Small cleanup on README.md/comments [ci skip]

Co-authored-by: Jerry Zhao <jerryz123@berkeley.edu>
This commit is contained in:
Abraham Gonzalez
2021-10-13 23:13:52 -07:00
committed by GitHub
parent 019435508b
commit 2fb3dc7c42
2 changed files with 2 additions and 3 deletions

4
.github/README.md vendored
View File

@@ -7,7 +7,7 @@ GitHub Actions Brief Explanation
--------------------------- ---------------------------
CI is executed by Github Actions (GA). GA is controlled by `.yml` files in the `.github/workflows/` directory. CI is executed by Github Actions (GA). GA is controlled by `.yml` files in the `.github/workflows/` directory.
In our case we have just one workflow named `chipyard-rocket-run-tests.yml`. In our case, we have just one workflow named `chipyard-run-tests.yml`.
It defines a number of `jobs` within it that do particular tasks. It defines a number of `jobs` within it that do particular tasks.
All jobs in the workflow must pass for the CI run to be successful. All jobs in the workflow must pass for the CI run to be successful.
In general, a job is run in parallel with others unless it depends on some other job. In general, a job is run in parallel with others unless it depends on some other job.
@@ -120,7 +120,7 @@ Note: On the remote server you need to have the `*.pub` key file added to the `a
Notes on CIRCLE CI Notes on CIRCLE CI
------------------ ------------------
This code is heavily based on the origin [CircleCI]() work. There a quite a few differences This code is heavily based on the origin [CircleCI]() work. There a quite a few differences
- CCI supports workflow level variables, in GA we must define thiing like `BUILDSERVER: ${{ secrets.BUILDSERVER }}` in every job - CCI supports workflow level variables, in GA we must define things like `BUILDSERVER: ${{ secrets.BUILDSERVER }}` in every job
- CCI allows a much larger cache. The entire CY directory with toolchains and RTL could be cached, with GA there is a 5Gb total cache limit - CCI allows a much larger cache. The entire CY directory with toolchains and RTL could be cached, with GA there is a 5Gb total cache limit
- GA support more parallel jobs 20 vs 4 - GA support more parallel jobs 20 vs 4
- GA seems to allow much longer run times - GA seems to allow much longer run times

View File

@@ -9,7 +9,6 @@ set -ex
SCRIPT_DIR="$( cd "$( dirname "$0" )" && pwd )" SCRIPT_DIR="$( cd "$( dirname "$0" )" && pwd )"
source $SCRIPT_DIR/defaults.sh source $SCRIPT_DIR/defaults.sh
# enter bhd repo
cd $LOCAL_CHIPYARD_DIR cd $LOCAL_CHIPYARD_DIR
# ignore the private vlsi submodules # ignore the private vlsi submodules