Add some basic benchmarks #27
Conversation
mordae
force-pushed
the
bench
branch
2 times, most recently
from
20dc3f9 to
2e6ccea
Compare
They measure almost all possible combinations of conversions for short, medium and large strings. String conversions were omitted, because they are by definition always hopelessly slow. Signed-off-by: Jan Hamal Dvořák <[email protected]>
There was a problem hiding this comment.
I'm not sure how much value benchmarks in string-conversions would have. This package just relays to other conversion functions from other libraries. So if one of the conversions is slow, there's not a lot that we could do about it.
Why do you think it'd be useful to add these?
I saw that you added e.g. altSBStoLT. Are these alternative conversions that are different from the ones we use currently? If yes, is any of the alternatives faster?
|
Alternative I believe that since we are picking a default of sorts for the user, we should make sure to pick the best conversion approach and this choice should be informed -- at least a little -- by some measurements. For easy to grok HTML output use: cabal run bench -- -o benchmark.html |
They measure almost all possible combinations of conversions for short, medium and large strings.
String conversions were omitted, because they are by definition always hopelessly slow.