Radeon Fury to be slower than GTX 980 Ti

[R-T-B]

It's just funny how I got everything right so far "without evidence". I mean, when evidence did come up, it turns out to be what I was saying all along...

I'm willing to admit you're right just as much as I'm willing to admit you're wrong (not what I'm seeing above, so not sure what you're reading, but that's beside the point). What's kinda annoying to me to be honest is your level of certainty on a product when the last generation were largely rebrands.

I hope you're right for the record... but I don't share your confidence.

And I also hope you don't mind 50/50 chance of AMD closing their door on their 300 series release day then...

Of course I do. Who in their right mind would want that? But I also highly doubt it would ruin them. We have to remember, they're squeezing every last drop into R&D right now and maybe it made sense to them to do this for a gen or two to develop a new tech. *shrugs*

 

[RejZoR]

Rx-200 series were largely a rebrand because they were able to afford doing that and because we were still within a DX11 era. It made no difference either way. Kepler was, as we all know less than stellar. They don't have that luxury this time around...

 

[Tatty_Two]

@Tatty_One
ill give you a eCookie if you lock this thread

Sadly I am not a fan of cookie's, you will have to do a little better than that

 

[Super XP]

There is also no proof that the reverse is true...so you're refuting an argument because of no proof with an argument that also doesn't contain proof. Did someone ask you to show a written example of irony?

Yet the R7 370 is almost certainly a rebrand of the R9 270 (which is in turn a rebranded HD 8860/8870 OEM, which is turn a rebranded HD 7870 GHz Edition.) - note the Crossfire finger in the slide.

In all honesty, I doubt AMD do either.

Lol,
Some sites claim no re-brands and others claim all the 370's are re-Branded.

At this point it's all speculation. And yes I'm speculating on a rumoured speculation. Why not.

"""The AMD Radeon R9 300 series will reportedlycontain no re-branded graphics cards from the R9 200 series. An entirely new lineup is coming at Computex. The report in question alleges that AMD’s original plan was centered around introducing a new flagship Radeon product this month and rebranding a few other R9 200 series products to R9 300 series products."""

This is where I read about it way back. Obviously they were wrong. "Re-Branding" should be illegal.

AMD Radeon 300 Series Won’t Be A Rebrand, New GPUs Coming in June
http://wccftech.com/amd-r9-300-series-not-rebrands/

 

[Solaris17]

Um...wouldn't this be something Best Buy would put on their website? Note how the pictures never show 390X either...just R7 370 and R7 360. As the rest already pointed out, none of this comes across as remotely credible.

No, I don't know if you were around when I was the first consumer to own a pair of 9600GTs but I bought them at a local bestbuy because while we did get deliveries of the cards weeks in advance we weren't allowed to put them out. However they are in the system and allowed to be rung out. That said someone put them on the shelf too early. I walked in without my uniform on. bought them and immediately told w1zz.

No they weren't on the site. thats controlled by the company itself those system are usually all automated.

 

[64K]

This is where I read about it way back. Obviously they were wrong. "Re-Branding" should be illegal.

True but wouldn't we just be bitching about the lack of shiny new cards otherwise? The bottom line is that neither AMD nor Nvidia can put much on the table at this point. Next year on the 16nm process we will have what we want.

 

[RejZoR]

If you want to make all 300 series cards fully DX12 capable, then we can't talk about rebranding anyway. It's impossibel to begin with then...

 

[Super XP]

True but wouldn't we just be bitching about the lack of shiny new cards otherwise? The bottom line is that neither AMD nor Nvidia can put much on the table at this point. Next year on the 16nm process we will have what we want.

If the price is right then I'll upgrade my DualX. But the issue is my current card is also a rebrand lol, and I was not aware of it at the time of purchase.
Just do NOT like deception that's all.

I checked CPUID and my stomach turned LOL. Felt like a sucker actually.

 

[the54thvoid]

fully DX12 capable

What does that even mean? AMD have already said GCN is not fully compliant with some aspect of DX12 and likewise, while Maxwell 2 (GM200) is compliant in that specific area, it is not Tier 3. So really, saying fully DX12 capable is very confusing.

Go back here for more confusion http://www.techpowerup.com/213191/n...t-supporting-direct3d-12-1.html?cp=2#comments

 

[RejZoR]

Erm, isn't that what I've said? Clearly existing GPU's aren't "enough" DX12 capable to deserve name DX12 GPU's. Clearly they have to do something about it for the upcoming series. How far they'll go with DX12 features, only they know, but they certainly can't support only what current DX11.2 cards support...

 

[Super XP]

Erm, isn't that what I've said? Clearly existing GPU's aren't "enough" DX12 capable to deserve name DX12 GPU's. Clearly they have to do something about it for the upcoming series. How far they'll go with DX12 features, only they know, but they certainly can't support only what current DX11.2 cards support...

