Glossary · Term

KernelBench

← all terms

Definition

Plain language

A benchmark for testing whether AI systems can write fast custom GPU code.

As stated in the literature

A benchmark of CUDA kernel generation tasks measuring whether generated kernels match or beat hand-written PyTorch references in correctness and speedup.

Why it matters: It measures whether AI systems can actually produce performant low-level GPU code, which is a high-leverage capability for ML systems work.

For example, a model is asked to write a CUDA kernel for matrix multiplication, and is scored on whether its output matches PyTorch's result and how much faster (or slower) it runs.

Heard on the show

“… On KernelBench — that's the standard benchmark, two hundred fifty kernel tasks across three difficulty tiers, …”
Episode 177 — Why Raw Profiler Data Made an AI Worse at Writing GPU Code

Mentioned in 3 episodes

  1. 177
    Why Raw Profiler Data Made an AI Worse at Writing GPU Code
  2. 124
    A Cheap Model With the Blueprints Beats Expensive Models Working Blind
  3. 065
    One Loop to Optimize Them All: A Universal API for LLM-Driven Discovery

Related terms