Benchmark · Directional
LaTeX conversion with BibTeX
superdocuments' reader converts LaTeX to Markdown about 6 times faster than Pandoc 3.1.3. That is 5 to 13 times per file. Pandoc could not convert three of five BibTeX files. The fixtures are hand-made. The Webfold command line does not route the reader.
superdocuments' LaTeX reader against Pandoc 3.1.3. BibTeX reader too. Hand-made fixtures.
What question does this benchmark ask?
How fast does each tool turn LaTeX into Markdown? How fast for BibTeX references? Do they agree on structure?
How was it run?
- 16 hand-made LaTeX files and 5 BibTeX files. Median of 20 runs per command, process start included, one core.
- Pandoc 3.1.3 (-f latex -t markdown, -f bibtex -t csljson) against the superdocuments command line.
What are the results?
| Measurement | superdocuments | Pandoc |
|---|---|---|
| LaTeX to Markdown, median of 16 files | 3.97 ms | 23.97 ms (6.0 times) |
| Per-file speed-up, median and range | 7.1 times | 5.3 to 13.4 times |
| Structural differences | 0 on 14 of 16 files | 2 on one figures file |
| BibTeX files converted | 5 of 5 | 2 of 5 |
| BibTeX, the two files Pandoc converted | 4.02 and 3.27 ms | 23.99 and 53.15 ms (6.0 and 16 times) |
About 6 times faster than Pandoc on LaTeX, and 5 to 13 times on every file. Pandoc could not convert three of the five BibTeX files at all (two syntax errors, one a parenthesis-delimited entry and one a missing comma, and one out-of-memory on a cross-reference cycle).
What does this not show?
Know the limits
- Hand-made fixtures, 16 LaTeX and 5 BibTeX files. Not a TeX engine: packages are not loaded, and Pandoc covers many more packages and macros.
- Only two BibTeX files could be timed on both sides. The sources do not say which tool is right on the two disagreeing figures counts.
- Hardware is not recorded. The command-line tool only: this reader is in superdocuments and in the revision Webfold pins, but the Webfold command line does not route .tex files. The run is from 6 November, before the Pandoc-layout rewrite of the reader, and the speed was not measured again.
Which other benchmarks are there?
- HTML to Markdown conversion Measured
- Document conversion Measured
- Main-content extraction quality Measured
- Web archive extraction Measured
- Footprint of a scrape or crawl Directional
- PII detection against Presidio Measured
- Localisation extraction Directional
- PDF text extraction with OCR Directional
- PDF extraction against other extractors Directional
- A whole site to Markdown Directional
- Agent-browser compatibility (Lexbench) Measured
- Headless engine over CDP Directional
- Search-results parsing Directional
- Folder search Directional
Check it on your own pages.
Run the command-line tool on your content.