* ci: replace OpenVINO installer action
To test wasi-nn, we currently use an OpenVINO backend. The Wasmtime CI
must install OpenVINO using a custom GitHub action. This CI action has
not been updated in some time and in the meantime OpenVINO (and the
OpenVINO crates) have released several new versions.
https://github.com/abrown/install-openvino-action is an external action
that we plan to keep up to date with the latest releases. This change
replaces the current CI action with that one.
* wasi-nn: upgrade openvino dependency to v0.4.1
This eliminates a `lazy_static` dependency and changes a few parameters
to pass by reference. Importantly, it enables support for the latest
versions of OpenVINO (v2022.*) in wasi-nn.
* ci: update wasi-nn script to source correct env script
* ci: really use the correct path for the env script
Also, clarify which directory OpenVINO is installed in (the symlink may
not be present).
* Refactor binary-compatible-builds for releases
I was poking around this yesterday and noticed a few things that could
be improved for our release builds:
* The centos container for the x86_64 builds contained a bunch of extra
tooling we no longer need such as python3 and a C++ compiler. Along
with custom toolchain things this could all get removed since the C we
include now is quite simple.
* The aarch64 and s390x cross-compiled builds had relatively high glibc
version requirements compared to the x86_64 build. This was because we
don't use a container to build the cross-compiled binaries. I added
containers here along the lines of the x86_64 build to use an older
glibc to build the release binary to lower our version requirement.
This lower the aarch64 version requirement from glibc 2.28 to 2.17.
Additionally the s390x requirement dropped from 2.28 to 2.16.
* To make the containers a bit easier to read/write I added
`Dockerfile`s for them in a new `ci/docker` directory instead of
hardcoding install commands in JS.
This isn't intended to be a really big change or anything for anyone,
but it's intended to keep our Linux-based builds consistent at least as
best we can.
* Remove temporary change
Updates the OpenVINO backend for wasi-nn to pre-configure all inputs as `NHWC`. This is not a long-term fix but no worse than the status quo, which configures all input tensors to the same layout. This change updates the CI script to use the latest version of OpenVINO. Closes#3948.
Co-authored-by: Andrew Brown <andrew.brown@intel.com>
This commit removes the publish step in GitHub actions, insteading
folding all functionality into the release build steps. This avoids
having a separately scheduled job after all the release build jobs which
ends up getting delayed for quite a long time given the current
scheduling algorithm.
This involves refactoring the tarball assembly scripts and refactoring the
github asset upload script too. Tarball assembly now manages everything
internally and does platform-specific bits where necessary. The upload
script is restructured to be run in parallel (in theory) and hopefully
catches various errors and tries to not stomp over everyone else's work.
The main trickiness here is handling `dev`, which is less critical for
correctness than than tags themselves.
As a small tweak build-wise the QEMU build for cross-compiled builders
is now cached unlike before where it was unconditionally built, shaving
a minute or two off build time.
Fixes a few issues that have been cropping up:
* Update `rustup` on Windows to latest to skip over the 1.24.1 installed
on GitHub Actions which can fail to install.
* Remove the no-longer-needed `define-llvm-env` action
* Install generic llvm/lldb packges instead of specific ones that may
migrate in versions over time.
CentOS 6 just went EOL at the end of November 2020; as of today, the
repository seems to have disappeared, so our CI builds are failing. This
PR updates us to CentOS 7, which should be usable until June 30, 2024.
