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

[JDG1980]

You have to be very careful here. Maximum could be the 100 microsecond blip depending on the granularity of the measurement.

In the same review the average power is 112W. AMD could easily counter with a SKU that drops clocks under very heavy loads to maintain a 130W envelope.

If you look at the top of the page, it doesn't appear that these were momentary blips. The description is as follows: "Maximum: Furmark Stability Test at 1280x1024, 0xAA. This results in a very high no-game power-consumption that can typically be reached only with stress-testing applications. The card was left running the stress test until power draw converged to a stable value."

It's worth pointing out that the R9 series rebrands were more power-hungry, in general, than their 7000-series counterparts. For example, the R9 270X's maximum power draw of 172W compares poorly to the 7870's maximum power draw of 144W. (The 7850 was even more frugal, barely breaking the 100W mark even with Furmark.) I don't know what happened here - did AMD just want their "new" cards to chart a bit higher, so they bumped the clocks even though efficiency went down the drain? Even the Tahiti-based 7950 wasn't that bad in terms of efficiency, maxing out at 179W, well short of the rated TDP of 200W. Even Tonga doesn't do that well.

As I said, I agree that AMD could get Pitcairn's TDP down to 110W-130W without butchering performance. As noted above, that's pretty much what they did with the 7850, and that was back in 2012. But I still don't think this is going to be a competitive product in 2015, not when it lacks modern features like FreeSync, Crossfire XDMA, and 4K video decoding.

 

[3DVagabond]

If you look at the top of the page, it doesn't appear that these were momentary blips. The description is as follows: "Maximum: Furmark Stability Test at 1280x1024, 0xAA. This results in a very high no-game power-consumption that can typically be reached only with stress-testing applications. The card was left running the stress test until power draw converged to a stable value."

It's worth pointing out that the R9 series rebrands were more power-hungry, in general, than their 7000-series counterparts. For example, the R9 270X's maximum power draw of 172W compares poorly to the 7870's maximum power draw of 144W. (The 7850 was even more frugal, barely breaking the 100W mark even with Furmark.) I don't know what happened here - did AMD just want their "new" cards to chart a bit higher, so they bumped the clocks even though efficiency went down the drain? Even the Tahiti-based 7950 wasn't that bad in terms of efficiency, maxing out at 179W, well short of the rated TDP of 200W. Even Tonga doesn't do that well.

As I said, I agree that AMD could get Pitcairn's TDP down to 110W-130W without butchering performance. As noted above, that's pretty much what they did with the 7850, and that was back in 2012. But I still don't think this is going to be a competitive product in 2015, not when it lacks modern features like FreeSync, Crossfire XDMA, and 4K video decoding.

It was indeed clocks. AMD had better hope they aren't going to have to release a cut down Pitcairn again at the same clocks they used on an immature node when it was first released.

 

[boozzer]

So, is there any reason for me to wait or can I just pick up a 290 now without missing out on anything? -_-

you really should take anything posted by someone who wants amd to die with a truck load of salt. wait for bench marks and reviews. 370 is rumored for mid april release right? just 20 more days.

 

[Blue_Max]

Need a new video card very soon... was so hoping there'd be something new to offer stiff competition.
[Holding on to fading hope...]

 

[Gikaseixas]

Even if the 380 is a rebranded 290, chances are it will be more efficient, faster, cooler, making it a better GTX 970/80 competitor. Plus it would have all the latest bullet point features.

Edit: GTX 480 to GTX 580, what a transformation, can be done again

 

[Erenhardt]

I would buy maxwell before it is too late. After mad releases their rebrands nv will increase prices of their cards by at least 50%. /FUD

Some people believe Pitcairn will be on the market for more than 5 years? Think about what you guys suggest in your dumb FUD campaign

 

[bystander36]

Even if the 380 is a rebranded 290, chances are it will be more efficient, faster, cooler, making it a better GTX 970/80 competitor. Plus it would have all the latest bullet point features.

Edit: GTX 480 to GTX 580, what a transformation, can be done again

If it was a "rebrand" that means they didn't change anything but clocks, which is more akin to the 680 to the 770. Unless they get a die shrink, it won't be more efficient, and only cooler if it has a better HSF.

