Alright folks, this one is hot off the mill. Whispers have reached us of AMD planning to announce its fastest Radeon graphics card ever. The fabled dual Fiji "Gemini" board, otherwise known as the R9 Fury X2 in press circles. This VR and 4K gaming behemoth featuring two Fiji GPUs with nearly 18 billion transistors is the company's latest creation and the last ever high-end 28nm graphics board to launch on the desktop before Polaris. AMD's next generation 14nm FinFET graphics architecture powering the company's next generation of Radeon graphics cards coming this summer.
Console-sized Virtual Reality and DirectX12 Radeon powered mini-supercomputers from MAINGEAR
We've already discussed AMD's "Gemini" board at length many times in the past. From its original announcement at E3 last year, to its latest appearance at VRLA inside the console-sized Falcon Northwest Tiki box and the many teases of said box on twitter by Roy Taylor. The Tiki is AMD's idea of the perfect virtual reality and DX12 developer box. Packing the highest performance graphics board in existence with 9 times the compute power of the XBOX ONE inside a form factor that's just as compact.
AMD's Fastest Graphics Card Yet To Debut At GDC - Radeon Gemini "Fury X2"
Burn in success 100%, shipping soon! @FalconNW @AMDRadeon #DX12 #VR @AMDDevCentral pic.twitter.com/qSHcgMQi6T
— Roy Taylor (@Roy_techhwood) February 28, 2016
AMD will be shipping these systems out to developers very soon and they will also be showcased at the company's upcoming red hot "Capsaicin" press event at GDC. An event that sources close to AMD have indicated will include a "very special" product launch. One that will culminate the company's graphics efforts into the last and most powerful 28nm Radeon product to date. In a sense it will be a celebration and a finale paving the way to Polaris.
Prepare to have your technology senses tantalized with #AMDCapsaicin at #GDC16: https://t.co/6YmMQc2jxC pic.twitter.com/gfcI4RLikc
— AMD (@AMD) March 7, 2016
Capsaicin Press Release @AMD
AMD (NASDAQ:AMD) today announced that it will webcast live from the historical landmark, Ruby Skye, during the Game Developers Conference in San Francisco. This marquee developer and press event, entitled "Capsaicin," will be hosted by Radeon Technologies Group's Senior Vice President and Chief Architect Raja Koduri on Monday, March 14 at 4:00 PM PT (6:00 PM CT/7:00 PM ET), and showcase AMD's world-class hardware, software and gaming partners.Named after the compound that delivers a pepper's spicy kick, "Capsaicin" will explore the inner essence of the graphics processing unit (GPU) and how it powers innovations in gaming and virtual reality. The webcast will feature an inside look at the latest technological advancements affecting the enthusiast and developer communities as they approach the burgeoning virtual reality market.
A real-time video webcast of the event will be accessible on AMD's Investor Relations home page: ir.amd.com. A replay of the webcast can be accessed a few hours after the conclusion of the live event and will be available for one year after the event. A replay of the webcast will be accessible on the AMD YouTube channel.
The launch will include the announcement of the fabled dual Fiji "Gemini" graphics card as well as the launch of several Gemini powered, compact, extreme high-performance systems. Of which the Tiki from Falcon Northwest will only be one. Boutique system builder MAINGEAR is also said to announce similar Gemini powered systems which are even smaller than the XBOX ONE. AMD's "Capsaicin" event will be live-streamed. So if you haven't found reason to tune in before, well... you just have.
