Back to browse
GitHub Repository

Single-file streaming JSON parser and query engine with GPU acceleration

1 starsC

jsoon, a streaming JSON parser and query engine in C

by enduku·Mar 27, 2026·3 points·0 comments

AI Analysis

●●●BangerWizardryBig Brain

GPU-accelerated JSON parsing with 91ns selective queries versus simdjson's 24ms.

Strengths
  • Streaming architecture skips irrelevant subtrees without building full DOM in memory
  • Optional CUDA path for structural indexing on large files with NVIDIA GPUs
  • Author is transparent about limitations and specific workload use cases honestly
Weaknesses
  • Limited query language compared to jq or full JSONPath implementations available
  • GPU path requires NVIDIA hardware and adds deployment complexity for teams
Target Audience

Backend developers, data engineers working with large JSON files

Similar To

simdjson · yyjson · jq

Post Description

Hi HN. I’ve been working on jsoon, a JSON query engine in one C file. The target use case is large JSON where I only need one or two fields out of the document. It doesn’t build a DOM. It scans for structure, skips subtrees when it can, and stops when it finds the result. The CPU path uses AVX2/ PCLMUL, and there’s also an optional CUDA path that does structural indexing on the GPU.

It has plenty of rough edges. The query language is limited, portability is limited, docs are still thin, and I would not treat it as a general-purpose replacement for simdjson, yyjson, or RapidJSON.

The benchmark numbers also need caveats. On selective queries jsoon can bail out early, while those libraries are doing full parsing and validation. So the large speedups are mostly about workload and architecture, not a claim that this is just a better JSON parser.

Posting because I think the implementation is interesting and I’d rather get criticism now than after spending more time on it. I’d especially like feedback on correctness, the SIMD/CUDA approach, and whether the benchmarks are framed in a fair way.

Similar Projects