OK, so there's a bit of a debate about it recently and a few tech YouTubers have also given their opinion, so it's only relevant that some random nobody with too many posts on a tech-forum also gives theirs.

I'll start off with what GCN cards AMD has released since 2012.

300 Series:

R9 390X - Hawaii XT 2816 Stream processors based on GCN 1.1
R9 390 - Hawaii PRO 2536 Stream processors based on GCN 1.1
R9 380 - Tonga PRO 1792 Stream processors based on GCN 1.2
R7 370 - Pitcairn PRO 1024 Stream processors based on GCN 1.0
R7 360 - Bonaire PRO 768 Stream processors based on GCN 1.1

200 Series:

R9 290X - Hawaii XT 2816 Stream processors based on GCN 1.1
R9 290 - Hawaii PRO 2536 Stream processors based on GCN 1.1
R9 285 - Tonga PRO 1792 Stream processors based on GCN 1.2
R9 280X - Tahiti XT 2048 Stream processors based on GCN 1.0
R9 280 - Tahiti PRO 1792 Stream processors based on GCN 1.0
R9 270X - Pitcairn XT 1280 Stream processors based on GCN 1.0
R9 270 - Pitcairn XT 1280 Stream processors based on GCN 1.0
R7 265 - Pitcairn PRO 1024 Stream processors based on GCN 1.0
R7 260X - Bonaire XT 896 Stream processors based on GCN 1.1
R7 260 - Bonaire PRO 768 Stream processors based on GCN 1.1

7000 Series:

7970 - Tahiti XT 2048 Stream processors based on GCN 1.0
7950 - Tahiti PRO 1792 Stream processors based on GCN 1.0
7870 - Pitcairn XT 1280 Stream processors based on GCN 1.0
7850 - Pitcairn PRO 1024 Stream processors based on GCN 1.0
7790 - Bonaire XT 896 Stream processors based on GCN 1.1
7770 - Cape Verde XT 640 Stream processors based on GCN 1.0
7750 - Cape Verde PRO 512 Stream processors based on GCN 1.0

Versions of GCN and improvements

GCN 1.0 - Initial release with the 7000 series in 2012, uses a SIMD (Single Instruction Multiple Data) arcitecture instead of the old VLIW design used in previous arcitectures like the 5000 and 6000 series. Each GCN compute unit totals 64 stream processors. AMD launched it as a long term building block for future GPUs, which is why they still use it today.

GCN 1.1 - GCN 1.1 was a minor iteration that added feature support for TressFX and TrueAudio, and improved PowerTune support. Most of the improvements were compute and HSA focused for the APU side of things.

GCN 1.2 - GCN 1.2 is a release we have only seen in Tonga and Fiji so far, the big focus with the release of the R9 285 was power efficiency improvements, AMD improved their instruction scheduling and added newer instruction support, the additional 16 bit Floating point & Integer instructions means that less power can be used on applications that do not need the mathematical accuracy of 32 bit instructions. The new Fiji GPU has a reworked memory architecture so it works with the HBM memory.

Improvements in manufacturing

So a term thrown around a lot is "refinements", unfortunately no-one has been too specific about the refinements but there are board level improvements and the process used has seen improvements.

Both AMD and Nvidia have used the same TSMC 28nm manufacturing process since 2012, over time as the process matures the ability to produce big chips consistently improves, and performance consistency also improves. We get less leaky chips. Which means we can have higher stock clocks, better overclocking and the better chips achieve this with the same or less voltage than before, which leads to a lower average power draw. This is very similar to Intel's Haswell Refresh/Devil's Canyons (note how people are fine with this being called refresh and not rebrand?).

The new cards have also received higher clocked VRAM, from 1250 to 1500MHz.

VRAM

AMD have generally been more generous than nvidia when it comes to VRAM in the past. This is still true for the most part, the Hawaii based products now offer 8GB of VRAM, AMD were saying a lot about Crossfire technology and have improved crossfire support in their new drivers as well, AMD would like people to pair multiple R9 300 series cards to play games at 4K and above resolutions and they've given us the framebuffer to achieve this, reviews have demonstrated that above 1080/1440p the R9 390 and the 970 begin to make more of a gap in the 390s favour and some games will cap out the 970s effective 4GB VRAM. They also offer a 4GB version of the R9 380 at a similar price to nvidias 960 4GB.

Pricing

Here I'll list the MSRP launch price of the products from the 200 and 300 series and the 900 series nvidia cards, a lot of people are saying you might as well pick up the 290X as it's very close to the 390X and is cheaper, which is true. You also need to consider that the 7970 was cheaper than the R9 280X until it went out of stock... very quickly. So if you want to buy you may want to buy now, as this will not hold true for long. Note I am only listing the 200 series cards that have a 300 series equivalent.

