Have you ever used Nvidia on Linux? It is a pain in the ass to properly get full hardware acceleration in Linux with an Nvidia card, where as Fire GL cards are amazing under Linux.

Yes, it's my main computer/os, but since the Linux/NVidia crisis few ago, things have changed (a lot). (BTW i have tried the two side of the force the green and the red).

And that Optimus crap likes being a nuisance no matter what. Windows has problems with it too. I mean come on I just install the damn drivers, it refuses to switch to the dedicated card. Then I run DRIVERLESS and it works fine. What!?

optimus is clearly an issue, and as stated on windows too 🙂 most of the time other, 3dr part driver can solve problem (it was this way i can make my AMD HD run nice on mint).

Only thing I'm going to say is Nvidia did huge rebrands with the G80 and G92, and they still sold a ton of those.

as AMD with HD rebranded in R... if you can sold longer product that cost nothing in dev... why don't you sell them 🙂 easy money 🙂 (if i where them i would do the same) vbetts 3/23/2015 10:18 AM EST - Please do not make posts in a row.

So.... nvidia does rebadging too (while still selling at a higher premium) and nobody bats an eye. AMD releases nearly twice the amount of GPUs at a time, roughly half of which are re-badges, and everyone loses their mind. But in the end - why do you care? The price point of the re-badges are still good, and driver stability will be better than they were when they were fresh architectures.

For everyone who thinks that Model ID correlation mean rebrands; read this. http://www.reddit.com/tb/3006o7

So.... nvidia does rebadging too (while still selling at a higher premium) and nobody bats an eye. AMD releases nearly twice the amount of GPUs at a time, roughly half of which are re-badges, and everyone loses their mind. But in the end - why do you care? The price point of the re-badges are still good, and driver stability will be better than they were when they were fresh architectures.

There were quite a few posts about the 770 rebrand when it came out here on Guru3D. Not to mention other news posts about lower end Nvidia cards being rebranded. I wouldn't say nobody bats an eye.

I highly doubt that the Pitcairn GPU will be the R9 370 chip. R9-370/X current 280/280X R9-380/X current R9-290/290X R9-390/X New cards It would be silly to have such a large jump from the R9-370 to the R9-370X/R9-380, if it were Pitcairn. We shall see.

Didnt we have a chart where the 380X was competing with the 980?

I completely missed out Tonga, the current R9-285 Speculation is rife that the actual tonga is a cut down chip. So it should now be : R9-370/X current 285/New 285X (Tonga/X) R9-380/X current R9-290/290X R9-390/X New cards (Fiji/X)

Only thing I'm going to say is Nvidia did huge rebrands with the G80 and G92, and they still sold a ton of those.

iirc, the G80 was a wicked card, great overclocker ( i had 2) the G92 was the die shrink of the G80, and that, was rebranded many many times. It was a damn good card though, so I understand why they built so many of them (and had to rebrand to not lose stock).

iirc, the G80 was a wicked card, great overclocker ( i had 2) the G92 was the die shrink of the G80, and that, was rebranded many many times. It was a damn good card though, so I understand why they built so many of them (and had to rebrand to not lose stock).

Oh hell yeah! My 9800GTX+ models that I had were both amazing!

At least this confirms that AMD cannot really [thread=5035136] magically redesign their graphics chips almost overnight[/thread] to be ready for Computex 2015 in June.

R9-370/X current 285/New 285X (Tonga/X) R9-380/X current R9-290/290X R9-390/X New cards (Fiji/X)

Had 28 nm to 20/14 nm transition worked as expected, that would be the case. AMD always introduce new architectures in the high end, where hot and expensive chips are not a problem, then gradually trickle it down to mid and low-level cards with respins and redesigns, so they can benefit economically from the improved process node. However this time even Intel suffered a major delay with 14 nm Broadwell/Skylake, after they've been rock solid in introducing new manufacturing processes for a a long time spending a considerable amount of money on research and testing. TSMC and GlobalFoundries had major problems with 28 nm chips and they haven't even shipped a 20 nm product yet, they're like 2-3 years late. So it would probably make sense for AMD to wait for a proper 20 nm / 14 nm process and spend their resources on a full range of redesigned chips, while using existing 28 nm designs for a couple more iterations.

That being said didn't the GCN 1.1 improvements come to the rebranded 2xx cards?

In what way? GCN 1.1/1.2 chips in R9 260/290 and 285 cards are new and have minor improvements in the GPU architecture, while rebranded cards use the older GCN 1.0 chips introduced with Radeon HD7000 series, and these have not been redesigned or updated - they are the same silicon.

In what way? GCN 1.1/1.2 chips in R9 280/285/290 cards are new and have minor improvements in the GPU architecture, while rebranded cards use the older GCN 1.0 chips introduced with Radeon HD7000 series, and these have not been redesigned or updated - they are the same silicon.

Well I guess my point was when the 280x came out everyone called it a rebrand, regardless to the fact that they did actually slightly modify the architecture. When I said rebrand I included the 280x in that term. Mostly because of how people viewed it. I wasn't really aware how far down the 2xx line the modifications went though, which is why I posed that in the form of a question. I haven't owned an AMD card since 4870X2.

The problem is if this is true; the 2xx Series was re-branded HD7000 Series, and if the 3xx Series is a re-branded 2xx series... then what on earth really? I mean I understand from a price point of view and business point of view why AMD is doing this, but the problem here is AMD is losing market share, and heavily; and it's because of mess-ups like this, even those who have been on the red side for over a decade like myself are now considering going green (because of a heap of issues). We know nVidia and AMD are not perfect (we know it costs more to go nVidia, and I've always loved the performance/price ratio of AMD) but it's time for AMD to step up and put itself in the game again, or risk doing something as stupid as the above; give a 10% to 30% performance increase, drivers lag behind, promising things and then not doing them... It's not going to get them their market share back. This is all rumors anyway, we shall wait and see.

