R9 300 cards listed in new driver - R9 370 is a rebrand

[Phynaz]

I will remind everyone that AMD have stated that every new GPU going from 2015 onwards will have FreeSync.

Freesync cannot work with Tahiti and Pitcairn, i dont believe they will rebrand 270 to 370 without Freesync in 2015.

I remind you that a rebrand isn't a new GPU. Don't fall for marketing spin.

 

[Cloudfire777]

I'm no AMD fan but what you've said in the last couple of posts is garbage.

Here's a place to start educating yourself. Do these chips sound like they are the same to you :

http://www.extremetech.com/gaming/1...cn-1-2-torpedo-that-takes-out-nvidias-gtx-760


"The R9 285 has a 359mm2 die and 5 billion transistors, compared to 352mm2 and 4.3 billion transistors for the old Radeon HD 7970/R9 280X. AMD has confirmed to us that the new chip has 32 CUs total (with 28 functional), implying that AMD has the option to bring a new version of the core out if it chooses to do so — but even the additional CUs don’t explain why a GPU with a smaller memory bus is both larger and more dense than its predecessor."

If you don`t have any arguments, you shouldnt resort to low blows like that.

Transistors density does not equal a new architecture. Its an improvement of fab process and a change in chip design. If you bothered to look at the slide I posted a few posts before your post, AMD say they focus on transistor density. Not just from GloFo but also their APUs. You don`t have to go back further than to Kaveri to see the density skyrocket compared to previous designs. AMD seems to be fond of getting the chips more denser. Even with Carrizo they bragged about 30% more transistors on the same die size as Kaveri.

Color compression is a step in the right direction, but the change is so small that its a revision rather than a new architecture. Which is why AMD call Tonga etc GCN 1.x, not GCN 2.0.
Did you know that there are certain parts where R9 280 beats R9 285 in tessellation btw?

Other than that, small changes like XDMA and TrueAudio again does not equal to a new architecture.

A new architecture is when you start changing processor layout, power consumption and efficiency, change in instruction sets.

AMD is still at the same architecture they were at since Radeon 7970 back in 2012. Adding communication through PCIe lanes instead of a dedicated crossfire cable, and adding a feature like a dedicated audio engine does not make the architecture any different. Feature wise its better than their first products, no doubt.

 

[Shehriazad]

I remind you that a rebrand isn't a new GPU. Don't fall for marketing spin.

But we still don't know if they're full rebrands?!? Can't it be possible for them to for example take Pitcairn and just slightly improve it?

Which would explain why that 370 packaging leak was showing 110w TDP.

Lowered TDP, added freesync and maybe some mhz tweaking....but it would still be a "Pitcairn". That would be a way to recycle effectively without having to waste tons of R&D funds.


But then again...I don't work in that industry, so I'm not sure how likely that is.

 

[Cloudfire777]

It shouldnt take long until we see the R9 370 I bet. Launch was in April so we should see the real deal soon.

That XFX 370 is coming in April while 390X is in June smells rebrandy to me

 

[SimianR]

Regardless of when we see these cards, I'm not sure it's worth debating about rumors or device ID's in drivers until we have the final word from AMD.

 

[Silverforce11]

But we still don't know if they're full rebrands?!? Can't it be possible for them to for example take Pitcairn and just slightly improve it?

Which would explain why that 370 packaging leak was showing 110w TDP.

Lowered TDP, added freesync and maybe some mhz tweaking....but it would still be a "Pitcairn". That would be a way to recycle effectively without having to waste tons of R&D funds.


But then again...I don't work in that industry, so I'm not sure how likely that is.

That's what they've been doing, minor iterations of GCN.

If they say their 2015 lineup has FS, it means they need to add support to it even if they "re-brand", as such, the die itself has to be tweaked.

Pitcairn with FS + Tonga architecture improvements, without the compute focus design of its bigger brothers, should be very efficient.

 

[shady28]

If you actually took the advice to educate yourself with that link, you'd not be saying yet more indefensible things.

If you don`t have any arguments, you shouldnt resort to low blows like that.

<snip>

A new architecture is when you start changing processor layout, power consumption and efficiency, change in instruction sets.

You mean like these kinds of things, from the linked article ?

"Support for 16-bit floating point and integer values. This is advantageous in low power settings, where 16-bit precision is more than good enough. Up until now, AMD simply used 32-bit values for all calculations — switching to 16-bit in specific contexts will save power and internal bandwidth without noticeably degrading image quality. AMD also notes that this new Tonga GPU supports sharing data between SIMD lanes and has an improved task scheduler."



