AMD launches Radeon 300 series rebrands for OEMs

Scorpus

Posts: 2,162   +239
Staff member

AMD has quietly launched a new range of graphics cards designed for OEMs and system builders that fall under Radeon 300 series branding. Like with the HD 8000 series a couple of years ago, these cards are essentially all rebrands of existing Radeon 200 series products, and won't be available through retail channels.

The highest end OEM product is the Radeon R9 380, the very same graphics card seen in HP's latest Envy desktops that we speculated could be a next-generation GPU. Unfortunately this isn't the case, as the R9 380 is a basic rebrand of the R9 285 from AMD's current range of graphics cards.

The rest of the OEM range are rebrands of either R7 or R5 series cards, the oldest of which is the R9 370 that uses a Pitcairn GPU first introduced in the Radeon HD 7850 released in early 2012.

Radeon 300 Series OEM Card Previously Known As Changes
Radeon R9 380 Radeon R9 285 VRAM increase from 2 GB to 4 GB
Radeon R9 370 Radeon R7 265 Clock speed increase from 925 MHz to 975 MHz
New 4 GB variant
Radeon R9 360 Radeon R7 260 Clock speed increase from 1000 MHz to 1050 MHz
VRAM increase from 1 GB to 2 GB
VRAM bandwidth increase from 96 GB/s to 104 GB/s
Radeon R7 350 Radeon R7 250 No change
Radeon R7 340 Radeon R7 240 New 1 GB GDDR5 and 4 GB DDR3 variants
Radeon R5 340 Radeon R5 240 Clock speed increase from 780 MHz to 825 MHz
Radeon R5 330 Radeon R5 240 Clock speed increase from 780 MHz to 855 MHz
No GDDR5 variant

As you can see from the table above, many of the rebranded cards have received minor spec upgrades, though performance shouldn't differ hugely from their predecessors. However not all of the cards support the same features: the R9 380 and R9 360 are the only cards not based on AMD's older GCN 1.0 architecture, meaning they're the only cards to support FreeSync, TrueAudio, and CrossFire without a bridge.

There's the possibility that AMD will release these GPUs to the retail market virtually unchanged from the OEM versions. It's unlikely that the company would release a retail R9 380, for example, with a different set of specifications, so any new GPUs from AMD will almost certainly be released under different names.

And as for those new graphics cards, AMD is gearing up to launch a selection of flagship GPUs this quarter, complete with High Bandwidth Memory. Computex is just around the corner, so expect to see these new GPUs in early June.

Also read: AMD Radeon R9 390X variant will come water-cooled with High Bandwidth Memory

Permalink to story.

 
So much for all the speculation on the next gen from AMD, looks like they're stalling and will now go straight to the 400 series for the retail segment when they finalize their silicon. This is just a game of bait and switch for the OEM's, fool the people who buy them in to thinking they're getting new hardware but in fact it's quite the opposite. Glad to not be one of those sucklers. I'll keep on enjoying my near silent GTX970 Strix from Asus, with just 3.5GB of usable VRAM...
 
All-Mighty Devices, we will remember you for the products like this one....and pray there won't be another. Amen!

 
This doesn't bode well for the top tier AMD cards IMO. They're running out of numbers in the 300 range and they haven't even finished reusing all their old parts yet. How fast can their new top dog really be if the 380 is really just a 285? Do they expect me to believe that they'll reuse all the parts from 240 up and not reuse 290 and/or 290X??? So I'm guessing 290X will be 385 or 390 and a single new board will be faster than that as 390 or 395 (single GPU?). Is the new design not fast enough that there is room to release a stepped down version that is still faster than the 290X?
 
These are OEMs, not official AMD retail cards. Remember HD 8000 series, 8xxm series? AMD hasn't released anything yet, OEMs did.
 
They are blowing a possible advantage.

Nvidia dropped the ball with their memory mistake/poor planning issue with the gtx 970 series. As well as the, imo, permanently annoying throttling in the 960,970,980 series, forcing users to have to use overclock utilities to keep their clocks up.