The 580 might not have been a lot different in spec's, it still was not the same chip.

 

[DarkKnightDude]

Yeah, GTX480 to 580 not a great comparison, GTX 680 to 770 is more akin to this situation.

 

[gamervivek]

Though 580 did get efficiency improvements despite being on the same process.
If AMD were going to rebrand, they should have done it with the inventory already. Doubt it would have been costlier than the reduced prices in order to get rid of it.

 

[Gikaseixas]

If it was a "rebrand" that means they didn't change anything but clocks, which is more akin to the 680 to the 770. Unless they get a die shrink, it won't be more efficient, and only cooler if it has a better HSF.

The 580 might not have been a lot different in spec's, it still was not the same chip.

I'm aware of the details, just wanted to show what can be done within the same node. I only meant that AMD can come up with a similar improvement.

 

[RussianSensation]

Even if the 380 is a rebranded 290, chances are it will be more efficient, faster, cooler, making it a better GTX 970/80 competitor. Plus it would have all the latest bullet point features[

That by definition is not a re-brand. If AMD retains the same 64 ROPs/2816 SPs on 380/380X cards and they are largely based on Hawaii but it's a whole new respin/more efficient 28nm transistors, it's not the same die. Secondly, what if AMD shifts them to GCN 1.2/1.3 and incorporates some of the benefits of Tonga into 380/380X?

It could easily be the case that the specs align just like they did with HD4870 and HD5770 but they were completely different chips. Reusing Pitcairn doesn't logically align with AMD having TrueAudio/FreeSync capable GPUs. The display portion of the controller responsible for FreeSync is inside the die itself, not on the PCB board/display controllers on the board (per Huddy). Therefore, to have FreeSync capable Pitcairn, it would also need to be redesigned/manufactured as a new chip.

The other problem with the re-branding theory as I mentioned before but it keeps getting ignored is that AMD could have easily re-branded R9 290/290X as 380/380X and priced them at $269/$329 but based on the sales of R9 290/290X, those cards are not selling well at even lower prices. As I mentioned above, R9 380/380X could share identical ROP/TMUs/SP/512-bit bus specs as R9 380/380X but the technical characteristics of the chip itself on a deeper (transistor and GCN revision level) could be different. Thus, even if the specs match perfectly with R9 290/290X on paper, that's not a 100% indication that the new range are re-brands. For example, AMD could have incorporated some of the new voltage saving technologies found in Carrizo. Legit Reviews reports this as a likely possibility:

"Adaptive Voltage-Frequency Scaling (AVFS) sensors that enable smart voltage monitoring for power savings of up to 10% are said to be in the next-gen graphics cards."
http://www.legitreviews.com/amd-radeon-r9-390x-video-card-powers-demos-gdc_159733#DzEluGY0qduUeDYV.99"

Everything I am typing has already been mentioned by various posters prior but it keeps getting ignored/brushed aside. Right now much of this is just speculation but we should expect more AMD doom and gloom threads/posts, and negative AMD spin as we get closer to AMD's next earnings report. This has become like clock work.

 

[StereoPixel]

Is 360/370 for OEM only?
http://www.techpowerup.com/gpudb/2661/radeon-r9-360-oem.html

 

[pepone1234]

Is there any possibility that those rebrands are ported to GF instead of continue fabbing at TSMC?

 

[dacostafilipe]

380/X would be 40% slower then 390/X ... that's a lot. We are not used to such a big performance gap between tiers.

I miss a 52CU/384bit card in the mix.

 

[Dribble]

380/X would be 40% slower then 390/X ... that's a lot. We are not used to such a big performance gap between tiers.

I miss a 52CU/384bit card in the mix.

Yes we are, look at 280 vs 290.

 

[Head1985]

Jumping to conclusions based on beta driver inf with potentially placeholder numbers & decimals...

What makes you think Hawaii, which cannot compete with 970/980, is going to be relaunched to compete against 970/980?

Also who in their right minds would actively prepare more of to sell a 512bit bus 4GB 300W TDP SKU at ~$250? The only reason its been selling that low is because nobody wants to buy it at similar prices to the 970 and they have to clear inventory.