| WCCFTech | Radeon Pro Duo | AMD Radeon R9 Fury X | AMD Radeon R9 Nano | AMD Radeon R9 Fury | AMD Radeon R9 290X |
|---|---|---|---|---|---|
| GPU | Fiji XT x 2 | Fiji XT | Fiji XT | Fiji Pro | Hawaii XT |
| Stream Processors | 8192 | 4096 | 4096 | 3584 | 2816 |
| GCN Compute Units | 128 | 64 | 64 | 56 | 44 |
| Render Output Units | 128 | 64 | 64 | 64 | 64 |
| Texture Mapping Units | 512 | 256 | 256 | 224 | 176 |
| Memory | 8GB HBM (4 GB Per GPU) | 4GB HBM | 4GB HBM | 4GB HBM | 4GB GDDR5 |
| Memory Interface | 8192bit (4096 Per GPU) | 4096bit | 4096bit | 4096bit | 512bit |
| Memory Frequency | 500Mhz | 500Mhz | 500 MHz | 500Mhz | 1250Mhz |
| Effective Memory Speed | 1Gbps | 1Gbps | 1Gbps | 1Gbps | 5Gbps |
| Memory Bandwidth | 1 Terabyte/s | 512GB/s | 512GB/s | 512GB/s | 320GB/s |
| TDP | TBA | 275W | 175W | 275W | 250W |
| GFLOPS/Watt | - | 31.3 | 47.1 | 26.2 | 19.3 |
| Launch Price | TBA | $649 | $649 | $549 | $549 |
| Launch Date | 2016 | 24th June 2015 | 7th September 2015 | 10th July 2015 | 24th October 2013 |
The Secret Behind AMD's Dual GPU Contraptions
In 2013 AMD introduced a brand new Crossfire hardware implementation by adding a DMA engine to the Crossfire composting block on the GPU die. This was first introduced with the Hawaii GPUs in September of 2013, and each subsequent GPU that AMD introduced also had this technology which AMD dubbed XDMA, short for Crossfire DMA.
AMD RADEON™ R9-‐290X, ONE BIG DGPU MIKE MANTOR
NEW AMD CROSSFIRE™ TECHNOLOGY -‐ XDMA
- XDMA is a modern approach to negoIaIng the communicaIon of mulIple GPUs for peak gaming performance with the ultra-‐high-‐resoluIon displays of tomorrow.
- Radeon™R9-‐290X includes a Hardware DMA engine in the AMD CrossFire™ composiIng block
- Designed for AMD Eyefinity and UltraHD resoluIons via DisplayPort™ ‒
- Allows for direct access between GPU display pipelines over PCI Express®
- No external connector required
- CompaIble on AMD Catalyst™ frame pacing technologies
- No performance penalty versus external bridge
XDMA, Crossfire DMA Engines, Enable Exceptional Multi-GPU Performance Scaling
The technology enables faster, lower overhead communication between multiple GPUs, this in turn translates to better Crossfire scaling. And that's what we're seeing take effect here. It also enables graphics cards in Crossfire setups to inter-communicate through the PCI Express interface which eliminates the need for a Crossfire bridge. But most importantly is that it enables more effective and efficient multi-GPU scaling. Allowing the performance of Crossfire, TriFire and QuadFire GPU setups to scale significantly better.
We've taken an in-depth look at Crossfire and SLI scaling in a previous article which you should check out if you want to learn more about how this technology can affect your gaming experience as well as reduce frametime variance, which is what causes the perceived micro suttering that gamers sometimes notice from multi-GPU setups.
AMD Radeon Graphics Architectures
| WCCFTech | Year | Consumer Product | GPU | Process | Transistors In Billions | Performance | Memory | Bandwidth |
|---|---|---|---|---|---|---|---|---|
| Southern Islands | 2012 | HD 7970 Ghz | Tahiti | 28nm | 4.3 | 4.1 TFLOPS | 3GB GDDR5 | 264GB/s |
| Volcanic Islands | 2013 | R9 290X | Hawaii | 28nm | 6.2 | 5.6 TFLOPS | 4GB GDDR5 | 320GB/s |
| Caribbean Islands | 2015 | R9 Fury X | Fiji | 28nm | 8.9 | 8.6 TFLOPS | 4GB HBM1 | 512GB/s |
| POLARIS | 2016 | RX 480 | Polaris 10 | 14nm | 5.7 | 5.2 TFLOPS | 8GB GDDR5 | 256 GB/s |
| VEGA | 2017 | RX Vega 64 | Vega 10 | 14nm | 12.5 | 13 TFLOPS | 8GB HBM2 | 480GB/s |
| VEGA Refresh VEGA 2 | 2018 2018 | TBA TBA | TBA Vega 20 | 12nm 7nm | 12.5 TBA | TBA TBA | TBA 32GB HBM2 | TBA 1 TB/s |
| NAVI | 2019 | TBA | Navi 10 | 7nm | TBA | TBA | Nextgen Memory | TBA |
How AMD Managed To Cram Unprecedented Levels Of Graphics Horsepower Inside Console-Sized Mini-Supercomputers
Fiji is AMD's largest ever graphics processing unit and the very first in the world to feature 3D structured, 2.5D stacked High Bandwidth Memory, or HBM for short. A standard that AMD and SK Hynix, one of the world's largest memory makers, co-invented. Because vertically stacking dies enables much greater densities and because HBM chip are smaller than GDDR5 chips to begin with there are immense area savings on the printed circuit board of the graphics card as a result. Allowing for the creation of far more compact form factors.
