Files
hermes-agent/.github/workflows/tests.yml
ethernet 0012dd1e0c perf(ci): set python test workers to one for each core, from measurement
`run_tests.sh` defaults to twice the core count, and the value this branch
started with came from a rule of thumb of 1.5x cores plus a measurement on a
16-core machine. A sweep on the real runner disagrees with both.

Run 32549672063 on the 96-core runner (EPYC 7763, 377GB) timed the whole suite
at six worker counts, two repetitions for each. A warmup run came first, and
retries were off:

    workers   x cores   rep 1   rep 2   mean
       48       0.5x     138s    139s   138s
       96       1.0x     127s    126s   126s   <- fastest
      144       1.5x     130s    134s   132s
      192       2.0x     132s    133s   132s
      240       2.5x     140s    139s   140s
      288       3.0x     143s    142s   142s

One worker for each core wins. Both repetitions agree on the order.

The shape is the more useful result. The range is 126s to 142s across a 6x
range of worker counts. The suite has sufficient concurrency at this machine
size, so nothing above the core count buys anything. The remaining time
belongs to the slowest individual files and to the setup. A future gain must
come from those, and not from this number.

The sweep ran from a temporary workflow that this branch does not keep.
2026-08-22 02:25:12 -04:00

197 lines
8.6 KiB
YAML

name: Tests
on:
workflow_call:
permissions:
contents: read
# Cancel in-progress runs for the same ref
concurrency:
group: tests-${{ github.ref }}
cancel-in-progress: true
jobs:
test:
name: Run tests
# One 96-core runner for the whole suite. There is no slicing. Slicing
# existed to spread the suite over 4-core runners. It cost a matrix job, a
# duration cache, a per-slice artifact and a merge job to do it.
#
# 96 cores clear the floor that the slowest single test file sets (about
# 82s). A second slice divides work that is already at that floor, and
# adds a second setup.
runs-on: ubuntu-latest-96-core
timeout-minutes: 30
steps:
- name: Checkout code
uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
- name: Install ripgrep (prebuilt binary)
run: |
set -euo pipefail
RG_VERSION=15.1.0
RG_SHA256=1c9297be4a084eea7ecaedf93eb03d058d6faae29bbc57ecdaf5063921491599
RG_TARBALL=ripgrep-${RG_VERSION}-x86_64-unknown-linux-musl.tar.gz
curl -sSfL --retry 3 --retry-delay 5 -o "$RG_TARBALL" \
"https://github.com/BurntSushi/ripgrep/releases/download/${RG_VERSION}/${RG_TARBALL}"
echo "${RG_SHA256} ${RG_TARBALL}" | sha256sum -c -
tar -xzf "$RG_TARBALL"
sudo mv "ripgrep-${RG_VERSION}-x86_64-unknown-linux-musl/rg" /usr/local/bin/rg
rm -rf "$RG_TARBALL" "ripgrep-${RG_VERSION}-x86_64-unknown-linux-musl"
rg --version
- name: Install uv
uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # 8.2.0
with:
# Pin the uv version: unpinned, setup-uv resolves "latest" by
# fetching a manifest from raw.githubusercontent.com on EVERY job —
# a transient fetch failure fails the whole job (2026-07-28 slice-5
# incident). Pinned, the binary downloads directly; no manifest hop.
version: "0.9.28"
# Persist uv's download/wheel cache (~/.cache/uv) across runs.
# Keyed on the dependency manifests, so the cache is reused until
# pyproject.toml or uv.lock changes. `uv sync` still runs every
# time, but resolves from the warm cache instead of re-downloading
# and re-building wheels.
enable-cache: true
cache-dependency-glob: |
pyproject.toml
uv.lock
- name: Set up Python 3.11
uses: ./.github/actions/retry
with:
command: uv python install 3.11
- name: Install dependencies
# `uv sync --locked` installs the exact pinned set from uv.lock (and
# fails if the lock is out of sync with pyproject.toml), giving a
# reproducible env. It also creates .venv itself, so no separate
# `uv venv` step is needed.
#
# The trailing extras beyond all/dev are the lazy-install features
# (tools/lazy_deps.py) that tests exercise for real: provider.anthropic,
# stt/tts.mistral, image.fal, terminal.modal, terminal.daytona,
# memory.hindsight, search.parallel. The hermetic test env forbids
# mid-run pip installs (HERMES_DISABLE_LAZY_INSTALLS=1 in
# tests/conftest.py), so the SDKs those tests need must be in the
# venv up front — resolved from uv.lock like everything else, which
# also honors the exact supply-chain pins these extras carry.
uses: ./.github/actions/retry