300 Series:

390X 8GB: $429 USD
390 8GB: $329 USD
380 2GB: $199 USD
370 2GB: $149 USD
360 2GB: $109 USD

200 Series:

R9 290X 8GB: $459 (note: the 4gb price had been cut to around 399 by the time this card launched)
R9 290X 4GB: $549
R9 290 4GB: $399
R9 285 2GB: $249
R9 265 2GB: $149
R9 260 2GB: $109

900 Series (including 700 series maxwell):

GTX980: $549
GTX970: $329
GTX960: $199
GTX750ti: $149
GTX750: $119

Performance:

R9 390X:
This card is priced a bit lower and performs a bit worse than the 980, but will beat out the 970. I'll use GTA5 as an example here due to the cards high end nature.

 

GTA V is tested with 'Very High' settings. This means putting all the default settings to their highest values, except for Grass Quality, Post FX and Reflection Quality, which have 'Ultra' options that we don't use, and Motion Blur which is disabled. We use FXAA, leaving the performance killing MSAA disabled, with 16x Ansiotropic Filtering applied as well. The settings in the Advanced Graphics menu are left off. We use the game's built-in benchmark, recording performance during the last 60 seconds of the longest scene, which is the most demanding part. It involves flying a jet over a city at high speed followed by a car chase through busy streets, complete with gunfire, crashes and explosions - everything you'd expect from a high action open world game.

 

 

GTA V is tested with 'Very High' settings. This means putting all the default settings to their highest values, except for Grass Quality, Post FX and Reflection Quality, which have 'Ultra' options that we don't use, and Motion Blur which is disabled. We use FXAA, leaving the performance killing MSAA disabled, with 16x Ansiotropic Filtering applied as well. The settings in the Advanced Graphics menu are left off. We use the game's built-in benchmark, recording performance during the last 60 seconds of the longest scene, which is the most demanding part. It involves flying a jet over a city at high speed followed by a car chase through busy streets, complete with gunfire, crashes and explosions - everything you'd expect from a high action open world game.

 

 

R9 390:
This card undercuts the 970 and matches/beats the performance, it also offers double the VRAM so it is better at higher resolutions particularly when paired with another card. I'll use GTA5 again as a demanding title suited to the hardware.

 

We tested the game with everything set to very high and MSAA off. Note the blue results are the reference design card tested at stock clock speeds.

 

 

We tested the game with everything set to very high and MSAA off. Note the blue results are the reference design card tested at stock clock speeds.

 

 

R9 380:
So a mid tier GPU going head to head against GTX960. There's a 2GB and 4GB version of each card priced similarly, so this is mostly a game of preference, although it is demonstrated that the 380 is usually the faster card. I'll only include a 1080p benchmark here, as that's the res these cards are aimed for. And the game I chose was bioshock infinite, a less demanding title than GTA5 used on the enthusiast cards. The settings are all maxed out, but MSAA is disabled.

 

 

R7 370:
So this is the oldest card of the lot and comes in at $149, directly challenging nvidias 750ti. AMD also released this card in the R7 series last generation as the 265 to answer nvidias 750ti price point. There is a 4GB version but I'll stick to the 2GB version, as if you can afford a 4GB 370 the 380 makes more sense. For this test I struggled to find one containing both the 370 and the 750ti, I found one on a polish (I believe) site and they are testing tomb raider on ultra with 2x SSAA and TressFX enabled.

 

 

So; refresh- or rebrand?

 

Well, a rebrand is when you take exactly the same thing and give it a new name. Logitech recently rebranded their entire company as Logi. Nothing changed but the name. On AMDs 300 series there haven't been any big changes however there have been a few improvements allowing them to perform better than previous generation cards. You can consider it a refresh of the same product line, if we look at Intel's Haswell refresh, that launched as a refresh and no-one challenged it being a refresh instead of a rebrand, it was the same chip with some additional capacitors, the rest was achieved due to a better manufacturing process. Much like AMDs revised cards,they have better chips, they gave us more VRAM for the price, cut the MSRP and GPU sellers have given their coolers improvements and facelift, and a lot of them improved their custom PCBs with improved power delivery etc.

 

So why didn't AMD move the product down a tier?

 

I've seen a lot of people slate AMD saying that when nvidia launch the same card/chip twice they move it down. Well strictly speaking there's a reason AMD didn't.

 

They didn't have to.

 

That's right, the cards still perform just fine at their price segment as we've seen so there was no point in a money strapped company like AMD producing an entirely new product to achieve the same performance level as they already had for the sake of it being a new thing. The 750 and 750ti haven't received an update so AMD kept their pricing the same on the refreshed cards, the 380 saw a price cut to make it competitive to the 960. And the 390 and 390X offer good performance in the price points between nvidias 970 and 980, they also make good sense as a pair for 4k crossfire gaming due to the generous framebuffer.

 

The example I see thrown about a lot is the 680 becoming the 770 when it was re-launched. They moved it down a tier because they had to, AMD were releasing a product (Hawaii) that would beat it easily so it would have been pure stupidity for nvidia to keep it as a x80 class product when both companies could beat it by a significant margin, but this time 390/390X still make sense at their performance segment and price because it's still very close.

 

Nvidia have re-launched the same product at the same performance segment/price as well- it was called the 9800GT. A re-released 8800GT with higher memory clock. AMD didn't have something that beat it massively so there was no point in nvidia producing a new product.

 

Exactly the same situation. Both companies have done it in the past and I imagine both companies will do it again.