"Color compression. Frame buffer data is now stored in a lossless compressed format and the GPU can read and write compressed data. This leads to substantial bandwidth efficiency gains and allows AMD to shrink the memory bus from 384 bits to 256 bits.
"


the modified GCN front end that AMD introduced with the Hawaii family. That means eight Asynchronous Command Engines (up from two) and support for four primitives per clock cycle (everything but the R9 290 cards top out at two primitives per clock).

AMD is still at the same architecture they were at since Radeon 7970 back in 2012. Adding communication through PCIe lanes instead of a dedicated crossfire cable, and adding a feature like a dedicated audio engine does not make the architecture any different. Feature wise its better than their first products, no doubt.

As referenced and linked to multiple times, this is not the same arch.

Tonga is as much a new arch as Maxwell is.

 

[boozzer]

Regardless of when we see these cards, I'm not sure it's worth debating about rumors or device ID's in drivers until we have the final word from AMD.

ha, it wouldn't be a smear campaign if they follow something sensible like that

 

[Cloudfire777]

You didnt have any arguments that proved it was a new architecture. You copy pasted things from their review, and reposted the things I refuted on my earlier post. You are seriously comparing Tonga and Maxwell?
wow...I think we will quit there.


------------------------------------------------------------------------

Anyone who knows what the blurred picture is of?
I dont think Ive seen this slide before

 

[Enigmoid]

I added more info to my original post. Please read it again. The iMac Retina itself throttles, both the CPU and the GPU. The cooling is insufficient and it's Apple's fault for choosing a 125W mobile part to integrate into a chassis built for an 85-100W one. Tonga XT is not to blame here. It's not like AMD advertised it as an 85-100W part but it uses 125W. AMD was transparent about what the chip was from day 1. It's 100% the manufacturer's responsibility to make sure they choose the appropriate product for a specific chassis/cooling setup. Apple is 100% to blame for this.

Yes, but that does not work around the fact that it is a 125W GPU and that 125W mobile GPUs tend not to be used by anyone. While it is apple's fault, its quite obvious that this chip is trying to be shoehorned somewhere that it doesn't fit too well (mobile) without being cut down.

It doesn't compete well. Apple should have added more cooling but its a very power hungry chip.

Seriously, after countless threads that discussed/proved Tonga's GCN 1.2 architecture's improvements over GCN1.0/1.1, people still think Tonga is the exact same architecture as a 280X? Tonga's problem is its both TMU and SP limited, which is why in 99% of games its architecture's advantages don't even show up. Take Tonga's advancements and put them inside a 4000+ SP and 256 TMU chip, and it's going to show up, guaranteed.

If we blow up Tonga by 2x it will still be SP and TMU limited. Again, we don't care about the 1%, we care about how well it does on average.

Its a mild IPC gain (comparing to the 280X). Full Tonga would likely compare favourably to the 280X (+10% would put it at identical performance) as it would be slightly bandwidth limited. But Tahiti is the worst performing gcn card on a per shader performance basis.

R9 285 1080p and 2160p index beats the R9 280X :

http://www.tomshardware.com/charts/...tml?prod[7297]=on&prod[7279]=on&prod[7276]=on


2160p R9 285 vs R9 280X vs R9 280X :68.81 65.71 64.60

1080p : 78.41 72.70 71.85


This is with 1792 SPs. With 2048 like the M295X it should be another 10% or so faster.

And no it doesn't win on every game. It depends on the game / benchmark. But there are enough where it wins that the aggregate (overall) is better.


Biggest gripe people had with the R9 285 was the 2GB. That negatively impacts some newer games and games at higher resolutions. That is an artificial limit though, its probably a marketing decision.

That a 256bit 2GB card can walk over their older 384bit 3GB card on anything should be saying something (including things like - this isn't the same architecture).

You are comparing to two 280s.

 

[Cloudfire777]

OH SNAP.
From VR-Zone (reliable source)

&#26681;&#25818;&#19968;&#20123;&#32819;&#35486;&#65292;AMD Radeon R9 390X&#65288;&#26283;&#23450;&#65289;&#65292;&#20063;&#23601;&#26159; Fiji XT &#34920;&#29694;&#26371;&#27604; GeForce GTX 980 &#24375; 20%&#12290;&#22914;&#26524;&#23526;&#38555;&#34920;&#29694;&#30495;&#26377;&#22914;&#27492;&#35480;&#20154;&#65292;&#23565;&#26044;&#28040;&#36027;&#32773;&#32780;&#35328;&#65292;&#20284;&#20046;&#20063;&#26159;&#20214;&#19981;&#37679;&#30340;&#20107;&#24773;&#12290;

