High DX11 CPU overhead, very low performance.

No, it is not, as you can see from the original post, in similar priced/tier cards like the 270x and the 750 Ti.
All these below are similar tier/spec cards, and their relative performance drops. You respond as if you didn't even read the original post.

Spoiler

Far Cry 4.
GTX 750 Ti Loss: 0%
R9 270x Loss: -27.27%


Far Cry 4 again.
GTX 750 Ti Loss: 10.52%
R9 270x Loss: -46.66%


Now the same game, with the R7 260x.

Far Cry 4.
GTX 750 Ti Loss: -10.52%
R7 260x Loss: -13.63%

You will notice that the more low-end an AMD GPU is, the less is bottlenecked. Keep it in mind until we get to the R9 280 numbers.

Far Cry 4 again.
GTX 750 Ti Loss: 0%
R7 260x Loss: -4.54%

Same here, almost no bottleneck, but you will notice that the 750 Ti is almost steadily at zero.

Ryse: Son of Rome
Back to the R9 270x now.
GTX 750 Ti Loss: -7.69%
R9 270x Loss: -35.29%


More Ryse: Son of Rome
This time with the R7 260x. The frame rate loss is not great, notice the horrible frame spikes though.
GTX 750 Ti Loss: -7.69%
R7 260x Loss: -8.33%


Call of Duty Advanced Warfare
The R9 270x again.
GTX 750 Ti Loss: 0%
R9 270x Loss: -34.14%


Call of Duty Advanced Warfare.
R7 260x.
GTX 750 Ti Loss: 0%
R7 260x Loss: -20.58%


Call of Duty Advanced Warfare moar.
This time the lower end processor is an A10 5800K/Athlon X4 750k. The NVIDIA CPU stutters more with the AMD CPU, something that the AMD GPUs do not. The higher-end R9 270x still loses more than 50% of the high end cpu frame rate though, which is (de)impressive.
GTX 750 Ti Loss: -16.66%
R9 270x Loss: -52.38%


Call of Duty Advanced Warfare final.
Still with the lower end AMD processors, the R7 260x. Although it gets GPU-bound much faster, it still manages to lose an impressive 41.17% with the lower CPU.
GTX 750 Ti Loss: -16.66%
R7 260x Loss: -41.17%

Because that is what you buy today with X amount of money from each company. Also, both companies have a unified driver architecture.

If all things equal, one bottlenecked card goes down 40-freakin-percent and the other 0-20. That is a big difference on similar price/tier cards.

Link? Percentages? Tests? I have a source at least.

The Ti and the 260x/270 are cards on the same tier, and this is where this was firstly observed. In the GTA V article they expanded it to the R9 280 which is even worse apparently. Bottom line, please read what is actually being said/posted instead of supposing and responding to the supposition.

There is no "standard" implementation. If you were actually following the thread you would have seen in the presentation given by people from all the companies that the GPU scheduler does whatever it wants really with DX11 code. NVIDIA's GPU scheduler is doing much better up to this point. Again, read the thread.

All GPUs have lower performance with lower CPUs, the issue lies again on the percentage of performance lost (ie driver efficiency). The R9 290/GTX 780 have similar behavior as their lower end siblings:
From the Eurogamer GTA V rig article:

Reading comprehension again.
Quote from the same article (which I have quoted also on the original post that you supposedly read):

I'm not bashing AMD, I'm just trying to make sure that they don't sweep this one under the rag like they did for SSAA, VSR etc. Your logic is completely flawed also. If I spend 60$ more on a GPU, I want it to AT LEAST perform as well as the cheaper alternative, not worse.

 

Hey man your first post was better, and I kid you not man, it definitely felt a lot smoother.. I was like wtfff? There's definitely an improvement, super awesome job. :thumbup:

I mean you can tell the smoothness from my pass 1 avg fps. Seems min fps has gone up, WHICH I EXPECTED ... this proves what we're saying here and what's going on, right PrMinisterGR? And the rest?

And also, take a look at all those under 16ms percentile fps passes etc, much higher numbers. Bigger percentage.

Anyone who could do the math and add all these up? Do the avg on them, vbs?

Tbh I have a calculator I could do it myself but whatever.. XD

 

Seems to be a common problem these days.

 

I read it all again, I believe I got it all.

The articles I posted (along with Anandtech overhead tests that nobody seems to comment on, I wonder why), are comparisons on both lower end cards, and higher end ones and the AMD cards are CPU bottlenecked in both instances.

What's also nice to see in this thread is people appearing with less than 10 posts to supposedly tell us how stupid and "soviet" we are, without a single piece of evidence apart from bad reading comprehension.
The reading comprehension explanation would be the best option really, because the other one makes me believe that this thread will go away again.

