in open air vs in case.

But my point is that this time AMD actually tested in somewhat realistic scenario compared to those horrendous Fury X internal benchmarks :D

EDIT: Btw, it seems that Async compute is enabled in this test, though it is used really lightly, like 5% per frame.
 
Last edited:
I am not looking through 8 pages but it sure is odd that the other reviews I looked at such as on Techreport have Nvidia clearly ahead and nothing like what the OP posted. :confused:
 
tybert7 said:
Why would you leave the air conditioner off if you lived in the tropics?
Click to expand...

Because if you open air conditioning for a few hours a day, the cost adds up to about 5-15% of the price of a 970 each month? Over here, quite a few are willing to splurge on better systems, even water cooling, but will not use air conditioning to save money to splurge on better hardware.

I have air conditioning on anywhere between 12-18 hours a day, and with power consumption at nearly 600W when cooling, that's adding on over $40/mo. Many would rather use fans and save the difference to buy better parts.

So yeah, not everyone can afford air conditioning. Though I do agree with you. Living in tropical climate, once you go air con you'd be hard pressed to go back.
 
5150Joker said:
Maybe in his part of the world? :confused:
Click to expand...

Not so far off just north of the US border. Locally here in Canada an XFX 390x will run about 570 plus tax the cheapest 980GTX made by Zotec will run you 625 plus tax. For a 980GTX made by one of the bigger names 660 and up. So yes close to a hundred dollar difference.
 
It would be interesting to know which 290x anandtech used in the benchmark as performance could even be better if they used a ref model.

Now I bought my Sapphire Tri X 290X New Edition brand new for $269 with 1020/1350 stock clocks and the card runs super cool and no throttling ..

So that is what all the cards benchmarked have to compare too and that is a huge price different's when looking at 980Ti /Fury X and what you get for the extra $400 you had to spend in just this one game.

Also to let you know the card runs around 64c with the Tri X cooler so that 970 GTX of yours is not going to run any cooler. but yes a ref Hawaii was 94c and hot as H*** AIB's have improved there cooling and phase power design on Hawaii PCB's and why we are seeing performance gains from Hawaii vs 2 years ago.
 
Last edited:
fanboy said:
It would be interesting to know which 290x anandtech used in the benchmark as performance could even be better if they used a ref model.
Click to expand...

I would bet on the ref. and when looking at the PCGH review, we can see that custom 290 (not X) is 3% slower than custom 980 which makes me :eek:

And original 7970 being over 30% faster than 960 judging from Anandtech article is another wtf moment.

Of course there's only 2 data points, the other being ashes benchmark where AMD gets similar gains so we can't really make any conclusions yet but for now, it looks mighty fine from AMD perspective in all price points except for true high-end where Fiji just can't beat 980Ti.

Fiji is definitely bottlenecked at frontend because custom 390X is only 5% slower than Fury X even though Fury X has 45% more shaders and a lot more bw at its disposal. I don't know what kind of stuff they were smoking when designing Fiji but it was really stupid decision not to increase ROP count even with the expense of having slightly less shaders.
 
Last edited:
The GTX 980 since release has had poor value even relative to Nvidia's own lineup in terms of gaming performance. It is a bit disingenuous to use it as a comparison point in terms of value due to the very low bar set for that metric. This also applies to Titan products.

If we look at the 390x while it may look good compared to the GTX 980 (which just about anything does) you are also paying 30% more for at most 10% more performance than the 390 (assuming completely shader bound). Neither is really the best value.

Edit: Just to add. For anyone actually looking for the best value in terms of pure gaming performance the best bet is still r9 280/x and r9 290/x at clearance prices while they last.
 
Last edited:
This is my value tree.

250 or below = 290X or used GTX 970
650 and above = 980 Ti

Everything in between is for ill informed people or people who are similar to ones quoted in post 123.
 
jwcalla said:
A 980 is $650?
Click to expand...

The prices in my country are as follows:

290 $351,49
290X 403,47
390 $400,75
390X $509,38
Fury $663,65
Fury X $828,47

970 $380,41
980 $589,23
980 Ti $751,79

(Location Denmark, Europe)
 
Last edited:
jwcalla said:
A 980 is $650?
Click to expand...

I just glanced at newegg for both of those prices, and yeah i didn't try to convert because the economy in my country is beyond whack.

And yes, new 980's non ti are 650 on newegg, take it with them if the price was not accurate for you.

Edit to add:

And yeah when i looked at newegg the 390x was also at 550$, so relatively to other places, both videocards seems equally overpriced. The point stands tho, the 390x is cheaper upfront with twice the memory but will require higher power equalling the price difference over time.

Edit to add 2:

Seems that my add blocker was making it think that i wanted canadian prices or something, will look them both up again.

450 vs 400, 50$ difference (cheaper) upfront for twice the memory, depending on use then indeed on a regular case scenario they are equal after a year.
 
Last edited:
Factum said:
The prices in my country are as follows:

290 $351,49
290X 403,47
390 $400,75
390X $509,38
Fury $663,65
Fury X $828,47

970 $380,41
980 $589,23
980 Ti $751,79

(Location Denmark, Europe)
Click to expand...

I envy those prices so damn much.

Which tells you how shitty prices are over here.
 
TaintedSquirrel said:
Both PCPer and AMD used the same X-axis in their graphs, so I scaled them to the same size and overlaid them.
In AMD's results, all of the red bars are slightly longer and green bars slightly shorter.

So yes there is something fishy. Either AMD is exaggerating their GPUs, or shrinking Nvidia's performance, or PCPer did the opposite.
Seems to be a discrepancy of about 5-10% at the most. So not exactly earth shattering.

OXtVm3c.png
Click to expand...

There is no conspiracy. PcPer is using an older driver AMD Catalyst Cat 15.201.1102, just like Anandtech for their testing. I'm going to jump the gun and asume that Extremetech and AMD used the latest driver which improved AMD scores across the board.
 
DeadSkull said:
There is no conspiracy. PcPer is using an older driver AMD Catalyst Cat 15.201.1102, just like Anandtech for their testing. I'm going to jump the gun and asume that Extremetech and AMD used the latest driver which improved AMD scores across the board.
Click to expand...

Bingo!

https://mobile.twitter.com/Thracks/status/647085787041910784

This is basic stuff for "review" sites to confirm latest information at time of review, really not impressed.
 
You must log in or register to reply here.
Back
Top