A 1.45GHz 52-to-162GFLOPS/W variable-precision floating-point fused multiply-add unit with certainty tracking in 32nm CMOS.
Himanshu KaulMark A. AndersSanu MathewSteven HsuAmit AgarwalFarhana SheikhRam KrishnamurthyShekhar BorkarPublished in: ISSCC (2012)
Keyphrases
- floating point
- variable precision
- high speed
- decision table
- fixed point
- rough sets
- reduction method
- knowledge structures
- power consumption
- low cost
- instruction set
- floating point arithmetic
- cmos technology
- decision trees
- dominance relation
- low power
- sensitivity analysis
- silicon on insulator
- set valued
- expert systems
- computer vision