A5101887660- Parallel Computing and Optimization Techniques
- Embedded Systems Design Techniques
- Advanced Neural Network Applications
- Seismic Imaging and Inversion Techniques
- Reservoir Engineering and Simulation Methods
- Medical Imaging Techniques and Applications
- Advanced Data Storage Technologies
- Interconnection Networks and Systems
- Distributed and Parallel Computing Systems
Universidade Estadual de Campinas (UNICAMP)
2019-2023
Convolution is one of the most computationally intensive operations that must be performed for machine learning model inference. A traditional approach to computing convolutions known as Im2Col + BLAS method. This article proposes SConv: a direct-convolution algorithm based on an MLIR/LLVM code-generation toolchain can integrated into machine-learning compilers. introduces: (a) Slicing Analysis (CSA)—a convolution-specific 3D cache-blocking analysis pass focuses tile reuse over cache...
Convolution is one of the most computationally intensive operations that must be performed for machine-learning model inference. A traditional approach to compute convolutions known as Im2Col + BLAS method. This paper proposes SConv: a direct-convolution algorithm based on MLIR/LLVM code-generation toolchain can integrated into compilers . introduces: (a) Slicing Analysis (CSA) - convolution-specific 3D cache-blocking analysis pass focuses tile reuse over cache hierarchy; (b) Optimization...