AMD has revealed their concept of a premium reference GPU for the upcoming Radeon Vega launch, with the "Frontier Edition" of the new graphics cards.
"Today, AMD announced its brand-new Radeon Vega Frontier Edition, the world’s most powerful solution for machine learning and advanced visualization aimed to empower the next generation of data scientists and visualization professionals — the digital pioneers forging new paths in their fields. Designed to handle the most demanding design, rendering, and machine intelligence workloads, this powerful new graphics card excels in:
- Machine learning. Together with AMD’s ROCm open software platform, Radeon Vega Frontier Edition enables developers to tap into the power of Vega for machine learning algorithm development. Frontier Edition delivers more than 50 percent more performance than today’s most powerful machine learning GPUs.
- Advanced visualization. Radon Vega Frontier Edition provides the performance required to drive increasingly large and complex models for real-time visualization, physically-based rendering and virtual reality through the design phase as well as rendering phase of product development.
- VR workloads. Radeon Vega Frontier Edition is ideal for VR content creation supporting AMD’s LiquidVR technology to deliver the gripping content, advanced visual comfort and compatibility needed for next-generation VR experiences.
- Revolutionized game design workflows. Radeon Vega Frontier Edition simplifies and accelerates game creation by providing a single GPU optimized for every stage of a game developer’s workflow, from asset production to playtesting and performance optimization."
From the image provided on the official product page it appears that there will be both liquid-cooled (the gold card in the background) and air-cooled variants of these "Frontier Edition" cards, which AMD states will arrive with 16GB of HBM2 and offer 1.5x the FP32 performance and 3x the FP16 performance of the Fury X.
Radeon Vega Frontier Edition
- Compute units: 64
- Single precision compute performance (FP32): ~13 TFLOPS
- Half precision compute performance (FP16): ~25 TFLOPS
- Pixel Fillrate: ~90 Gpixels/sec
- Memory capacity: 16 GBs of High Bandwidth Cache
- Memory bandwidth: ~480 GBs/sec
The availability of the Radeon Vega Frontier Edition was announced as "late June", so we should not have too long to wait for further details, including pricing.
There is that machine
There is that machine learning again.
Suggest IMHO that the days of consumers playing 3D games on desktop PC pushing graphics cards forward are coming to a close.
The Pozidrive brain of Azimuth’s fiction is in fact the TensorFlow GPU. Instead of explaining the universe or how love works it will stack shelves and drive van loads of cheaply made consumer goods from China.
Will it spare some cycles to examine our face’s micro muscles to guage our current emotional state, offering a robot hug to those it finds sadden by their life?
Or instead spew robotic platitudes “How are you? …Good, glad to hear it. Have a nice DAY”
“The Pozidrive brain of
“The Pozidrive brain of Azimuth’s fiction is in fact the TensorFlow GPU”
I’m assuming you meant “positronic” and “Asimov”, or otherwise were referencing some obscure story about telescope angle adjustments.
Maybe it is a sentient screw
Maybe it is a sentient screw that is current in orbit?
nVidia already cemented their
nVidia already cemented their position in AI, I don’t understand what AMD is even hope to accomplish
not even taking into account that nVidia is getting ready to release the kraken, Volta with Tensor Units
and they haven’t said a peep about consumer Vega – Q3 then?
Considering how much Nvidia
Considering how much Nvidia has invested into professional and HPC, I doubt it is a war they even hope they are winning, but just enough to get their own products out there as an alternative.
they cemented squat.
they cemented squat.
why do ppl forget about Google’s TPU for targeted tensor workload.
They’ve now confirmed
They’ve now confirmed consumer Vega will be announced at Computex on the 31st.
They are trying to accomplish
They are trying to accomplish the same thing that any other company is trying to accomlish, to make that $$$!!!
I know this is petty but in
I know this is petty but in the press image they have drawn in light coming out of the “R cube” on the hybrid. Really? GMC marketing cars with lines drawn trailing off the back will be next? This is something a 3 year old would do in a coloring book.
However big props for not going with the age old dollar store red and black ATI plastic.
AMD is pushing their company
AMD is pushing their company to compete where the money is, and that’s always been enterprise. These days, enterprise is driving GPU progress because of the ubiquitous use of GPGPU acceleration in almost everything: AI, Big Data, and biomedical to name just a few uses. Just because nVidia’s been there already doesn’t mean AMD should just give it up. Just because video games aren’t driving GPU development doesn’t mean video games won’t benefit.
AMD is competing on a lot of different fronts now and I can’t wait to see what the next few years will look like! Could they sh!t the bed like before? Sure, it’s possible, but considering their recent successes and their current leadership I doubt we’ll see that level of ineptitude.
Looking at that water cooled
Looking at that water cooled Vega card, the size doesn’t make any sense if it’s a single GPU card.
*Puts on Tinfoil gat*
If it was purely an HBM2 GPU it would be around the size of the Fury X. Instead it’s a full length PCB card.
The card has 16GB but only 480GB/s bandwidth. That means either 4 stacks of 4GB, which would mean significantly higher bandwidth. Or they’re using two 8GB stacks, which is also unlikely. Neither Samsung nor SK Hynix has shown any working.
Even Ryan Smith at Anandtech doesn’t know how AMD might have done that.
Calling it here folks, Vega 16GB is actually 8GB HBM2, and 8GB GDDR5 using the HBCC to use it all.
They don’t list the card as having 16GB HBM2, they state and list 16GB HBC – High Bandwidth Cache.
That was the first thing
That was the first thing about the specs that jumped out to me as well — seems wayy too low for it to be all HBM2.
They’ve confirmed (on the
They’ve confirmed (on the Reddit AMA) that it’s two 8GB stacks of HBM2.
As the other poster pointed
As the other poster pointed out, it is indeed two 8GB stacks of HBM2.
here is the link to the AMD reddit AMA: