Conversation
thomasp85
left a comment
thomasp85
left a comment
There was a problem hiding this comment.
Maybe add a new commit here to see if it picks up the cache? If it does you can merge at your leisure
|
Just noting down results of run 1: build: 14m49s Restarting the jobs manually for run 2. |
|
Second run has: build: 14m48s This is not a noteworthy improvement. |
|
Times for the uncached run with the new step: build: 15m25s For the cached run: build: 3m33s |
|
it seems the build at least was succesful |
|
but even with the cache the Jupyter kernel needs a full rebuild... |
|
Yes, I'm going to try to give each job individual keys next. So they don't all need to build from the same cache, but each can cache its own stuff. |
|
Uncached run: build: 14m |
|
Cached run: build: 4m15s I'm going to call this a success |
|
Fantastic! |
2 participants
Current GHA takes a long time, maybe we can shave off a good chunk?