 

Ok, lets take Project cars for example, performance on AMD gpus sucks on high settings (higher, the worse amd performs, suck on low too, it just doesn't bother 60fps gamers), plain and simple because of AMD driver overhead. And in the aforementioned game, 750ti really is AS FAST AS 280/280X! (FullHD) And not with an i3, but i7-4790K@4,5 GHz!
Don't believe me, take a look at this article: http://pclab.pl/art63572-7.html
Good thing it really ain't a problem, as it looks gorgeous with low/med settings too



E: I've begun to wonder if AMD is keeping this trick under their belt and waiting for W10 release. Then they will go "Looke here, we made magic!" And releasing it W10 only ofc.

 

Glad I'm not the only one that felt the smoothness. The results don't show that but who cares

 

If those numbers are true, then... uke2:

The Digital Foundry performance analysis also found that the main problem with the game is CPU utilization. Unless they pull the numbers out of their bottom orifices, the 970 is 100% faster than the R9 290x. WTF.

 

Which drivers were used for that benchmark? AMD users don't buy Project CARS!

 

macmac9 install the modded VBS DRIVER'S THEY'RE AWESOME MAN!! BETTER OVERHEAD, you'll see much more efficiency. GTA V definitely felt a lot smoother for me, I was quite amazed.

Give it a go and see how it does in project cars. You should see improvements everywhere lol, everywhere where there's cpu..

dude, super good point lol, they might even do that, WTF LOL.... amd style. XD hahahaah.. rofl

 

Nothing was improved actually judging by numbers, so it's just a placebo. Microscopic difference in min fps is just a usual variance. And how about testing some other games, which has a lot of drawcalls? Theres only overhyped crappy gta5 in this world? Horrible game nobody should care about.

 

Never mind, got it.

 

Don't come and tell me what's placebo and what's not. I know what my eyes see and I know what I saw. So stfu. Look at the min fps in the 1st pass, they're much higher. Look at all the percentages, they're much higher. Pass 2 is just an exception.

Either you're a troll or work for amd, so go away fool.

edit: sorry correction, the min fps is much higher in the first row of passes (pass 0-4).

 

I said this a few time. GTA V with 15.4 and crossfire is UNPLAYABLE because of micro-stutter. Single card is fine.
With the modded drivers I would never know that both cards are in crossfire (thank you G15). It feels just as smooth as 15.4 with both cards (if not smoother).

 

Judging by your frames under 16ms scores, these drivers give tangible improvements to judder. Most of your tests are in the 5-10% better range, and considering we talk about frametimes they matter quite a lot on the smoothness level.

According to his numbers, it did and quite a lot at that (if you consider that it is percentile frametimes we're talking about).
As Vbs explained before, my numbers were very close to the statistical deviation, so I could test again. Rome 2, Heroes of the Storm and StarCraft 2 have tangible improvements for me when I play, but I don't know how to standardize a bench for them to show results. The Rome 2 benchmark is useless, since the greatest lag is on the campaign map and in combat with fire effects that I cannot precisely duplicate.
In Heroes of the Storm even the menu before I join games is tangibly smoother, and in SC2 the SCVs don't have an almost invisible microstutter they always had when they were collecting minerals.
These are all games I've been playing for years (except Heroes, where it's only months), and I could tell the difference without even looking for it. Even my wife commented on how the Heroes unit selection screen/loading was much smoother.

15.4b
Frames under 16ms (for 60fps):
Pass 0: 446/633 frames (70.46%)
Pass 1: 345/585 frames (58.97%)
Pass 2: 443/644 frames (68.79%)
Pass 3: 741/790 frames (93.80%)
Pass 4: 4421/7393 frames (59.80%)

1018.1
Frames under 16ms (for 60fps):
Pass 0: 469/640 frames (73.28%)
Pass 1: 419/622 frames (67.36%)
Pass 2: 504/671 frames (75.11%)
Pass 3: 688/783 frames (87.87%)
Pass 4: 5469/7914 frames (69.11%)

 

I'm migrating my games over to a newer quicker HDD so once that's done I'll try get some benchmarks done comparing 15.4 to the modded drivers. I'll do crossfire and no crossfire ones too. Give about 42 hours to do it lol...

 

I'm not a troll nor an amd employee, but I see your tests and don't see any difference really. What about testing those "mega"drivers in other games? Arma 3 for ex., ACU, Dying Light, DAI etc? Let this crappy gta die already.

 

Yeah, the numbers don't lie, and it's a good thing that games are starting to use cpu too, eh?

There really is no reason to not buy project cars because amd sucks. If it would bother me too much I would boycott amd

I've tried them and they fall in between W10+tech preview drivers and W8.1+15.4b But best option in W8.1 for sure.

E: AC:U for me, in a few hours

 

*Adds Dying Light benchmarks to the list.

Does it even have a benchmark tool?

 

I agree that he was bad mannered towards you, but this thread is suddenly full of people with less than ten posts in the forum that start to tell us how we're doing it wrong without providing any kind of evidence.

You have a point about the other games, but GTA V is what most people have. I have AC Unity installed, do you have any good idea for me on how to test it? It doesn't have a benchmark mode unfortunately.

 

Crossfire is unplayable in most games actually. Don't think that those placebo drivers will change anything. Can you record 60 fps video with 15.4 and modded ones so aI can see if there any difference with my own eyes?