PassNet provides a dataset of 18K graphs and PassBench for LLM-generated compiler passes, with fine-tuned models achieving 2.67x gains on long-tail tasks where TorchInductor underperforms.
Compilergym: Robust, performant compiler optimization environments for ai research
7 Pith papers cite this work, alongside 74 external citations. Polarity classification is still indexing.
7
Pith papers citing it
74
external citations · external index
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
A new abstract interpretation algorithm enables sound optimistic analysis of e-graphs during equality saturation, unifying it with non-destructive rewriting and improving precision on cyclic SSA programs.
CTA-pipelining reduces 2-layer GEMM latency up to 31.8% vs micro-batching and 29.6% vs Tensor Parallelism on 8-GPU H200/B200 systems by CTA-level cross-GPU pipelining.
WingSpan adds concurrency wrappers and a dependence theory to the Finch tensor compiler, enabling parallel sparse tensor kernels with sparse outputs.
R^3 optimizes full scientific applications on GPUs better than tuning kernel parameters or compiler flags alone while running nearly an order of magnitude faster than modern evolutionary search methods.
AEGIS reduces inter-GPU communication by up to 81.3% in self-attention and reaches 96.62% scaling efficiency with 3.86x speedup on four GPUs for 2048-token encrypted Transformer inference.
Clef compiler applies fixed-point scaffolding and a functor from compilation poset to target category to preserve dimensional, grade, escape and numeric structure through MLIR lowering while adding compact-closed negative and fractional types.
citing papers explorer
-
PassNet: Scaling Large Language Models for Graph Compiler Pass Generation
PassNet provides a dataset of 18K graphs and PassBench for LLM-generated compiler passes, with fine-tuned models achieving 2.67x gains on long-tail tasks where TorchInductor underperforms.
-
Optimism in Equality Saturation
A new abstract interpretation algorithm enables sound optimistic analysis of e-graphs during equality saturation, unifying it with non-destructive rewriting and improving precision on cyclic SSA programs.
-
CTA-Pipelining: A Latency-Oriented Spatial Scaling Method for Multi-GPU Systems
CTA-pipelining reduces 2-layer GEMM latency up to 31.8% vs micro-batching and 29.6% vs Tensor Parallelism on 8-GPU H200/B200 systems by CTA-level cross-GPU pipelining.
-
WingSpan: Concurrency and Dependence for Sparse and Structured Tensor Compilers
WingSpan adds concurrency wrappers and a dependence theory to the Finch tensor compiler, enabling parallel sparse tensor kernels with sparse outputs.
-
Record-Remix-Replay: Hierarchical GPU Kernel Optimization using Evolutionary Search
R^3 optimizes full scientific applications on GPUs better than tuning kernel parameters or compiler flags alone while running nearly an order of magnitude faster than modern evolutionary search methods.
-
AEGIS: Scaling Long-Sequence Homomorphic Encrypted Transformer Inference via Hybrid Parallelism on Multi-GPU Systems
AEGIS reduces inter-GPU communication by up to 81.3% in self-attention and reaches 96.62% scaling efficiency with 3.86x speedup on four GPUs for 2048-token encrypted Transformer inference.
-
Fixed-Point Scaffolding in the Clef Programming Language
Clef compiler applies fixed-point scaffolding and a functor from compilation poset to target category to preserve dimensional, grade, escape and numeric structure through MLIR lowering while adding compact-closed negative and fractional types.