FastKernels is a production-aligned benchmark covering 96.2% of HuggingFace Transformers that reveals state-of-the-art kernel agents deliver at most 0.94x aggregate speedup.
hub
Kevin: Multi-turn rl for generating cuda kernels
14 Pith papers cite this work. Polarity classification is still indexing.
14
Pith papers citing it
hub tools
citation-role summary
baseline 2
background 1
other 1
citation-polarity summary
years
2026 14representative citing papers
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.
Step-TP is a dataset providing grounded, atomic step-level IR transitions and CoT supervision to enable reliable multi-step LLM-guided tensor program optimization instead of end-to-end imitation.
CUDAHercules benchmark demonstrates that leading LLMs generate functional CUDA code but fail to recover expert-level optimization strategies needed for peak performance on Ampere, Hopper, and Blackwell GPUs.
Hawk raises NPU kernel generation accuracy from 49.4% to 80% and yields up to 2.2× speedups by retrieving and distilling structured hardware-aware knowledge without any model training.
SpecGen introduces speculative generation to fork non-reasoning kernel candidates during LLM reasoning traces, enabling early termination and parallel profiling to reduce end-to-end optimization time on H200 GPUs.
KLineage derives verified optimization skills from backward lineages of expert GPU kernels to guide LLM agents toward higher-quality and more efficient kernels than memory-based baselines.
CUDABEAVER benchmark and pass@k(M,C,A) metric show LLM CUDA debugging success drops by up to 40 percentage points under strict performance requirements.
A diagnosis-driven evolutionary search with retrieval-augmented expert initialization improves Triton kernel correctness and speed on KernelBench, reaching 99–100% correctness on Level 2.
AdaExplore improves correctness and speed of Triton kernel generation by converting recurring failures into a memory of rules and organizing search as a tree that mixes local refinements with larger regenerations, yielding 3.12x and 1.72x speedups on KernelBench Level-2 and Level-3 within 100 steps.
InCoder-32B-Thinking uses error-feedback synthesized thinking traces and a code world model to reach top open-source scores on general and industrial code benchmarks including 81.3% on LiveCodeBench and 84.0% on CAD-Coder.
Kernel-Smith combines evolutionary search with RL post-training to generate optimized GPU kernels, achieving SOTA speedups on KernelBench that beat Gemini-3.0-pro and Claude-4.6-opus on NVIDIA Triton and generalize to MetaX MACA.
The FIL Hypothesis claims that inductive biases outperform purely data-driven methods on GPU programming tasks with non-trivial feedback loops.
AscendOptimizer combines kernel rewinding for reusable experience with evolutionary search on hardware feedback to optimize Ascend NPU operators, delivering 1.21x geometric-mean speedup and faster performance on 53.47% of 101 tested operators versus baseline.
citing papers explorer
-
FastKernels: Benchmarking GPU Kernel Generation in Production
FastKernels is a production-aligned benchmark covering 96.2% of HuggingFace Transformers that reveals state-of-the-art kernel agents deliver at most 0.94x aggregate speedup.
-
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.
-
Step-TP: A Grounded, Step-Level Dataset with Chain-of-Thought Reasoning for LLM-Guided Tensor Program Optimization
Step-TP is a dataset providing grounded, atomic step-level IR transitions and CoT supervision to enable reliable multi-step LLM-guided tensor program optimization instead of end-to-end imitation.
-
CUDAHercules: Benchmarking Hardware-Aware Expert-level CUDA Optimization for LLMs
CUDAHercules benchmark demonstrates that leading LLMs generate functional CUDA code but fail to recover expert-level optimization strategies needed for peak performance on Ampere, Hopper, and Blackwell GPUs.
-
Hawk: Harnessing Hardware-Aware Knowledge for High-Performance NPU Kernel Generation
Hawk raises NPU kernel generation accuracy from 49.4% to 80% and yields up to 2.2× speedups by retrieving and distilling structured hardware-aware knowledge without any model training.
-
SpecGen: Accelerating Agentic Kernel Optimization with Speculative Generation
SpecGen introduces speculative generation to fork non-reasoning kernel candidates during LLM reasoning traces, enabling early termination and parallel profiling to reduce end-to-end optimization time on H200 GPUs.
-
Learning When to Optimize: Verified Optimization Skills from Expert GPU-Kernel Lineages
KLineage derives verified optimization skills from backward lineages of expert GPU kernels to guide LLM agents toward higher-quality and more efficient kernels than memory-based baselines.
-
CUDABeaver: Benchmarking LLM-Based Automated CUDA Debugging
CUDABEAVER benchmark and pass@k(M,C,A) metric show LLM CUDA debugging success drops by up to 40 percentage points under strict performance requirements.
-
Kernel Foundry: A Diagnosis-driven Evolutionary Kernel Optimizer with Multi-Experts
A diagnosis-driven evolutionary search with retrieval-augmented expert initialization improves Triton kernel correctness and speed on KernelBench, reaching 99–100% correctness on Level 2.
-
AdaExplore: Failure-Driven Adaptation and Diversity-Preserving Search for Efficient Kernel Generation
AdaExplore improves correctness and speed of Triton kernel generation by converting recurring failures into a memory of rules and organizing search as a tree that mixes local refinements with larger regenerations, yielding 3.12x and 1.72x speedups on KernelBench Level-2 and Level-3 within 100 steps.
-
InCoder-32B-Thinking: Industrial Code World Model for Thinking
InCoder-32B-Thinking uses error-feedback synthesized thinking traces and a code world model to reach top open-source scores on general and industrial code benchmarks including 81.3% on LiveCodeBench and 84.0% on CAD-Coder.
-
Kernel-Smith: A Unified Recipe for Evolutionary Kernel Optimization
Kernel-Smith combines evolutionary search with RL post-training to generate optimized GPU kernels, achieving SOTA speedups on KernelBench that beat Gemini-3.0-pro and Claude-4.6-opus on NVIDIA Triton and generalize to MetaX MACA.
-
The FIL Hypothesis: Inductive Biases Help with Kernel Engineering
The FIL Hypothesis claims that inductive biases outperform purely data-driven methods on GPU programming tasks with non-trivial feedback loops.
-
AscendOptimizer: Episodic Agent for Ascend NPU Operator Optimization
AscendOptimizer combines kernel rewinding for reusable experience with evolutionary search on hardware feedback to optimize Ascend NPU operators, delivering 1.21x geometric-mean speedup and faster performance on 53.47% of 101 tested operators versus baseline.