This work targets the middle ground between per-operator hand-written kernels and bespoke implementations synthesized for an entire workload or execution engine, and finds that strong code models can often generate buildable and equivalent fused Triton kernels, and that these kernels can improve end-to-end runtime when deployed opportunistically with fallback.
A hierarchical search-space planning framework for GPU kernel optimization that delivers stronger overall implementation validity, sample efficiency, and optimization performance than existing training-free methods, while remaining competitive with the training-based CUDA-L1 without additional model training is propose...
Jing-Hao Wang, Qiqi Gu, Chenpeng Wu et al.· 0 citations
DataKernelBench is introduced, which translates SQL into validated PyTorch TorchPlan programs and evaluates LLMs that optimize either the core tensor-bounded snippet or the full query in CUDA or Triton through execution-guided repair and finds that higher-performing implementations commonly use kernel fusion and execut...
Large language models show promise for code generation, yet their effectiveness in producing high-performance GPU kernels remains insufficiently understood. Recent agentic workflows improve over direct prompting through specialized generation roles and execution feedback, but our characterization of a representative wo...
Yang-Jie Zhou, Xing-Chen Ma, Zi-Han Liu et al.· Proceedings of the 41st IEEE...· 0 citations
The CUDA Code-General is presented, a query compilation framework for GPU database systems that allows selecting any of the 22 TPC-H benchmark queries or entering arbitrary SQL, to choose between runtime or static compilation and to inspect the generated CUDA code for three kernel fusion strategies side by side.
Maha Alwahibi, Bachana Jangebashvili, Maximilian E. Schüle· Proceedings of the VLDB Endo...· 0 citations
In-Copy Fusion (ICF), a runtime optimization of OpenMP’s accelerator model, implemented in LLVM, that gathers eligible map clauses into staging buffers, orchestrating an optimal transfer pipeline through fusion without modifying application or kernel code, preserves OpenMP semantics with negligible overhead is presente...
Dionisis-Odysseas Sotiropoulos, Sara Royuela Alcázar, Eduardo Quiñones et al.· Proceedings of the Internati...· 0 citations
KernelOPT is a multi-agent system that treats compiled models as structured artifacts and exclusively targets generated Triton sub-kernels using five profiling-guided LLM agents and achieves geometric mean speedups over PyTorch.
Aheli Poddar, Sanskar Prasad, Arindam Samanta et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