AMD may as well sell nothing because it would a waste of their time and their AIBs time to sell such a low profit part compared to the competition.

Why prepare to launch a dud that's already obsolete and failing to compete? Zero logic. Do you honestly believe AMD is THAT stupid?

They are already that stupid with PR.Why we dont know already something about 300 series?
Maxwell GTX970/980 is already OUT 6-7 Month and we only knows 390x will have HBM.
They didnt say r300 series will be brand new cards with better performance than maxwell to tease pepole.Or something to stop people buying Maxwell.They are just silent and wtching how good maxwell sell.
AMD is just already stupid retarded.

 

[Techhog]

you really should take anything posted by someone who wants amd to die with a truck load of salt. wait for bench marks and reviews. 370 is rumored for mid april release right? just 20 more days.

The person I replied to wasn't the person who desperately wants every CPU company besides Intel to die. On top of that, there's no arguing with the driver.

 

[n0x1ous]

Seems very logical what RS posted that if they are indeed having new codenames (grenada) that theres a good chance it will have the same specs as hawaii but respun to tidy up some transistors and perhaps incorporate the 1.2/1.3 gcn features.

Seems odd to change the die name otherwise unless they are trying to intentionally mislead which i doubt.

480 to 580 i think is a good comparison as one was GF100 then respun to a refined GF110 with slightly better power use and thermals etc etc.

 

[Silverforce11]

They are already that stupid with PR.Why we dont know already something about 300 series?
Maxwell GTX970/980 is already OUT 6-7 Month and we only knows 390x will have HBM.
They didnt say r300 series will be brand new cards with better performance than maxwell to tease pepole.Or something to stop people buying Maxwell.They are just silent and wtching how good maxwell sell.
AMD is just already stupid retarded.

So what exactly do you expect would happen if AMD starts their own hype train?

People would just say they are desperate and its pure BS hype.

All they need to say or leak is the ETA for the launch. That way people on the fence who can wait will wait and see.

 

[Head1985]

So what exactly do you expect would happen if AMD starts their own hype train?

People would just say they are desperate and its pure BS hype.

All they need to say or leak is the ETA for the launch. That way people on the fence who can wait will wait and see.

Everything is better than been silent for 7 monts.
If they have better and faster cards than NV and they are silent they are just stupid.
If they rebrand most of cards then AMD is dead and you have right no PR will HELP there.

 

[TemjinGold]

So a few folks have suggested the driver lines could be placeholders instead of proof of rebranding. Has AMD done that in the past before to anyone's knowledge? Or has history showed that "the driver doesn't lie"?

 

[Silverforce11]

So a few folks have suggested the driver lines could be placeholders instead of proof of rebranding. Has AMD done that in the past before to anyone's knowledge? Or has history showed that "the driver doesn't lie"?

We don't even know if its a 100% rebrand or a re-tweak die to add in more modern features etc. AMD has publicly said its entire 2015 lineup are all FS compatible for instance. That alone would rule out a straight re-brand.

 

[TemjinGold]

We don't even know if its a 100% rebrand or a re-tweak die to add in more modern features etc. AMD has publicly said its entire 2015 lineup are all FS compatible for instance. That alone would rule out a straight re-brand.

Yep, I get that. All I'm asking is, has AMD (or nVidia even) ever in the past did this in drivers only to later change that because they were placeholders. We know the driver happened. All I'm wondering is how conclusive is that based on history.

 

[Kenmitch]

We don't even know if its a 100% rebrand or a re-tweak die to add in more modern features etc. AMD has publicly said its entire 2015 lineup are all FS compatible for instance. That alone would rule out a straight re-brand.

Those in the know aren't talking. Nothing but rumors, fud, and the usual death to AMD all over the web.

 

[Kenmitch]

Yep, I get that. All I'm asking is, has AMD (or nVidia even) ever in the past did this in drivers only to later change that because they were placeholders. We know the driver happened. All I'm wondering is how conclusive is that based on history.

Disclaimer: Not an expert on the subject.

I'd imagine you could find examples that go both ways. Probably even examples of gpus that never even saw daylight.