Publication | Closed Access
AKG: automatic kernel generation for neural processing units using polyhedral transformations
66
Citations
62
References
2021
Year
Unknown Venue
EngineeringMachine LearningComputer ArchitectureGpu ComputingComplicated Memory HierarchyParallel ComputingAutomatic Kernel GenerationNeurocomputersTensor CompilersComputer EngineeringComputer ScienceDeep LearningNeural Architecture SearchGpu ClusterPolyhedral TransformationsDeep Neural NetworksHardware AccelerationComputational NeuroscienceReproducing Kernel MethodParallel ProgrammingBrain-like ComputingKernel Method
Existing tensor compilers have proven their effectiveness in deploying deep neural networks on general-purpose hardware like CPU and GPU, but optimizing for neural processing units (NPUs) is still challenging due to the heterogeneous compute units and complicated memory hierarchy.
| Year | Citations | |
|---|---|---|
Page 1
Page 1