With the last version 0.33b5 i'm not that sure the best multi_hash is the max any longer. To be tested, but on my two RX560 one 2G one 4G, the 4G is no longer the fastest. And high values tend to make the hashrate less stable on RX.

Linux version: unlikely, by lack of time. I'm struggling at keeping the dev up and i still have to do support on the CPU version.
Again, little support today sorry

nVidia version (question i skipped before): very very unlikely, i dev in OpenCL, not Cuda, and nVidia's OpenCL somehow sucks.

I'm experimenting a very different implementation, JCE Hybrid Miner, with the AES job done by the CPU in a split second (because an AES CPU is basically an AES one-cycle asic) and only the scratchpad done by the GPU, with the requirement the CPU is aes-compatible, and the GPU memory relocated is above 4G (option in most BIOSes).
If it works, performance should skyrockets The bottleneck is the synchronization, which is fine on OpenCL for FPGA (like Intel or Xilix) but not on AMD GPUs.