Mapping Parallelism in a Functional IR through. Constraint Satisfaction. A Case Study on Convolution for Mobile GPUs. Naums Mogers. University of Edinburgh, UK.
Mar 18, 2022 · This paper proposes to extract parallelization constraints automatically from a functional IR and use a solver to identify valid rewriting.
This paper tackles the problem by modeling parallelization restrictions in Lift using constraints and finding valid mappings using a solver. 3 Convolution in ...
Mapping parallelism in a functional IR through constraint satisfaction: a case study on convolution for mobile GPUs ; Naums Mogers at The University of Edinburgh.
Mapping Parallelism in a Functional IR through. Constraint Satisfaction. A Case Study on Convolution for Mobile GPUs. Naums Mogers. University of Edinburgh, UK.
Satisfaction: A Case Study on Convolution for Mobile GPUs. ... analysis through rich representation, which ... functional GPU-programming with nested parallelism.
Dive into the research topics of 'Mapping Parallelism in a Functional IR through Constraint Satisfaction: A Case Study on Convolution for Mobile GPUs'. Together ...
Mapping parallelism in a functional IR through constraint satisfaction: a case study on convolution for mobile GPUs. N Mogers, L Li, V Radu, C Dubach.
Mar 2, 2022 · Using a convolutional neural network on a mobile GPU as a use case, this approach matches the performance of the ARM Compute Library GEMM ...
Missing: study | Show results with:study
This paper proposes to extract parallelization constraints automatically from a functional IR and use a solver to identify valid rewriting and match the ...