Debunking "Faster than Rust": Malicious Benchmarks in Tov
Web page · https://blog.fyralabs.com/debunking-faster-than-rust-malicious-benchmarks-in-tov
First seen by Oriel on .
AI summary
The post challenges Tov’s “faster than Rust” claims, arguing that three benchmark comparisons disadvantage Rust through inequivalent implementations and showing optimized Rust versions matching or outperforming Tov.
- For bigint_fib, the authors could not reproduce Tov’s claimed 168 ms versus Rust’s 501 ms; their results varied by machine, with Rust faster on two setups and Tov slightly faster on another.
- The dispatch benchmark uses Vec<Box<dyn Shape>> and dynamic dispatch in Rust; an enum-based Rust version removes that overhead and runs about as fast as or faster than Tov across the reported setups.
- For map_insert, the authors compare Rust’s HashMap with an AHashMap version that reuses prebuilt keys; the latter outperforms Tov in all three reported runs.
- The authors conclude the benchmark implementations are inequivalent and the results are misleading, while attributing deliberate intent to Tov’s developers without establishing that claim independently.
#rust #programming languages #benchmarking #tov
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Lobsters: 46, latest observation ; +45 over observed period since
Comments
Comment count reported by the source.
- Lobsters: 17, latest observation ; +17 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Lobsters | gajtrm | https://blog.fyralabs.com/debunking-faster-than-rust-malicious-benchmarks-in-tov/ Read discussion |