Comments (2)
Normal CI (i.e. cargo check && cargo test
) takes about 10 minutes and runs on a single Linux host for every push to a PR. This can normally drop to 2-5 minutes per run if the cache is warm, but the cache is invalidated after a day because we use nightly.
These builds are by far the most important and need to be run. That said, it's in our best interest to minimise build times so everyone gets rapid feedback on their changes.
The benchmark action takes about 30 minutes on a Linux host and is set to run every time there is a commit to master
.
We don't really use the results from these runs, so I'd be happy with changing the trigger to be either weekly or just for tags.
Notably, #76 (if merged) will introduce another set of builds because it builds the entire repo and packages it up so people aren't compiling things from source all the time.
Ideally we'd run the action every night (assuming there have been changes) and for Windows/Linux/Mac, but once you consider each build takes about 15 minutes and that Windows and Mac builds consume 2x and 10x minutes, that job alone will cost us (15 + 15*2 + 15*10) * 7 = 1365
CI minutes/week.
It could be tweaked so we run Linux nightly and only do the other builds when we do a release (i.e. a tagged commit is pushed to master
). That would drop the nightly
action to a more realistic 105 minutes/week.
from rune.
GitHub Actions are all set up and we get it for free now the Rune repo is public.
from rune.
Related Issues (20)
- Enable Apple targets for librune_native in Github actions
- Consistent Tensor element types
- Allow people to attach a description to a capability
- Allow each Rune to have a description
- Make the compiler's internals deterministic
- Missing dependency libclang causes build to fail and hang HOT 3
- Let people specify supported runtime environments/backends in the Runefile
- Split the rune CLI up so there are individual binaries for each sub-command
- Caching strategy for long running processes like BERT QA inference HOT 3
- Add support for inference on ONNX and TensorFlow models HOT 1
- Make it easier to pass inputs through the builder API
- The "rune build" command completely ignores the "image" field
- The web runtime should try to cache Runes where possible
- Add rune serve to the cli
- Structured Logging HOT 1
- Load tf.js models using a custom loader
- BigUint64Array is not supported on some versions of Safari
- Random capability should be float
- Build error when connecting image source with normalisation block HOT 2
- Adding Observability Probes (Behind the scenes) to inputs to Runes
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from rune.