Reusing GEMM Hardware for Efficient Execution of Depthwise Separable Convolution on ASIC-Based DNN Accelerators.
Susmita Dey ManasiSuvadeep BanerjeeAbhijit DavareAnton A. SorokinSteven M. BurnsDesmond A. KirkpatrickSachin S. SapatnekarPublished in: ASP-DAC (2023)
Keyphrases
- efficient execution
- single chip
- hardware implementation
- hardware architecture
- field programmable gate array
- query optimization
- parallel execution
- computing systems
- query processor
- query processing
- database operations
- database systems
- query execution
- image processing algorithms
- computer systems
- embedded systems
- graphics processing units
- data partitioning
- data model
- data warehouse
- data sources
- spatial join
- database
- signal processing
- efficient implementation
- parallel processing