with:
command: uv sync --locked --python 3.11 --extra all --extra dev --extra anthropic --extra mistral --extra fal --extra modal --extra daytona --extra hindsight --extra parallel-web
- name: Minimize uv cache
# Optimized for CI: prunes pre-built wheels that are cheap to
# re-download, keeping the persisted cache small and fast to restore.
run: uv cache prune --ci
- name: Run tests
# Per-file isolation via scripts/run_tests.sh: each test file runs
# in its own freshly-spawned `python -m pytest <file>` subprocess
# with bounded parallelism. No xdist, no shared workers, no
# module-level state leakage between files.
#
# No --files: the runner discovers the suite itself. The discovered
# set is identical to the list the removed matrix job used to pass in.
run: |
source .venv/bin/activate
scripts/run_tests.sh
env:
# This is the maximum number of test FILES that run together.
# run_tests_parallel.py starts one pytest subprocess for each file
# from a single ThreadPoolExecutor, so this value IS the limit. The
# default is cpu_count*2, which is 192 here.
#
# Measured on this runner (96-core EPYC 7763, 377GB). Whole suite,
# two repetitions for each value. See run 32549672063:
#
# workers x cores mean
# 48 0.5x 138s
# 96 1.0x 126s <- fastest
# 144 1.5x 132s
# 192 2.0x 132s
# 240 2.5x 140s
# 288 3.0x 142s
#
# One worker for each core wins. The curve is shallow: 126s to 142s
# across a 6x range. The suite has sufficient concurrency at this
# size. The remaining time is the slowest files plus the setup.
# Workers above the core count only add contention.
HERMES_TEST_WORKERS: 96
# Ensure tests don't accidentally call real APIs
OPENROUTER_API_KEY: ""
OPENAI_API_KEY: ""
NOUS_API_KEY: ""
e2e:
runs-on: ubuntu-latest
timeout-minutes: 15
steps:
- name: Checkout code
uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
- name: Install ripgrep (prebuilt binary)
run: |
set -euo pipefail
RG_VERSION=15.1.0
RG_SHA256=1c9297be4a084eea7ecaedf93eb03d058d6faae29bbc57ecdaf5063921491599
RG_TARBALL=ripgrep-${RG_VERSION}-x86_64-unknown-linux-musl.tar.gz
curl -sSfL --retry 3 --retry-delay 5 -o "$RG_TARBALL" \
"https://github.com/BurntSushi/ripgrep/releases/download/${RG_VERSION}/${RG_TARBALL}"
echo "${RG_SHA256} ${RG_TARBALL}" | sha256sum -c -
tar -xzf "$RG_TARBALL"
sudo mv "ripgrep-${RG_VERSION}-x86_64-unknown-linux-musl/rg" /usr/local/bin/rg
rm -rf "$RG_TARBALL" "ripgrep-${RG_VERSION}-x86_64-unknown-linux-musl"
rg --version
- name: Install uv
uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # 8.2.0
with:
# Pin the uv version: unpinned, setup-uv resolves "latest" by
# fetching a manifest from raw.githubusercontent.com on EVERY job —
# a transient fetch failure fails the whole job (2026-07-28 slice-5
# incident). Pinned, the binary downloads directly; no manifest hop.
version: "0.9.28"
# Persist uv's download/wheel cache (~/.cache/uv) across runs.
# Keyed on the dependency manifests, so the cache is reused until
# pyproject.toml or uv.lock changes. `uv sync` still runs every
# time, but resolves from the warm cache instead of re-downloading
# and re-building wheels.
enable-cache: true
cache-dependency-glob: |
pyproject.toml
uv.lock
- name: Set up Python 3.11
run: uv python install 3.11
- name: Install dependencies
# `uv sync --locked` installs the exact pinned set from uv.lock (and
# fails if the lock is out of sync with pyproject.toml), giving a
# reproducible env. It also creates .venv itself, so no separate
# `uv venv` step is needed.
#
# Same extras as the test job's sync above: the hermetic test env
# forbids mid-run pip installs (HERMES_DISABLE_LAZY_INSTALLS=1 in
# tests/conftest.py), so lazy-install SDKs exercised by tests must be
# in the venv up front.
uses: ./.github/actions/retry
with:
command: uv sync --locked --python 3.11 --extra all --extra dev --extra anthropic --extra mistral --extra fal --extra modal --extra daytona --extra hindsight --extra parallel-web
- name: Minimize uv cache
# Optimized for CI: prunes pre-built wheels that are cheap to
# re-download, keeping the persisted cache small and fast to restore.
run: uv cache prune --ci
- name: Run e2e tests
run: |
source .venv/bin/activate
python -m pytest tests/e2e/ -v --tb=short
env:
OPENROUTER_API_KEY: ""
OPENAI_API_KEY: ""
NOUS_API_KEY: ""