Blog

Measurements and how-tos on GPU energy: what jobs really use, read from the hardware's own meters.

fp16 vs fp32 on a Tesla T4: 81 to 84% less GPU energy

17 October 2026 · The same matrix multiply job in fp16 and fp32, five alternating runs each, under three setups. fp16 used 81 to 84% less GPU energy and finished 4.3 to 4.8× faster.