search

ETLCPP / etl

Embedded Template Library
https://www.etlcpp.com
MIT License
2.26k stars 393 forks source link

Speed up builds #911

Closed jaskij closed 3 months ago

jaskij commented 4 months ago

Something I noticed is that current GHA workflows simply call make - which defaults to running the jobs sequentially. Meanwhile, GitHub's public runners have more than one CPU core.

Your builds could probably be significantly sped up by changing the make call to something like make -j4, or more universally make -j $(nproc). It could also be possible to switch over to ninja (which defaults to one job per CPU thread), but I don't think it's installed by default in GHA images.

jaskij commented 4 months ago

I'll take care of this one once #907 is merged.

It also seems to me that there's a lot of opportunity for deduplication in workflow code, utilizing matrices and environment variables.

jwellbelove commented 4 months ago

It also seems to me that there's a lot of opportunity for deduplication in workflow code, utilizing matrices and environment variables.

You're probably correct. I look forward to seeing what you propose.