XFX Teases Custom RX Vega 56 and RX Vega 64 Double Edition Graphics Cards

Subject: Graphics Cards | December 1, 2017 - 02:48 PM |
Tagged: xfx, vega 10, Vega, RX VEGA 64, RX Vega 56, double edition, amd

Not content to let Asus have all the fun with X shaped products, graphics card manufacturer XFX is prepping two new Vega graphics cards that feature a cut-out backplate and cooler shroud that resembles a stretched-out X. XFX has, so far, only released a few pictures of the card but they do show off most of the card including the top edge, cooler, and backplate.

View Full Size

XFX has opted for a short PCB that extends slightly past the first cooling fan. The card is a dual slot design with a large heatsink and two large red fans and a bit less than half of the cooler extends past the PCB as a result. Cooling is not an issue thanks to liberal use of heat pipes (I think there are five main copper heat pipes), but the cooler hanging so far past the PCB has resulted in the two 8-pin PCI-E power connectors ending up in the middle of the cooler (the middle of the X shape) which is not ideal for cable management (still waiting for someone to put the PCI-E power connectors on the back edge closest to the motherboard!) but with a bit of modding maybe it would be possible to hid the wires under the shroud and route them around the card as one of the photos it looks like there is a bit of a gap between the heatsink and the shroud/backplate heh).

The design is sure to be divisive with some people loving it and other hating it, but XFX has put quite a bit of work into it. The red fans are surrounded by a stylized black shroud with a carbon fiber texture while the top edge holds the red XFX logo. The backplate specifically looks great with a black and grey design with red accent that features numerous cutouts for extra ventilation.

Display outputs are standard with three DisplayPort and one HDMI out.

TechPowerUp along with Videocardz are reporting that the card will come in both RX Vega 56 and RX Vega 64 variants. Unfortunately, while XFX has gone all out in the custom cooling and backplate, they are not pushing any of the clockspeeds past factory settings with the RX Vega 56 Double Edition clocking in at 1156 MHz base and 1471 MHz boost on the GPU and 1600 MHz on the 8GB of HBM2 memory. The XFX RX Vega 64 Double Edition is also stock clocked at 1247 MHz base, 1546 MHz boost, and 1890 MHz memory. It is not all bad news though, because with such a beefy cooler, enthusiasts should be able to overclock the chips themselves at least a bit (depending on how lucky they are in the silicon lottery) but it does mean that XFX isn’t guaranteeing anything. Also, overclocking might be more top-end overclock limited on the Vega 64 version versus other custom cards due to it only including two 8-pin power connectors (which does make me wonder what they have done as far as the VRMs versus reference if anything).

View Full Size

XFX has not yet revealed pricing or availability for their custom RX Vega cards.

What are your thoughts on the X design? 

Also read:

Source: TechPowerUp

December 1, 2017 | 03:56 PM - Posted by remc86007

It's a real shame Vega wasn't ready this time last year. Just a lower clocked, GDDR5X "big Polaris" would have been nice to see. Even just releasing the 40cu version that is in the XB1X with a 320bit GDDR5 setup below $300 would have given AMD something to work with. My guess is that AMD is plagued by poor fab output at every turn.

December 1, 2017 | 06:24 PM - Posted by ItsROPsThatGetTheGamersTripping (not verified)

Not poor fab output it's more like AMD only had the funds to design one Vega 10 discrete GPU base die design and that design was really tuned for professional compute/AI workloads.

Nvidia has more base die designs GP100, GP102, GP104, and GP106/107/108 and each of those designs has either more compute for the professional markets GP100, or more compute and ROP resources GP102 for professional graphics and high end professional graphics, or GP104 gaming, and GP106/107/108 lower end gaming. So Nvidia can take one of its base GPU die designs for say professional graphics that has an excess of ROPs and spin out a GTX 1080 TI with 88 ROPs and hit the highest FPS metrics.

So AMD's Vega tops out at 4096:256:64 Shaders:TMUs:ROPs and 16GB of HBM2 for the Radeon Pro WX 9100/Radeon Instinct MI25, and the Radeon Founders Edition SKUs. Also using the same Vega 10 base die with 8GB of HBM2 AMD creates the Vega 64 and its binned variant the Vega 56 with its 8GB of HBM2 and 3584:224:64 Shaders:TMUs:ROPs.

