300 series Rebadge & Refresh confirmed

[Silverforce11]

Tonga, still the cut down since full-dies => Apple.

This is the fail right here, lowest SKU, still needs 1x6 pin.

That's really all we were awaiting on to confirm a straight rebadge, no major power usage change, just a bump in clocks and more vram which is useless for the performance class.

They are seriously expecting these 380/390/X to actually sell for $199, $329 and $429, against the 960/970/980 which clobbered R280/X/285 & R290/X and dragged marketshare down towards 20%. Must be out of their minds.

It's clear they had no plan to invest in refreshing the entire lineup, this renaming and extra vram is simply to keep AIBs happy with new SKUs and "moar vram" gimmick salespoints.

I sure hope they will have plenty of Fury Nanos to capture the 970/980 segment cos Hawaii with 275W TDP with 8GB sure as heck won't cut it.

Edit: It appears the 380 and 390 series are refreshes while the 370 is a rebadge.

http://nl.hardware.info/reviews/613...90x-review-nieuwe-line-up-met-bestaande-chips

With such gains at similar TDP, its wrong to label 390/X a rebadge.

 

[railven]

Tonga, still the cut down since full-dies => Apple.

This is the fail right here, lowest SKU, still needs 1x6 pin.

That's really all we were awaiting on to confirm a straight rebadge, no major power usage change, just a bump in clocks and more vram which is useless for the performance class.

They are seriously expecting these 380/390/X to actually sell for $199, $329 and $429, against the 960/970/980 which clobbered R280/X/285 & R290/X and dragged marketshare down towards 20%. Must be out of their minds.

It's clear they had no plan to invest in refreshing the entire lineup, this renaming and extra vram is simply to keep AIBs happy with new SKUs and "moar vram" gimmick salespoints.

I sure hope they will have plenty of Fury Nanos to capture the 970/980 segment cos Hawaii with 275W TDP with 8GB sure as heck won't cut it.

Especially now @ $330 and $430 price points. THey were struggling to sell when they were at <$250 $300 each.

If they are confident about Nano, probably should slot that at $450, and give the other cards a $100 bump down.

390X with 8GBs sound waaaaay more appealing at $330. And just the gimmick of "new" to the uninformed 390 w/8GB @ $230 be super popular. I mean, if they're just using overstock chips it's not like they're paying up front for inventory.

 

[Silverforce11]

Nano definitely would be a winner at $449, compared to Fury and X prices. Just the efficiency and form factor itself is a good justification for potentially 980 + 20% performance (Captain Jack anyone?).

The problem with that is they won't have anything competitive below that and the $200-$350 is a huge sweetspot. It's where the 960 and 970 did the major damage to AMD.

If they have lots of volume, Fury Nano at $399 makes more sense, it could entice a lot of gamers who are planning to spend ~$339-349 for a 970 to pay a bit more and get heaps more performance at similar power usage.

 

[raghu78]

you guys are damn right. these R9 390 and R9 390X cards are just going to pile up at retail. AMD is not serious about selling these cards at these prices. This is a stop gap money grab attempt to fool the buyers who are not well informed. Once Fury Nano launches at USD 429-449 I expect AMD to revise R9 390X to USD 349 and R9 390 to USD 299.

Nano on the other hand is going to be a damn exciting card. My guess is its a heavily salvaged Fury chip for maximizing yields. 25% CUs are disabled per shader engine. I would guess 4 x 768 = 3072 sp and 850-900 Mhz at stock with 175W TDP. The chip will have the full 4 shader engines, 4 tesselation engines and 64 ROP, 4 GB HBM. Its going to be an excellent competitor against GTX 980 with all the latest architectural enhancements to performance and efficiency and that blazing fast HBM. :thumbsup:

 

[Dribble]

It is looking pretty dire, trying to continue to sell cards that already weren't really selling at lower prices before the rebadge. That probably 98% of their market that is going to continue to go downhill. There must be so many of these cards kicking around now that most people who want one already have them, and the rest can probably pick up an old one on some amazing deal. The OEM's weren't selling these before the badge and are just going to continue not to sell them.

 

[Stuka87]

I am wondering if the Fury Nano will actually replace the 390X entirely. And the 390X is only a stop gap for a few months, which is why nothing was really done to it outside of more memory.

 

[Flapdrol1337]

No full tonga makes me a saaadd panda

 

[Creig]

Tonga, still the cut down since full-dies => Apple.


This is the fail right here, lowest SKU, still needs 1x6 pin.


That's really all we were awaiting on to confirm a straight rebadge, no major power usage change, just a bump in clocks and more vram which is useless for the performance class.

