An RL agent with a graph neural network learns loop nest optimizations for the Tiramisu compiler and generalizes to unseen benchmarks, reporting 2.02x and 3.36x geometric mean speedups over Tiramisu and Pluto.
PENCIL: Towards a Platform-Neutral Compute Intermediate Language for DSLs
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
We motivate the design and implementation of a platform-neutral compute intermediate language (PENCIL) for productive and performance-portable accelerator programming.
fields
cs.PL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Pearl: Automatic Code Optimization Using Deep Reinforcement Learning
An RL agent with a graph neural network learns loop nest optimizations for the Tiramisu compiler and generalizes to unseen benchmarks, reporting 2.02x and 3.36x geometric mean speedups over Tiramisu and Pluto.