AMD revealed Radeon Vega earlier this year and followed up with news of its Vega Frontier Edition at its 2017 Financial Analyst Day. AMD claims that the Frontier Edition is the "world's first" GPU geared for AI (Artificial Intelligence), creatives, and science pioneers.
The Radeon Vega Frontier edition is set to take the wraps off in June 2017 so more details prior to that at Computex, where AMD is expected to host a keynote will be on the 31st of May. As far as the specs go, the card features 4096 stream processors and clocked beyond 1550 MHz to deliver 13 TFLOPs of FP32 and 25 TFLOPs of FP16 compute performance. There's also 16 GB of HBM2 VRAM which comes in two stacks (8 GB per stack). The graphics card has a total rated bandwidth of 480 GB/s which is lower than 512 GB/s on Fiji. It also comes pixel fill rate of 90 GPixels/s.
More so, AMD's Senior Vice President and Chief Architect at RTG, Raja Koduri, said in a slide that the Radeon Vega Frontier edition comes with four times more memory than their Radeon R9 Fury X graphics card that was rolled back in 2015. The graphics card featured 4 GB of HBM1 VRAM compared to 16 GB HBM2 on the upcoming Vega card. The new card also equips with 3 times more half-precision and 1.5 times more single-precision performance.
Furthermore, AMD Radeon Vega Frontier edition will be hitting the market in June in two variants. One would be an air cooled model which comes in a nice blue industrial texture with a glowing yellow "R" emblem placed on the corner side. Moreover, the graphics card comes with a "Radeon" label and logo on the sides and front while the back is wrapped by a nice blue textured backplate. The whole card is powered by a dual 8-Pin connector configuration that confirms a TDP around 300W. The card is cooled off by a single blower style fan that throws air out of the I/O side exhaust vents.