According to some whisper, AMD Radeon R9 390X (tentative), which is Fiji XT performance will be stronger than the GeForce GTX 980 by 20%. If the actual performance really so attractive, for consumers, but also pieces seem good thing.

http://chinese.vr-zone.com/146560/a...inidad-and-tobago-for-computex-2015-03232015/

 

[firesword2007]

Pretty dissapointing

 

[Gloomy]

OH SNAP.
From VR-Zone (reliable source)

http://chinese.vr-zone.com/146560/a...inidad-and-tobago-for-computex-2015-03232015/

Meh, Titanix managed 33%.

 

[SolMiester]

OH SNAP.
From VR-Zone (reliable source)

http://chinese.vr-zone.com/146560/a...inidad-and-tobago-for-computex-2015-03232015/

That sounds okay!...I never believed Fiji would best Titan anyway, but 20% more than 980 at same price would be a winner!

 

[Silverforce11]

OH SNAP.
From VR-Zone (reliable source)

http://chinese.vr-zone.com/146560/a...inidad-and-tobago-for-computex-2015-03232015/

More reliable than the same leaks (which VR-Zone didn't even have any info about before its launch!) that put Titan X at ~30-35% faster than 980, with very accurate power levels at ~780ti (which some NV fans refused to believe, thinking Maxwell is some miracle in efficiency!)? Same leaks have 390X faster than Titan X. -_-

Selective believe & dissing based on rumors/leaks is... selectively silly?

 

[tolis626]

That sounds okay!...I never believed Fiji would best Titan anyway, but 20% more than 980 at same price would be a winner!

As much as I'd like to see AMD do well with something they release, what you said is not true at all. In June it will be what? 9-10 months since the 980 came out. If by that time they only manage a 20% improvement, no matter the price, I call it a failure. No, it should trade blows or even beat the Titan X if they want to have SOME success. And that's for the high-end only. With everything else being a rebrand (Again, I don't believe they can be THAT stupid, but they can always suprise me), they are gonna have a baaaaad time. And it's not just that their "new" GPUs won't sell, it will destroy their brand image even more. That's what they can't afford. They've got to convince the audience that their products are at least of equal quality/performance to NVidia. Or better yet, that they're better. Selling well or badly in one particular generation of cards is meaningless compared to the bigger picture.

Also, has anyone noticed a shortage of Hawaii/Tonga/Tahiti etc? If yes, that could be a sign that they aren't made at large quantities anymore. That's a long long shot, but it could be something. Many Hawaii models haven't been available for sale here for the past couple months.

Other than that, while I wouldn't really be surprised if the 390x crashes the Titan X or if it only trades blows with the 980, that's not really what I care about. What I do care about, and what all of you should care about, is whether we're gonna see this continue. I mean, Maxwell is good and all, but Pascal in 2016, on a 14nm(or is it 16nm?) process, with HBM2 and all that other stuff is going to be serious stuff for NVidia. I expect a performance jump quite bigger than Kepler-Maxwell. AMD will desperately need something to compete with Pascal. If they do have a 400 series ready in time, that is also competitive with Pascal, I am ready to forgive whatever crap they come up with now, especially seeing as they are in a period of restructuring. But sadly, I don't think that that will be the case. I can see the gap between the two companies grow larger. I hope I'm proven wrong, because seeing AMD go to hell would be a bad, bad, BAD thing for the industry. And before some NV-fanboy says "Good riddance" to AMD, I would like to ask the question, how does a possible 800$ mid-range card sound? Or a 1500$ high-end card? Disposable income only goes so far even for the more wealthy among us. And seeing as I'm not one of the wealthy ones, I would be among the first to go away from PC gaming. Then the guys who now own CF/SLI setups will go slowly. And then what? Nope, AMD has to keep up for the foreseeable future.

 

[Silverforce11]

Also, has anyone noticed a shortage of Hawaii/Tonga/Tahiti etc? If yes, that could be a sign that they aren't made at large quantities anymore. That's a long long shot, but it could be something. Many Hawaii models haven't been available for sale here for the past couple months.

In my region yes, its definitely a winding down and clearing of inventory.

285 in particular has rarely been available at all since launch, suggesting the harvested Tonga die that don't make the cut for M295X is a small proportion.

Either way, looking forward to the new launch. I am optimistic, because if AMD performs, NV will lower their prices, and we all win. When the time comes to upgrade, I'll weigh in both side pros (GameWorks) & cons (NV $ Premium).

