What is Rapid Packed Math?
One of the Exclusive feature of RX Vega is called Rapid Packed Math technology, Which allows 16-Bit operation to be used between 32-bit operating effectively enabling mixed-compute workloads. This basically means both of these games should run faster with F16 compute as compared to standard F32 ( All games currently use) .
This will alow frames to be completed at a faster rate and increase the framerate and responsiveness of future gaming titles, Keep in mind not all in-game calculations can be used with Rapid Packed Math. The performance impact of this technology will depend on how much game code can be simplified with rapid packed math now it depends on developers how much time that they’re going to invest on a feature. This feature is being supported exclusively by the AMD RX Vega graphics cards that will be released later this month, Polaris, Fiji and previous GCN owners wont be able to benefit from it.
AMD’s RX Vega GPUs will be able to close the gap between the GTX1080 and the GTX1080Ti via Rapid Packed Math and Futuremark tool will support this feature as well.
Not only this, AMD’s RX Vega GPUs offer the most complete support for Microsoft’s DX12 API, These features will be able to boost AMD cards in order to compete with Nvidia. Again we dont know if developers will take advantage of all these DX12 features.
Also Read: AMD Giving Up on CrossFire with RX Vega