Interesting you bring this point up. But depends on when they plan on implementing DX12 Fully. ATM seems it's going to take long just like it's counter-parts.

 

[lilhasselhoffer]

12 days of this thread. Just, for one moment, let that sink in. For nearly two weeks people have argued that Nvidia is better, AMD is better, and all of this argument is based off of conflicting information and complete fabrications of "fact" that are demonstrably idiotic.

This is how religion start.


I will make this plea one last time. Can we please wait until some benchmarks appear?

DX 12 is a pipe dream right now. Despite the fact that MS is selling Windows 10 with it, it isn't ready. There's no hardware that supports all its features, it doesn't have a reasonably sized test sample showing its benefits, and even when it's introduced the consoles are unlikely to support its features 100% (so developer adoption is likely to stagnate again, similar to DX11). Becoming frothing mad, on either side of this debate, is insane. Keep calm, and let's see some actual test data before we jump to any conclusions.

 

[arbiter]

12 days of this thread. Just, for one moment, let that sink in. For nearly two weeks people have argued that Nvidia is better, AMD is better, and all of this argument is based off of conflicting information and complete fabrications of "fact" that are demonstrably idiotic.

Yea its gotten pretty sad, We know for fact what Nvidia has. Everything from AMD side has been super hype that it will be a super gpu that nothing can match, even though history says other wise. yea it does have HBM memory but just cause it has it doesn't mean performance scales 1:1 with it. There is a point were performance benefits level off. You can look at CPU performance with diff speed of ram as proof of that. As for where that of diminishing return is yet to be known.

 

[Steevo]

12 days of this thread. Just, for one moment, let that sink in. For nearly two weeks people have argued that Nvidia is better, AMD is better, and all of this argument is based off of conflicting information and complete fabrications of "fact" that are demonstrably idiotic.

This is how religion start.


I will make this plea one last time. Can we please wait until some benchmarks appear?

DX 12 is a pipe dream right now. Despite the fact that MS is selling Windows 10 with it, it isn't ready. There's no hardware that supports all its features, it doesn't have a reasonably sized test sample showing its benefits, and even when it's introduced the consoles are unlikely to support its features 100% (so developer adoption is likely to stagnate again, similar to DX11). Becoming frothing mad, on either side of this debate, is insane. Keep calm, and let's see some actual test data before we jump to any conclusions.

Years ago we watered at the mouth for DX10 and it took a long time to get a worthwhile title out that meant anything, same thing will happen here, we get DX12 and all of a sudden we have what to do with it? Need games to support the standard to take advantage of it, and that will require a year for something that doesn't run like it came flying out a monkeys ass and done for the sake of saying durr hurr, it DX12, you buy....

So just like when tessellation became the new huge toy, that did jack shit for a long time, except allowing one dickhead company to pay off a developer for shit like Crysis 2 underground Tessellation on non rendered water, overabundant tessellation on concrete barriers, shitty looking rock roads in demos, and specifically porting a OpenCL piece of software back to a in house proprietary code so that on AMD systems the call would have to go back through the CPU to make it seem like AMD drivers have issues...... well by the time this card is mainstream they will have found and fixed the bugs in DX12 and be on to a standard that no current cards support, making all this moot.

 

[R-T-B]

Erm, isn't that what I've said? Clearly existing GPU's aren't "enough" DX12 capable to deserve name DX12 GPU's.

Yes, they are. They only need to support DX_12_0 to be DX12 "compliant" and benefit from it.

The rest is optional standards crap, much like DX9 capability bits were.

EDIT: And quit trying to lock a thread just because there's speculation you guys. If there wasn't speculation 90% of tech sites would vaporize.

 

[Steevo]

FYI

http://physxinfo.com/news/12197/introducing-nvidia-hairworks-fur-and-hair-simulation-solution/

"
PhysXInfo.com: It was really a big surprise to see NVIDIA HairWorks utilizing DirectCompute API, while other NVIDIA hardware physics effects is typically exclusive to CUDA capable GPUs. What was the reasoning behind this decision?

Tae-Yong Kim: One of the goals of NVIDIA GameWorks is to solve hard visual computing problems in a way that balances implementation efficiency and time-to-market, runtime performance, and ease of integration. This requires choosing the right technologies, and sometimes that will lead to CUDA solutions, other times to DirectCompute, and other times to solutions using completely different approaches.

With NVIDIA HairWorks, the balance landed in favor of DirectCompute, partly because the simulation portion of the algorithm is a small part of overall runtime cost, which is dominated by rendering"

Nvidia was using DirectCompute with Hairworks until they discovered they could screw AMD users, and owners of their older series and hope no one noticed it runs like shit unless you have a new 9XX series GPU.

http://www.tomshardware.com/reviews/call-of-duty-ghosts-pc-performance,3683-8.html

 

[MxPhenom 216]

even if fury is a faster card, I may opt for nvidias 980ti... the reason I support amd usually is because I don't want them to go under... the core on the 980ti is amazing though, kidney anyone?????

