Limiting the memory footprint when dynamically scheduling DAGs on shared-memory platforms.
Loris MarchalBertrand SimonFrédéric VivienPublished in: J. Parallel Distributed Comput. (2019)
Keyphrases
- shared memory
- memory footprint
- parallel machines
- heterogeneous platforms
- parallel algorithm
- message passing
- distributed memory
- memory usage
- parallel computing
- scheduling problem
- multi processor
- directed acyclic graph
- parallel programming
- scheduling algorithm
- massively parallel
- parallel architecture
- parallel architectures
- address space
- single processor
- shared memory multiprocessors
- graphic processing unit
- parallel execution
- memory access
- compute unified device architecture
- real time database systems
- parallel computers
- database systems