The Current Radeon RX Vega 64 / Radeon VII Linux OpenCL Performance Against NVIDIA
The texture read bandwidth test of SOC was substantially faster with the NVIDIA Turing GPUs on GDDR6 video memory compared to the current Vega cards. It will be interesting to see where AMD's Navi ends up fitting into this equation.
In the single-precision FFT test of SHOC, the Radeon VII performance absolutely dominates thanks to the single-precision capabilities not being crippled.
But in a more realistic workload like MD5 hashing, the Radeon VII performance was short of the RTX 2080.
The HBM2 memory of VII did pay off in some of the OpenCL memory tests. The Radeon VII performance was around that of the GeForce RTX 2080 in an OpenCL memory copy benchmark.