 

[JDG1980]

I remind you that a rebrand isn't a new GPU. Don't fall for marketing spin.

In the past, AMD has been "very explicit" in insisting that their rebadges aren't rebadges. Furthermore, everything we've heard indicates that the 370-series GPU will be called "Trinidad", not one of the previous names. No matter which way you slice it, if AMD were to bring out another GCN 1.0 release as the R9 370, it would be a violation of their promise for FreeSync on all new products. More to the point, it wouldn't make any sense; they just can't get away with this kind of lateral rebrand, especially with the competition from Maxwell. I suspect we're either dealing with junk in the INF files, or these are OEM cards which have completely different specs from retail cards (it wouldn't be the first time either GPU vendor did that).

 

[DarkKnightDude]

Supposedly he was referring to the rebrands and other speculation about releases.

Basically we'll see if its true or not when its released.

 

[JoJoman88]

The 390X thru 380 will be the same chip with lower chips being disabled/defective Fuji GPU's. The rest will be reworked current GPU's, added features, voltage tweaks, ect. What's so suprising about any of this?

If the 390X is 10 to 20% off a Titan X but it's 15 to 20 % cheaper, then it's a trade off. Lets see where the price falls before saying it's DOA. If it's 25 to 30 % slower than a Titan X but only 15 to 20 % cheaper, then Houston we have a problem.

Having much room left in my PSU, I have been waiting for the 290X's to drop. I want the exact same GPU for Crossfire but the price for it is floating back up. What is up with that?

 

[desprado]

The 390X thru 380 will be the same chip with lower chips being disabled/defective Fuji GPU's. The rest will be reworked current GPU's, added features, voltage tweaks, ect. What's so suprising about any of this?

If the 390X is 10 to 20% off a Titan X but it's 15 to 20 % cheaper, then it's a trade off. Lets see where the price falls before saying it's DOA. If it's 25 to 30 % slower than a Titan X but only 15 to 20 % cheaper, then Houston we have a problem.

Having much room left in my PSU, I have been waiting for the 290X's to drop. I what the exact same GPU for Crossfire but the price for it is floating back up. What is up with that?

Rumor are suggesting that R9 390X is just 20% faster GTX 980 than let alone GTX Titan X.

 

[Enigmoid]

I expect relative parity with the TitanX. But we shall see.

 

[JDG1980]

While GCN 1.0 definitely has to die, it's worth pointing out that GCN 1.1 could potentially be more competitive against Maxwell than is commonly realized. This TechReport comment makes for interesting reading. For a long time, AMD has had the bad habit of cramming way too much voltage into its GPUs, and pushing the clock speeds further than efficiency would dictate. Now that Maxwell has put power consumption back in the spotlight, perhaps they will reconsider. This poster managed to make Hawaii (R9 290X) run as low as 145W by dropping the clocks to around 850 MHz. Is the little bit of added performance from boosted clocks really worth double the power consumption?

Regarding the rumors that R9 390X will only beat GTX 980 by 20%, I wonder if these rumors were true at one point, and that's why Fiji's release was delayed. Perhaps AMD wanted to make sure they had a product that could beat the Titan X, and decided to do some minor tweaks (which could be as simple as factory overclocking or tighter binning on the watercooled model) to ensure that they would get the single-GPU performance crown before going forward.

 

[Silverforce11]

Regarding the rumors that R9 390X will only beat GTX 980 by 20%, I wonder if these rumors were true at one point, and that's why Fiji's release was delayed. Perhaps AMD wanted to make sure they had a product that could beat the Titan X, and decided to do some minor tweaks (which could be as simple as factory overclocking or tighter binning on the watercooled model) to ensure that they would get the single-GPU performance crown before going forward.

We've had multiple leaks from various sources (including developers on twitter & journalists) which claims 390X is a lot faster than R290X and even faster than Titan X a long time before this VR-Zone leak comes out with 20% more than 980.

So the range is pretty wild out there in rumor-land. Anything from 20% above 980 to 50% and demolishing Titan X. It's fun to speculate and that's about all it should be.

ps. My R290s @ 800mhz and undervolted is around 150W ea (manually), I haven't tested it via Power Limiting in CC.

 

[geoxile]

OH SNAP.
From VR-Zone (reliable source)

http://chinese.vr-zone.com/146560/a...inidad-and-tobago-for-computex-2015-03232015/

Have these been debunked:
http://videocardz.com/55146/amd-radeon-r9-390x-possible-specifications-and-performance-leaked