The 980ti wont cure your intolerance to hackers though.

 

[GhostRyder]

Yea its gotten pretty sad, We know for fact what Nvidia has. Everything from AMD side has been super hype that it will be a super gpu that nothing can match, even though history says other wise. yea it does have HBM memory but just cause it has it doesn't mean performance scales 1:1 with it. There is a point were performance benefits level off. You can look at CPU performance with diff speed of ram as proof of that. As for where that of diminishing return is yet to be known.

You have 0 right to talk especially on the whole "History says otherwise" argument. Have you ever looked at actual benchmarks on this site recently because the GTX 770/680 versus R9 280X/HD 7970 and R9 290X versus GTX 780ti, because they show things under a different light. So give it up on the "History says otherwise" argument. While there is no proof of how good it is which everyone acknowledges (mostly) however you have no facts either...

Erm, isn't that what I've said? Clearly existing GPU's aren't "enough" DX12 capable to deserve name DX12 GPU's. Clearly they have to do something about it for the upcoming series. How far they'll go with DX12 features, only they know, but they certainly can't support only what current DX11.2 cards support...

Yea nothing is and will be for awhile. it honestly makes very little difference in my book because its going to be so long before DX12 is actually used.

Either way, what we need is some actual facts which will come hopefully very soon (16th will have something).

 

[Schmuckley]

After a quick skim;This thread is typical TPU.I will make a prediction: There will be many butthurt Nvidia fanboys when the AMD card releases.
PS:EVGA Titan-X is..I haven't checked lately but..around..
It is.. well-known the Nvidia fanboyness of TPU.
Myself,Whatever I can get the most performance out of for less than $300 is what I'm getting.
I am an EOL buyer..
780 Ti you say?

 

[HumanSmoke]

After a quick skim;This thread is typical TPU

Well if it wasn't before, I'm sure your post put it over the top.

Myself,Whatever I can get the most performance out of for less than $300 is what I'm getting.
I am an EOL buyer.. 780 Ti you say?

Get the best of both worlds. Buy a 290X and BIOS flash it to 390X. Sorted.

 

[arbiter]

Nvidia was using DirectCompute with Hairworks until they discovered they could screw AMD users, and owners of their older series and hope no one noticed it runs like shit unless you have a new 9XX series GPU.

Hairworks uses a standard in DX11, not nvidia's fault AMD's cards suck in doing it.

Have you ever looked at actual benchmarks on this site recently because the GTX 770/680 versus R9 280X/HD 7970

When gtx680 was released it beat the 7970 by a bit. AMD had to re-release the 7970 with clocks upp'ed to 1ghz to make it compete.

 

[Tatty_Two]

Hairworks uses a standard in DX11, not nvidia's fault AMD's cards suck in doing it.


When gtx680 was released it beat the 7970 by a bit. AMD had to re-release the 7970 with clocks upp'ed to 1ghz to make it compete.

To be accurate, according to this sites initial review by 6% across all resolutions, I acknowledge that "quite a bit" is subjective but 6% hardly falls within it in my subjectivity, to be fair slightly more in favour of the 680 than that in the lower resolutions, less than that (4%) in the higher. I do seem to recall though that when the 290/290X came out, despite them being acknowledged as running power hungry and hot an absolute legion of people, many from here went out and bought one.

 

[RejZoR]

I hate % values in mentions with graphic cards. They are absolutely pointless and misleading. 6% sounds like a lot, but at 60fps, it means a 3.6fps difference. And we all know this doesn't mean it universaly extrapolates into 3.6fps less as a whole. It just means that much of a difference with that exact scenario (usage of HairWorks) And then you know it's all BS.

Besides, how is it with TressFX which was the first feature to utilize this anyway?

 

[SonicZap]

When gtx680 was released it beat the 7970 by a bit. AMD had to re-release the 7970 with clocks upp'ed to 1ghz to make it compete.

Actually, while AMD did up the clocks on the 7970, most of the performance difference has come from AMD's so-called horrible drivers improving the performance of all GCN cards more than Nvidia improved Kepler's performance. Back in the day GTX 680 beat HD 7970, now the R9 280X beats the GTX 770 (which is an overclocked 680) by a fair margin while having 1 GB more VRAM. With current drivers even the original 7970 beats the GTX 770 (= overclocked 680), as you can see in GhostRyder's link. So, in this case AMD didn't really need to raise clocks to compete. They wanted to get the most powerful GPU crown and a higher price (= fatter margins) for Tahiti, and it was succesful. The 7970 GHz Edition was the most powerful single-GPU card until the Titan was released.

About tesselation performance, the R9 285 (Tonga) is decent at it, so Fury likely won't exactly suck in it either.

 

[RejZoR]

The problem is, ONLY Fury. The rest will be the same shit as almost 2 years ago. Yay...