AMD had a position that if they released something that was 100% equal to the NVIDIA power, without the mistakes and with an option to turn off throttling (setting to max power doesn't do it, it's a bios thing), there would have been a mass exodus from nvidia.. from those who weren't fanboys of either.

But no, they keep dragging their feet. The re-branding of this series was realized in their driver from last month. There has been rumors and hints on testing boards, of the 390x being on the titan level of performance. But it's not out yet.
 
So there is r9 380 (partially disabled chip), 370 (partially disabled) and 360 (partially disabled). Where the hell do they pile all those completely working chips? And Tonga (r9 380, r9 285) was not yet sold as completely enabled... are they putting it aside for 4xx series or what?!
 
So much for all the speculation on the next gen from AMD, looks like they're stalling and will now go straight to the 400 series for the retail segment when they finalize their silicon. This is just a game of bait and switch for the OEM's, fool the people who buy them in to thinking they're getting new hardware but in fact it's quite the opposite. Glad to not be one of those sucklers. I'll keep on enjoying my near silent GTX970 Strix from Asus, with just 3.5GB of usable VRAM...

Yes, here we have a guy criticizing AMD for re-branding when he praises Nvidia for their own.
 
Yay, typical AMD these days. They just tack a new number on the same old products, and expect people to line up and buy them. What the hell is wrong with you AMD???? Where is the incentive to buy their products when they are nothing but re-badges. Since they do this, where is the incentive for Nvidia or Intel? They have no reason to spur advancements when AMD doesn't provide competition.
 
Yes, 9xx series is not a rebrand. But before that, there are generations of rebrands and "tweaked" arhitectures. And even now this new arhitecture is great for gaming but nvidia is behing in compute workloads as they reduced dedicated die area percentage for those workloads (and this is happening every generation after 5xx series)
 
I used to be a huge AMD (and ATI )fan. I just loved their value offers in the days of Phenom 2 X4 980 black edition and ATI radeon 9000 series. But they went down a wrong path with their APUs and GPUs apparently. And while the GCN architecture is delivering decent performance it still consumes way too much power and generating more than acceptable heat.
For long years AMD has only been trying to give answers to Intel's and Nvidia's innovation but has not given any big innovation of it's own.
I feel really sorry for this company and it would be a dark day for all of us to see them close the doors. In that case both the CPU and GPU market would become monopolistic and the already pricy Intel and Nvidia products would just become even more expensive without motivation for innovations...
 
Last edited:
AMD has a hard time competing with NVidia and Intel, well because of a lot of reasons!!!

1) Intel's SMT Hyper-Threading was, (SO I HAD THOUGHT) was a proprietary Innovation that Intel them selves had patented, so in turn AMD, Nor any other processor designer/developer could use that specific cpu feature, so AMD designed the CMT ( or the so called Bulldozer Clustered-Multi threading Modules, which in turn became the AMD FX Bulldozer Core pairs, aka, .... for example....
*an,..

2) with amd VS NVidia, NVidia . for one had to buy out the PhysX, GPU hardware accelerated Physics Processing, which when Ageia Technologies first designed the GPU Hardware add-on Graphics Physics processing unit, at which could be used with either AMD/ATI OR NVidia Graphics Cards, and once NVidia had purchased/bought out ageia technologies,
NVidia had released a driver update, not only for NVidia's gpus with PhysX, but even the older ageia Hardware Accelerated PhysX Processor cards wouldn't work. :( .. and if an AMD/ATI graphics card was detected/installed, even the AGEIA PhysX Accelerators wouldn't function, if an AMD/ATI GPU was detected, and I have checked this,. my ageia PhysX card worked until NVidia's new Driver release, then bam, got very slow/poor PhysX Acceleration, until I reverted back to Ageia's older driver.. and also,

Ageia's PhysX had to be not only bought out by NVidia, NVidia also re patented the Hardware-PhysX
 
Back