* Add an initial wasi-nn implementation for Wasmtime
This change adds a crate, `wasmtime-wasi-nn`, that uses `wiggle` to expose the current state of the wasi-nn API and `openvino` to implement the exposed functions. It includes an end-to-end test demonstrating how to do classification using wasi-nn:
- `crates/wasi-nn/tests/classification-example` contains Rust code that is compiled to the `wasm32-wasi` target and run with a Wasmtime embedding that exposes the wasi-nn calls
- the example uses Rust bindings for wasi-nn contained in `crates/wasi-nn/tests/wasi-nn-rust-bindings`; this crate contains code generated by `witx-bindgen` and eventually should be its own standalone crate
* Test wasi-nn as a CI step
This change adds:
- a GitHub action for installing OpenVINO
- a script, `ci/run-wasi-nn-example.sh`, to run the classification example
* Move most wasmtime tests into one test suite
This commit moves most wasmtime tests into a single test suite which
gets compiled into one executable instead of having lots of test
executables. The goal here is to reduce disk space on CI, and this
should be achieved by having fewer executables which means fewer copies
of `libwasmtime.rlib` linked across binaries on the system. More
importantly though this means that DWARF debug information should only
be in one executable rather than duplicated across many.
* Share more build caches
Globally set `RUSTFLAGS` to `-Dwarnings` instead of individually so all
build steps share the same value.
* Allow some dead code in cranelift-codegen
Prevents having to fix all warnings for all possible feature
combinations, only the main ones which come up.
* Update some debug file paths
We don't need full debug information but rather line tables
(debuginfo=1) should suffice for backtraces if truly necessary. Note
that this doesn't actually work on stable Rust just yet due to it being
an unrelease feature of Cargo. With the Rust release next week though
this'll work on all of stable/beta/nightly.
This should save us about 3GB of target directory disk space and it may
also be a tiny speed boost. There's no real benefit to using incremental
builds on CI because we're not changing code anyway!
This will delete a same-name of a previous release for all tags, not
just the dev tag. That way if we need to retry a tagged release we'll
delete it and recreate it as usual.
Closes#978
Rust's recent update to libstd of the wasm32-wasi target turned out to
be buggy with respect to fetching the process arguments, so we'll need
to wait on a fix there before we can run these tests with nightly again.
* use setuptools_scm for python version management
* add git dep for python wheel building
* if no tag is defined, default to dev
* any untagged version default to 0.0.1
This commit migrates wasmtime's CI infrastructure from Azure Pipelines
to Github Actions. Using Github Actions has a few benefits over other
offerings:
* Being natively integrated with Github means that there's no degree of
user account configuration or access control management, it's all
inherent via already existing Github permissions.
* Github Actions gives 20 parallel builders instead of Azure's 10 by
default, which is a nice boost to have!
Overall I've found Github Actions to feel a bit cleaner than Azure
Pipelines as well. Subjectively I've found the configuration to be more
readable and more pleasant to work with, although they're both just as
"powerful" I think. Additionally Github Actions has been pretty solid in
my own personal testing for a number of other projects.
The main trickiness with wasmtime's CI is the rolling `dev` release of
the master branch as well as binary releases for tags. Github Actions
doesn't have quite as much built in functionality as Azure Pipelines,
but Github Actions does have a nice feature where you can define the
code for an action locally rather than only using built-in actions.
This migration adds three local actions with some associated JS code to
run the action (currently it looks like it basically requires JS)
* An `install-rust` action papers over the gotchas about installing
Rust, allowing Rust installation to be a one-liner in the configuration.
* A `binary-compatible-builds` action allows easily configuring the
wheels and the binaries to be "more binary compatible" and handles
things like compilation flags on OSX and Windows while handling the
`centos:6` container on Linux.
* The `github-release` action is the logic using the `@actions/github`
JS package to orchestrate the custom way we manage rolling releases,
ensuring that a new release is made for the master branch under `dev`
(deleting the previous tag/release ahead of time) and then also
manages tagged releases by uploading them there.
I'm hoping that most of the inline actions here will largely go away.
For example `install-rust` should be simply `rustup update $toolchain`
once various environment issues are fixed on Github Actions runner
images. Additionally `github-release` will ideally migrate to something
like https://github.com/actions/create-release or similar once it has
enough functionality. I'm also hoping that the maintenance in the
meantime of these actions is pretty low-cost, but if it becomes an issue
we can look into other solutions!
