https://neper.dev/blog/ Notes from measuring Neper: memory, WebAssembly, desktop toolchains, text rendering, charts, databases and LLM cost, with the columns where it loses. en Fri, 30 Oct 2026 00:00:00 +0000 https://neper.dev/gtm/blog-og.png https://neper.dev/blog/ https://neper.dev/blog/compiler-memory.html https://neper.dev/blog/compiler-memory.html How a Neper program grows into memory with no flag: an arena reserved, not allocated, and the measured limits on Windows and Linux. Fri, 30 Oct 2026 00:00:00 +0000 https://neper.dev/blog/neper-webassembly.html https://neper.dev/blog/neper-webassembly.html Neper's own compiler writes wasm32-wasi modules with no LLVM: an 18 KB hello world, a 1.7 MB window and the compiler running in a browser. Tue, 27 Oct 2026 00:00:00 +0000 https://neper.dev/blog/desktop-toolchains.html https://neper.dev/blog/desktop-toolchains.html One compiler for Windows and Linux desktop programs, nothing to install, a 2.7 MiB window: compared with ten languages install guides. Fri, 23 Oct 2026 00:00:00 +0000 https://neper.dev/blog/why-neper-is-optimized-for-llms.html https://neper.dev/blog/why-neper-is-optimized-for-llms.html Search, generate, edit, check, repair: what each step costs a model and the language rules, JSONL commands and diagnostics that make it cheap. Wed, 21 Oct 2026 00:00:00 +0000 https://neper.dev/blog/cleartype-text.html https://neper.dev/blog/cleartype-text.html What subpixel ClearType text costs: measured glyph rendering against native Windows captures, and the trade-offs. Sun, 18 Oct 2026 00:00:00 +0000 https://neper.dev/blog/charts-speed-and-png-size.html https://neper.dev/blog/charts-speed-and-png-size.html How fast 234 chart types render, and why PNG size and render speed trade off: measured times and file sizes for every chart. Tue, 13 Oct 2026 00:00:00 +0000 https://neper.dev/blog/charts-field-guide.html https://neper.dev/blog/charts-field-guide.html A field guide to all 234 charts e.gfx.chart can draw: which question each answers, when to pick it, and speed against matplotlib. Sun, 04 Oct 2026 00:00:00 +0000 https://neper.dev/blog/sc1m-eight-compilers.html https://neper.dev/blog/sc1m-eight-compilers.html The sc1m program (1,000 modules, 920,554 lines) in eight compilers: compile and run times with verified output, and where each language lands. Fri, 02 Oct 2026 00:00:00 +0000 https://neper.dev/blog/six-algorithms.html https://neper.dev/blog/six-algorithms.html A backtracking regex, mixed-integer programming, model checking, crash-dump parsing, Windows and Kerberos sign-on, and AC-2001: one program each. Tue, 29 Sep 2026 00:00:00 +0000 https://neper.dev/blog/sqlserver-openssl.html https://neper.dev/blog/sqlserver-openssl.html x.microsoft.tds hands TLS to the host's OpenSSL at run time: SQL Server scans 3.2-3.6x faster, and Neper's scan leads C, Go and Rust. Tue, 29 Sep 2026 00:00:00 +0000 https://neper.dev/blog/e-db-drivers.html https://neper.dev/blog/e-db-drivers.html e.db, Neper's SQL contract, and its SQLite, PostgreSQL, MySQL, ODBC and SQL Server drivers: what differs, and one measured workload. Sun, 27 Sep 2026 00:00:00 +0000 https://neper.dev/blog/neper-vs-dart-llm-cost.html https://neper.dev/blog/neper-vs-dart-llm-cost.html Seven days of Claude transcripts measured byte by byte: what Neper and Dart cost per KB, and where each wins on edits, context and repair. Sun, 27 Sep 2026 00:00:00 +0000