They are seriously expecting these 380/390/X to actually sell for $199, $329 and $429, against the 960/970/980 which clobbered R280/X/285 & R290/X and dragged marketshare down towards 20%. Must be out of their minds.

It's clear they had no plan to invest in refreshing the entire lineup, this renaming and extra vram is simply to keep AIBs happy with new SKUs and "moar vram" gimmick salespoints.

I sure hope they will have plenty of Fury Nanos to capture the 970/980 segment cos Hawaii with 275W TDP with 8GB sure as heck won't cut it.

So much drama... We don't know if any improvements have been made to the GPU of the 300 series yet. Therefore, your pricing doom predictions are meaningless. We need to wait for detailed reviews to come in before anyone can make informed decisions regarding pricing.

 

[Braznor]

I think of this AMD generation as a trial run for HBM. They sure don't have the cash to redesign chips at every price point and that too with two differing memory technologies at the same time.

The biggest benefit this generation will bring to AMD's stable is the experience in integrating HBM into their product lineup. Plus with all the design changes made to the Fury X/Pro/Nano chips, AMD is on a much stabler path on the roadmap towards its next generation of products. In contrast Nvidia has to deal with two great unknowns in its next generation Pascal, the 14nm process and HBM at the same time. AMD played it smart technology wise by crossing one of these risky moves right at this point. This leaves AMD one less thing to worry about.

I don't think AMD itself considers sales for the rest of the 300 series to bring anything to its table at this point. At best, these products are meant in a serious manner only for OEMs and oh, Apple. This particular generation for AMD is all about regaining its name as a serious and premium technology brand under the flagship of the Fury.

This is a very good thing for informed consumers like us in this forum. The next generation of AMD's cards in all price tiers should be even more exciting.

 

[tviceman]

I sure hope they will have plenty of Fury Nanos to capture the 970/980 segment cos Hawaii with 275W TDP with 8GB sure as heck won't cut it.

No one knows for sure how well the Fury Nano will actually do in real world games. Hell, no one even knows how for sure how Fury X and Fury lite will do in games yet. When companies tout their product improvements, they almost always state best case out-of-the-norm scenarios. Hence when Nvidia intro'd the gtx 750 TI, they compared it to cutdown Kepler sku's to make it look better and were still slightly fudging their 2x perf/w increase. When Nvidia intro'd GTX 980 they compared it to GTX 680 (instead of the GTX 770, which was the same chip but had better avg. perf/w), and again they "fudged" the perf/w figures by rounding up from the avgerage ~1.66x to 2x increase. AMD did the same when they released their past dual GPU cards, claiming 2x perf over Nvidia's current best.

Lets actually wait and see how in-game results pan out. I'm sure the improvements are substantial but CEO's are supposed to hype up their company and products (which Lisa Su did a great job of with Fury, btw. Much better than bore-fest Huddy and even JHH).

 

[Gikaseixas]

So much drama... We don't know if any improvements have been made to the GPU of the 300 series yet. Therefore, your pricing doom predictions are meaningless. We need to wait for detailed reviews to come in before anyone can make informed decisions regarding pricing.

Fully agree
We have to see reviews to see how much improvement vs older R9 200 series AMD was able to get. Remember that both AMD and Nvidia align performance with prices

 

[happy medium]

You know they are straight rebrands when the topic of conversation turns to the only bright spot, the released in a few months Nano.

 

[LTC8K6]

Did TSMC do a new 28nm die or something? The specs and the boards appear identical.

Any performance improvements other than clock speeds will surprise me.

 

[LTC8K6]

Yeah, Nano is the problem for 390X.

There's no reason to even look at 390X unless you need a card right now.

 

[exar333]

Yeah, Nano is the problem for 390X.

There's no reason to even look at 390X unless you need a card right now.

Totally agree. Unless the 8GB is required, the Nano is hands-down the better choice unless prohibitively more $$$.

 

[SPBHM]

why is the memory "up to XXXXMHz" on the 380 and lower, and not a fixed clock like on the 390s? they going to allow versions with slower memory?

also the 360... is not even the full GPU, like the 7790 was lol (at least clocks are higher)

 

[exar333]

On the plus side, AMD consolidated the 280/285/280x into the single 380. Three SKUs all performing closely was silly. This makes more sense.

The 390 would be better, IMHO, with just 4GB ram. Maybe cut the price to $299 and lower TDP a bit too. That would help differentiate better between that and the 390x. With current prices, at $100 more, who would seriously buy the 390x? It is not quite as bad as the 970 vs 980 on the NV side, but close.

 

[Subyman]

I would have thought AMD would have put a better cooler on those instead of the same blower with a new shroud. Hopefully there is a major overhaul of the reference cooler under the new plastic or they will be in the same situation they were in with the original R9 290 and R9 290X reviews where the cards ran at 80C+ and were loud. Those reviews hurt them, so hopefully they learned from it.

