Performance of Floating-point Intensive Kernels on Low-power Processor - A Case Study with Geodesic Distance Kernel.
Zheming JinPaulius VeleskoHal FinkelPublished in: IGSC (2019)
Keyphrases
- low power
- floating point
- geodesic distance
- single chip
- high speed
- instruction set
- kernel function
- riemannian manifolds
- euclidean space
- reproducing kernel hilbert space
- gate array
- low cost
- positive definite
- power consumption
- floating point arithmetic
- geometric structure
- support vector
- shortest path
- euclidean distance
- feature space
- fixed point
- distance transform
- manifold learning
- floating point unit
- distance metric
- shape analysis
- vector space
- gaussian mixture
- high dimensional
- cmos technology
- real time
- kernel matrix
- low dimensional
- data points
- pattern recognition