| commit | e3cb102c03ef01588cd4913d537f70a73928c34c | [log] [tgz] |
|---|---|---|
| author | Geoffrey Martin-Noble <gcmn@google.com> | Thu Jan 05 10:49:26 2023 -0800 |
| committer | GitHub <noreply@github.com> | Thu Jan 05 10:49:26 2023 -0800 |
| tree | ffb0091ec1c407ca775480bee8fd4be0452d2a35 | |
| parent | a13b60a3ad2257a901068e08ba95014d3caca06f [diff] |
Enable custom scale-down of runner instances (#11688) Currently, instances only delete themselves when they finish their job. This doesn't help if the group has scaled up too much and then a bunch of instances sit idle. This introduces a periodic check if the runner is idle and the autoscaler thinks the group should scale down and stops the runner if so. There's a few refactors in here. Commits should mostly be reviewable individually (I think some earlier commits may have some syntax errors, so perhaps overlook those till the end). Of note, this adds cloud logging for systemd services from these VMs, which is helpful when debugging. Tested: Deployed to test environment, including a test instance-deleter service. Deliberately over-scaled instance group and confirmed that it scaled back down within ~30 minutes  Ran jobs against test environment and confirmed no issues: https://github.com/iree-org/iree/actions/runs/3810920085/jobs/6488260971 and that group scaled up and then back down again:  Note that when the presubmit run had finished, the group was still at seven instances, so it was this new functionality that caused the scale down.
IREE (Intermediate Representation Execution Environment, pronounced as “eerie”) is an MLIR-based end-to-end compiler and runtime that lowers Machine Learning (ML) models to a unified IR that scales up to meet the needs of the datacenter and down to satisfy the constraints and special considerations of mobile and edge deployments.
See our website for project details, user guides, and instructions on building from source.
IREE is still in its early phase. We have settled down on the overarching infrastructure and are actively improving various software components as well as project logistics. It is still quite far from ready for everyday use and is made available without any support at the moment. With that said, we welcome any kind of feedback on any communication channels!
See our website for more information.
IREE is licensed under the terms of the Apache 2.0 License with LLVM Exceptions. See LICENSE for more information.
