Benchmark: gzipped JSON matches binary formats in size, decompresses at ~270MB/s per core
cis_female · x · 2026-09-21
Arguing against "JSON is unsuitable for performance-critical paths," the author shows that after GZIP compression JSON's size difference vs binary formats like protobuf/msgpack is negligible, with decompression + parsing at a median 270MB/s per core on an M4 — while many binary formats top out around 500MB/s anyway, so compression is mandatory regardless. On the browser side, options are limited: direct casting is unsafe, struct-type solutions too custom, XML effectively dead, and protobuf/msgpack workable but complex. A reproducible benchmark gist is included.
More from Infra
- Qdrant experiments: 10→500 candidate depth lifts best-possible nDCG by 0.28 but real score by ≤0.01 — qdrant_engine · 2026-09-21
- gemini-cli PR Fixes Process Hang on Exit via stdin and MCP Child Process Cleanup — Pcmhacker-piro · 2026-09-21
- Daniel Lemire Tests Whether CPUs Can Take More Than One Branch Per Cycle — lemire · 2026-09-21
- Qwen3.8-27B in native 8-bit hits 37-55 tok/s on Apple Silicon, avoiding the 4-bit reasoning cliff — SnooPredictions515 · 2026-09-21
- Google's Orphaned VMs Patches Keep VMs Running While Host Kernel Goes Offline — jedisct1 · 2026-09-21
- djev-run Deploys DiffusionGemma on Cloud Run's RTX PRO 6000 Blackwell for ~$3/hr — bodonoghue85 · 2026-09-21