GFlink: An In-Memory Computing Architecture on Heterogeneous CPU-GPU Clusters for Big Data.
Cen ChenKenli LiAijia OuyangZhuo TangKeqin LiPublished in: ICPP (2016)
Keyphrases
- big data
- heterogeneous computing
- compute intensive
- memory bandwidth
- multithreading
- memory access
- commodity hardware
- level parallelism
- social computing
- high volume
- intel xeon
- graphics processors
- memory hierarchy
- single instruction multiple data
- data management
- graphics processing units
- cloud computing
- real time
- data analysis
- data processing
- big data analytics
- social media
- vast amounts of data
- gpu implementation
- data transfer
- business intelligence
- processing units
- unstructured data
- parallel programming
- graphic processing unit
- knowledge discovery
- parallel computing
- management system
- memory management
- data access
- general purpose computing
- data science
- data mining
- parallel implementation
- grid computing
- parallel processing
- data warehousing
- shared memory
- general purpose
- information retrieval
- massive datasets
- main memory
- information processing
- end users
- database systems
- social networks
- data driven decision making