 

[n0x1ous]

I would have thought AMD would have put a better cooler on those instead of the same blower with a new shroud. Hopefully there is a major overhaul of the reference cooler under the new plastic or they will be in the same situation they were in with the original R9 290 and R9 290X reviews where the cards ran at 80C+ and were loud. Those reviews hurt them, so hopefully they learned from it.

I fully expect you will never see one of those new blower shrouds in the wild or in reviewers hands. All the cards showed on stage yesterday were aftermarket cards for the non Fury stuff.

 

[bryanW1995]

Why does every assume that nano will close in price to 390x? They are completely different cards with completely different aims. I'd guess nano will be priced more like Fury Pro, say ~ $500. Besides which, it won't be out for several months at least, whereas the rest of the lineup is either available now or will be soon.

 

[RussianSensation]

Technically 390/390X are refreshes like 680-> 770, not a re-badge since official GPU and memory clocks do go up. Re-badge/rebrand would mean exact same specs, just a different name on the box. That aside, $329 and $429 pricing is too much. Smart mid-range buyers will be hunting for $230-240 290 or $270 XFX 290X on Newegg or even a $290-300 970. Once those cards run out though, AMD will need to drop prices. Worldwide, the situation will be worse for AMD since a lot of gamers said 290/290X didn't cost substantially less for them than a 970. That means 390/390X will be priced higher than 970 in many countries. Also, once R9 290 runs out, the $150-250 buyer will be stuck with 960 or 285/380 level card - both garbage for the $. Awful situation this gen for budget gamers. This is the only generation I can think of where a $150-250 space will be an automatic skip straight to the 970 because I just can't recommend a 960 or 380 2GB or spending $230+ on a 4GB version of these GPUs. AMD could try to differentiate by pricing 380 4GB at $199. That would put a lot more pressure on the 960 2GB card.

 

[Vesku]

Either AMD's roadmap is to spend the money saved to beat Nvidia to the node shrink by ~6 months or they have made a very bad cost cutting decision with the 300 series.

 

[RussianSensation]

Either AMD's roadmap is to spend the money saved to beat Nvidia to the node shrink by ~6 months or they have made a very bad cost cutting decision with the 300 series.

There are other possibilities. They could have needed that money for 14nm HBM2 GPUs overall. It might not even be about beating NV to the node shrink by 6 months but making sure the next gen has sufficient funding. Financially speaking, it could have resulted in more negative cash flow if they had invested in designing a top-to-bottom stack of new GPUs and then NV just lowered prices a bit. This would have been disastrous - spend $250 million+ to redesign each new stack and then lose even more $ from NV's aggressive marketing/price drops. Right now AMD spent minimal marketing and R&D budget on R9 300 series which means Lisa Su probably focused on profitability. Think about it, if it were to cost them $100 per each new fully redesigned GPU in R&D, and they profit $80, they are losing $20. If they spent $10 on marketing for R9 300 series and they profit $20 only from a sale, they are making $10. Lisa Su probably calculated the rate of return on investment for resigning an entire stack on GDDR5 and the CFO/financial guys kept telling her it would be impossible to recoup the cost of investment and profit. I don't think companies make these types of decisions so lightly and if Lisa Su actually figured it's financially safer to go this route, it's akin to Nintendo making $ off each Wii U sold and the Wii U games, which is really all they need to try again with their next console. Sometimes the strategy of making a little bit of money on outdated hardware coast to next gen while allowing the firm to set aside the key resources for next gen is the exact direction many companies choose instead of investing more sunk costs into old tech.

 

[Vesku]

Porting to GF 28nm isn't 100s of millions of dollars, this is a <$100 million cost savings decision. Most likely a few $10s of millions or even less. Now there may be some issues regarding GF capacity and not wanting to completely abandon TSMC (even though they publicly tell their shareholders they plan to go full GF). They already did the heavy lifting engineering when they ported the GCN units on their APUs to GF 28nm.

Imo a very bad decision unless they are confident they will have the next series ready in 12 months or less. Or perhaps TSMC 28nm is cheap enough they plan to push Fiji lower and lower down the stack as the months tick by, unlikely given HBM cost and availability.

 

[Stuka87]

Why does every assume that nano will close in price to 390x? They are completely different cards with completely different aims. I'd guess nano will be priced more like Fury Pro, say ~ $500. Besides which, it won't be out for several months at least, whereas the rest of the lineup is either available now or will be soon.

Because we were told the performance is equivalent to 290X but at a much lower TDP. I highly doubt it will be priced at 500 if you can get a much faster Fury for ~550.