Bend

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

19808 stars · 494 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Bend is a high-level, massively parallel programming language that feels like Python or Haskell but automatically parallelizes programs across multi-core CPUs and NVIDIA GPUs without explicit concurrency annotations. It is powered by HVM2, an interaction-combinator-based runtime that achieves near-linear speedup with core count.

Use cases

• write parallel code without locks, mutexes, or atomics
• run programs massively parallel on NVIDIA GPUs
• parallelize recursive and higher-order functional programs
• achieve near-linear speedup on multi-core CPUs
• compile Bend programs to standalone C or CUDA for maximum performance
• experiment with interaction net computation models

When to choose

• you want implicit, automatic parallelism on CPUs or NVIDIA GPUs
• your workload scales with core count and single-core performance is less critical
• you prefer a Python/Haskell-like high-level language over writing CUDA
• you want to explore interaction combinator-based computation

When to avoid

• you need Windows support natively (only WSL2 works)
• you need non-NVIDIA GPU support
• you need mature, optimized single-core performance comparable to GCC or GHC
• you need a production-stable language toolchain

Facets

cli-tool · maturity experimental

compiler interpreter concurrency gpu-computing programming-languages compilers gpu-computing developer-tools cli rust interaction-combinators functional-programming implicit-parallelism cuda hvm2 high-level-language programming-language linux macos gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="HigherOrderCO/Bend")