Also unlike GDDR5, HBM is packaged alongside the host processor, in this case the GPU, on a single interposer. The closer proximity to the GPU enables significantly wider memory interfaces and reduces latency. The smaller, shorter connections also enable greater power efficiency. This means that HBM will also require less voltage to operate allowing for even more power savings.
HBM allowed AMD to create significantly smaller and more power efficient graphics cards. Starting with the popular R9 Nano and all the way up to the dual Fiji board inside the Tiki. A consumer version of the dual GPU board was supposed to land sometime late last year, but it was delayed to Late Q1, early Q2 of this year to align it with major VR headset launches according to the company.
Next Chapter Of Radeon - Polaris & The 14nm Revolution
As impressive as these systems are, more exciting perhaps for ordinary gamers is the advent of AMD's Polaris powered Radeon 400 series graphics cards launching this summer. Featuring cutting edge 14nm technology, 4th generation GCN graphics architecture and a whole new set of user and developer features.
AMD has already demoed the smallest member of the Polaris family. The GPU demonstrated performance and power efficiency results that were by far the most impressive we had seen from one generation of GPUs to the next to date. The company revealed that there will be multiple GPUs based on the Polaris architecture to address the entry level, mid-range and high-end segments of the discrete graphics market.
So far it has unveiled two 14nm FinFET Polaris GPUs, given the names Polaris 11 and Polaris 10. One of which is a very small GPU, estimated to be around the same size as AMD’s Cape Verde which is 123mm² large. The goal for this chip is to deliver console-class gaming performance in thin and light notebooks. This is the GPU that AMD chose to demo at CES. The other member of the Polaris family is a large GPU, described as a successor to the Radeon R9 Fury X.
Polaris 11 And Polaris 10 Are AMD’s Fastest And Most Power Efficient GPUs To Date
Raja Koduri, Chief Architect and head of AMD's Radeon Technologies Group, asserted that both Polaris 11 and Polaris 10 are “extremely power efficient”. He went on to state that these GPUs and the Polaris architecture in general will deliver “our most revolutionary jump in performance so far”. A historical leap that's achieved through the combination of cutting edge 14nm FinFET manufacturing technology and the company's 4th generation Graphics Core Next architecture.
Raja Koduri, head of Radeon Technologies Group at AMD
Image Credit: Dean Takahashi via venturebeat.com
Raja also spoke briefly about the makeup of the new graphics architecture. Featuring a new geometry processor, a redesigned command processor and completely new 4th generation Graphics Core Next compute units – the main component of all modern AMD GPUs – with “a very high increase in performance”. He also spoke about the new multimedia cores and the new display engine in Polaris. Which will enable gamers to record and stream 4K gameplay at 60 FPS with no performance cost.
Suffice to say the next generation of Radeon graphics cards are shaping up to be everything that a gamer would ever want. Significantly better performance , a plethora of new features and power efficiency to boot. AMD Polaris graphics cards are set to launch this summer, before the back to school season, for both desktops and gaming notebooks
