3 ms·
Bench-marking a fat native call, rather than a purely-interpreted function, is a bit bait-and-switch.
by mwkaufma 7y ago
Bench-marking a fat native call, rather than a purely-interpreted function, is a bit bait-and-switch.
- chrisseaton 7y agoThey do explain it very clearly though, and give a reason why that is what they are interested in for business purposes. Benchmark and optimise for what you're running, right? Better than a Mandelbrot?
- superdimwit 7y agoYeah, but they're just benchmarking a wrapper around gzip. It looks to be around 300 LOC and does very little.. It's a bit deceptive. A large table of benchmarks across different domains and workloads would be more honest.
- the_duke 7y agoThey explain it, but it's still deceptive. I would have expected a honest benchmark result. Which will be much worse than 6%. Which is totally fine! an interpreter will of course be much slower. But this seems like the wrong kind of marketing by omission to me.