Evolving a CUDA kernel from an nVidia template.
William B. LangdonMark HarmanPublished in: IEEE Congress on Evolutionary Computation (2010)
Keyphrases
- parallel implementation
- graphics processors
- compute unified device architecture
- graphics hardware
- gpu implementation
- parallel algorithm
- kernel function
- template matching
- matching algorithm
- kernel methods
- general purpose
- cpu implementation
- shared memory
- sparse kernel
- parallel computing
- neural network
- kernel space
- multiple kernel learning
- times faster
- feature space
- support vector
- parallel computation
- convolution kernel
- mutual subspace method
- gpu accelerated
- component analysis
- efficient implementation
- feature vectors
- feature selection