DeepUM: Tensor Migration and Prefetching in Unified Memory.
Jaehoon JungJinpyo KimJaejin LeePublished in: ASPLOS (2) (2023)
Keyphrases
- prefetching
- response time
- access patterns
- web prefetching
- access latency
- hit rate
- user perceived latency
- web documents
- cache misses
- web caching
- memory requirements
- caching scheme
- web page prediction
- higher order
- hit ratio
- web logs
- cache replacement
- web objects
- memory space
- computational power
- prediction accuracy
- replacement policy
- machine learning
- main memory
- mobile devices
- database systems