Considering at the minute the prices for 290/290X cards have actually gone UP here (OCUK lists the 290 Vapor X at £275 at the minute), having a rebrand which would knock this down to £200 as a 380 Vapor X doesn't seem like a bad thing to me... EDIT - Quick check, at today's rate, that is $410.

Same here.

So.... nvidia does rebadging too (while still selling at a higher premium) and nobody bats an eye. AMD releases nearly twice the amount of GPUs at a time, roughly half of which are re-badges, and everyone loses their mind. But in the end - why do you care? The price point of the re-badges are still good, and driver stability will be better than they were when they were fresh architectures.

That not true, I bashed nvidia for the 770 as well as other people too. If these amd cards are rebrands expect more of the same.

At least this confirms that AMD cannot really [thread=5035136] magically redesign their graphics chips almost overnight[/thread] to be ready for Computex 2015 in June. Had 28 nm to 20/14 nm transition worked as expected, that would be the case. AMD always introduce new architectures in the high end, where hot and expensive chips are not a problem, then gradually trickle it down to mid and low-level cards with respins and redesigns. So 390 would be a new 20 nm chip, and 370/380 would be based on 280/285/290. However this time even Intel suffered a major delay with 14 nm Broadwell/Skylake, after they've been rock solid in introducing new manufacturing processes for a a long time spending a considerable amount of money on research and testing. TSMC and GlobalFoundries had major problems with 28 nm chips and they haven't even shipped a 20 nm product yet, they're like 2-3 years late. In what way? GCN 1.1/1.2 chips in R9 280/285/290 cards are new and have minor improvements in the GPU architecture, while rebranded cards use the older GCN 1.0 chips introduced with Radeon HD7000 series, and these have not been redesigned or updated - they are the same silicon. Having a new PCI\VEN&DEV is not an indication - these numbers are programmed on the ROM.

I amhoping we might have a surprise on the cards : intel couldn't get 14/16nm up and running as they liked, but GloFo and Samsung have. (I don't remember which is doing 14 or 16nm). GloFo might be able to fab a first gen 390/X, and still cover the costs, even with HBM. It would be likely a very expensive card though. ~700 minimum($/€). The 390's might be a good deal, if the yields are bad 🙂

The thing is that every rebrand is a slight redesign. Each rebranded gpu that gets put out gets tweaked so its better than its previous version. Both Nvidia and AMD do this. Everyone seems to forget this each time a new generation comes out though. And to the few who have been saying 'drivers are horrible blah blah blah" stop that has not been the case for about 2-3 years now AMD drivers are much better now. I am not a fanboy of their cards I got mine because it was cheaper than the alternative at the time. I have not been disappointed yet by it. I have heard more from my Nvidia user friends complaining about their drivers lately.

AMD is really taking rebadging to the extreme. No wonder they lose market share. I wonder when they release a new flagship card that way...

And you act like nvidia didnt. Hopefully your not blind with fanboyism and just didnt know. G92 is the most whored out gpu ever. 8800gt>9800gt>8800gs>9600gso>8800gts 512>9800gtx>9800GX29800gtx+>GTS 250.

I completely missed out Tonga, the current R9-285 Speculation is rife that the actual tonga is a cut down chip. So it should now be : R9-370/X current 285/New 285X (Tonga/X) R9-380/X current R9-290/290X R9-390/X New cards (Fiji/X)

I think you are going to be right on the money with this. Does this mean the 390 is the only card that will have "full" DX12 support?

The thing is that every rebrand is a slight redesign. Each rebranded gpu that gets put out gets tweaked so its better than its previous version. Both Nvidia and AMD do this. Everyone seems to forget this each time a new generation comes out though. And to the few who have been saying 'drivers are horrible blah blah blah" stop that has not been the case for about 2-3 years now AMD drivers are much better now. I am not a fanboy of their cards I got mine because it was cheaper than the alternative at the time. I have not been disappointed yet by it. I have heard more from my Nvidia user friends complaining about their drivers lately.

I am so sick of the crying and pure ignorance about drivers. I use both, and honestly have zero issues with both. Half luck, and half not being an ignorant idiot when using my PC. I have a R9 290 in my main rig, and GTX 650ti in my second rig. Im with you on the price point. I paid $240 for a non reference 290. Add in $80 for water cooling parts, and its still barely the price of a reference gtx 780 3gb and faster. No way was I paying $200+ for a GTX 680 2GB. I want a GTX 660 2GB for my second rig. Cause I use it for game streaming to my S4. I can get a HD 7950 3gb for the same price that walks the 660 bad, but doesnt have the needed nvidia tech sadly.

Sorry but I don't believe any of this proves that the 300 series will be rebrands. AMD has stated they won't be and that is what I believe. Even the 200 series cards are not straight rebrands either. They all have improvements over and perform better than the earlier cards even if the main chip has the same specs.

I highly doubt that the Pitcairn GPU will be the R9 370 chip. R9-370/X current 280/280X R9-380/X current R9-290/290X R9-390/X New cards It would be silly to have such a large jump from the R9-370 to the R9-370X/R9-380, if it were Pitcairn. We shall see.

Agreed, and my HD 7850 2GB is also Pitcairn...;) (Stock clocks to 1GHz without difficulty.) Do you notice how some folks reading the thread are behaving as if HH's speculation here (which HH freely admits to--nothing wrong with that at all) is an actual announcement from AMD? Why people invest so much in speculation is beyond me...beginning to think people are too lazy to read these days...;)