Closed PragTob closed 9 months ago
Mental notes:
Name ips average deviation median 99th %
sequential_output 240.79 4.15 ms ±5.70% 4.13 ms 4.34 ms
format & write 240.00 4.17 ms ±3.71% 4.13 ms 4.82 ms
Formatter.output 228.23 4.38 ms ±16.58% 4.07 ms 6.34 ms
Comparison:
sequential_output 240.79
format & write 240.00 - 1.00x slower +0.0138 ms
Formatter.output 228.23 - 1.06x slower +0.23 ms
might want to provide "input_names" to use instead of the broken list :eyes:
adjust docs & Changelog
:partying_face:
Name ips average deviation median 99th %
Formatter.output 248.71 4.02 ms ±4.58% 4.03 ms 4.59 ms
sequential_output 248.38 4.03 ms ±3.16% 4.00 ms 4.53 ms
format & write 247.77 4.04 ms ±6.76% 4.02 ms 4.35 ms
Comparison:
Formatter.output 248.71
sequential_output 248.38 - 1.00x slower +0.00539 ms
format & write 247.77 - 1.00x slower +0.0153 ms
I definitely should get exguard working again... the amount of force pushes due to minor credo things are embarrassing ._.
:wave: Hey @hrzndhrn @NickNeck as you have some of the most Benchee plugins I'd appreciate it if you took a look here, mainly from the PoV on impact for plugins. If you want to review the PR as a whole of course I'm happy too :grin:
TLDR;
Scenario.function
, Scenario.input
and Configuration.inputs
will only retain the names, not the valuesThanks!
@PragTob thanks for the hint. Thanks for Benchee and all your work. I will take a look during the weekend.
@NickNeck thanks for all your great OSS work and also specifically of creating Benchee plugins, it was designed to facillitate and allow this but seeing all the things people do warms my heart! 💚
@NickNeck thanks a lot! :green_heart:
Should combat memory consumption issues caused by sending too much data to processes.
edit: To highlight the severeness of this:
:save
to compare performance across elixir version, before this PR each individual save file was 200+MB. With this MR, it's 1MB (as it removes the data from formatters and:save
is a formatter under the hood). The file size isn't the major thing, but it illustrates how much memory we save and how much time copying memory we avoid.I wanna write up all of this in like 2 blog posts but as per usual my speed is abysmal 😁