Now Vega 56 and 64 can get close to the GP104 base die based GTX 1080 with its 2560:160:64 Shaders:TMUs:ROPs but not the GP102 base die based GTX 1080Ti with its 3584:224:88 Shaders:TMUs:ROPs. And the GP102 base die design also is used for the Titan X with 3584:224:96 Shaders:TMUs:ROPs and the Titan XP with its 3840:240:96 Shaders:TMUs:ROPs. So you can see that Nvidia has a lot more hold cards than AMD does at this time and look at all the extra ROPs that Nvidia has to play with with its GP102 based die design to get more FPS flung out there compared to the Vega 10 base die design that is mostly AMD's only choice currently for discrete GPUs.

Had AMD increased its Vega 10 base die design's full complement of ROPs to 96 then AMD could have had more options to match Nvidia's ROP counts. And Vega 10 is a very compute/Shader heavy base die design that competes in compute with the GP100 based Nvidia SKUs. Vega 10 is a shader heavy and TMU heavy base die design with even Vega 56 getting almost the full complement of TMUs available from the Vega 10 base die design.

So Nvidia wins the ROP/FPS race and AMD's Vega 10 die base design Vega 64 variant can best a Quadro P6000 in some compute workloads and even the Titan X and Titan XP. If Vega 56 could clock a little higher it could beat the GTX 1080 on more benchmarks besides some DX12 benchmarks as the Vega 56, 64, and the GTX 1080 all have the same number of ROPs it's just that Nvidia get some higher clocks on its less densely packed TSMC 16nm process than AMD can get on that 14nm process node and AMD's more densely packed Vega 10 based dies design(4096 shader cores take up some densely packed die space and those deeper Vega pipelines on those shaders take more transistors to implement to get better performance at higher clocks, deeper shader core pipelines are needed for higher clocks).

And Thus is the answer to why Vega is great for compute but too ROP deficient to compete with Nvidia in frame flinging per second. AMD's TMU/texture fill rates are much higher than Nvidia's TMU/Texture fill rates and AMD's current flagship GPUs are great at compute with their excess of shader cores and that will come in handy as more games port over fully to DX12 and Vulkan. So Vega's longer term improvement potential is better on its new Vega SKUs.

Vega 64/Vega FE is great for non gaming Graphics rendering and Ray Tracing computation acceleration on those Vega FEs with their 16GB of HBM2 to hold large multi-million polygon count meshes/scenes and super high resolution textures for Blender 3D animation rendering where FPS is not the goal. And each frame of animation can take up to 1/2 of an hour or longer to render if all the Ray Tracing settings are turned up but Vega 10 needed more ROPs to compete in gaming workloads with the GTX 1080 and its 88 ROPs and mush higher pixel fill rates that is the very thing that produces the FPS metrics that gamers so swoon over.

The Radeon Pro WX 9100s and Radeon Instinct MI25s are getting most of the best binned Vega 10 dies with the consumres getting to not so good Vega 10 binns and consumers are competing with the coin miners to get the not so great Vega 10 dies. AMD is damn sure going to make sure that the Professional Vega 10 SKUs get priority as those markups are actually very profitable for AMD while the consumer market margins are so low that it's almost a loss leader for AMD in the consumer gaming market what with HBM2 costing a little higher than other options at the moment.

Consumers cant/wont pay for Vega so consumers will have to wait for it, wait for it...Money Talks and AMD needs to get its average corprate margins above that 40% metric!

December 1, 2017 | 06:30 PM - Posted by ItsROPsThatGetTheGamersTripping (not verified)

Edit: GTX 1080 and its 88 ROPs
To: GTX 1080Ti and its 88 ROPs

December 1, 2017 | 06:29 PM - Posted by ItsROPsThatGetTheGamersTripping (not verified)

Edit: GTX 1080 and its 88 ROPs
To: GTX 1080Ti and its 88 ROPs

December 1, 2017 | 06:30 PM - Posted by ItsROPsThatGetTheGamersTripping (not verified)

Remove this post, posted in the worng place!

December 1, 2017 | 06:45 PM - Posted by CNote

I like XFX, my current card is one, but that's the ugliest card I've ever seen. Looks like a cheap Chinese knock off of Zotac.

December 2, 2017 | 10:19 AM - Posted by CrazyRobotDontGiveARatsAzzBoutNoTheLawsOfRobotics (not verified)

It's XFS's Roberto, the Criminally insane, psychotic robot, obsessed with stabbing people in different places with different objects, Edition. With Plenty of stab holes for great GPU ventilation! Look at those crazy Robot's Eyes in that GPU's design! Just wanting to grab and stab with knives, forks, or even shish kabob skewers!

Bender: He'll get all stabby in less than a nanosecond his clock is wound up so tightly!

Roberto: What F-ing Laws of Robotics! I'm getting all 5150 on your Azz! And it's stabby time!

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

CAPTCHA
This question is for testing whether you are a human visitor and to prevent automated spam submissions.