Jul 2026
CONQuER: Hardware-Aware Mixed-Precision Quantisation with Online-Calibrated Surrogates
Evaluation across mobile and laptop CPUs, and server GPUs demonstrates that optimal quantisation policies are hardware-dependent, so CONQuER shifts quantisation into the compiler pipeline at the TOSA level, enabling intelligent configuration handling based on compiler support.
Aidan Dakhama, Ajitha Rajan
· arXiv.org · 0 citations