1. Quantization means using low resolution formats for some variables some of the time: binary (e.g. BinaryConnect), ternary, etc.
  2. The Turing architecture recently introduced by Nvidia is much faster with half-precision calculations than the previous one.

Intuitively, there is a symbiosis between the two, but how exactly and to what extent can the GTX 16xx and RTX 20xx GPUs speed up these calculations?


Browse other questions tagged or ask your own question.