FlinkCL: An OpenCL-Based In-Memory Computing Architecture on Heterogeneous CPU-GPU Clusters for Big Data.
Cen ChenKenli LiAijia OuyangKeqin LiPublished in: IEEE Trans. Computers (2018)
Keyphrases
- big data
- graphics processing units
- heterogeneous computing
- compute intensive
- commodity hardware
- multithreading
- memory bandwidth
- memory access
- general purpose computing
- shared memory
- memory hierarchy
- general purpose
- parallel computing
- cloud computing
- graphic processing unit
- high volume
- real time
- graphics processors
- data management
- level parallelism
- big data analytics
- social computing
- data analysis
- gpu implementation
- business intelligence
- parallel processing
- parallel programming
- data processing
- processing units
- computing systems
- single instruction multiple data
- knowledge discovery
- memory management
- vast amounts of data
- unstructured data
- social media
- data mining
- intel xeon
- data science
- data transfer
- parallel computation
- parallel algorithm
- data access
- efficient implementation
- massive datasets
- massively parallel
- floating point
- parallel implementation
- high performance computing
- main memory
- data warehousing
- grid computing
- instruction set
- parallel execution
- case study
- decision making
- social networks
- data sets
- database