chore(ci): add dedicatd param-prod run

This commit is contained in:
Thomas Montaigu
2025-12-10 15:41:13 +01:00
parent aaec3fd109
commit 40e96d4c84
4 changed files with 337 additions and 20 deletions

View File

@@ -0,0 +1,280 @@
# Run a small subset of tests to ensure quick feedback.
name: aws_tfhe_param_prod_tests
env:
CARGO_TERM_COLOR: always
ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
RUSTFLAGS: "-C target-cpu=native"
RUST_BACKTRACE: "full"
RUST_MIN_STACK: "8388608"
SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
SLACKIFY_MARKDOWN: true
IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
PULL_REQUEST_MD_LINK: ""
CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN || secrets.GITHUB_TOKEN }}
# Secrets will be available only to zama-ai organization members
SECRETS_AVAILABLE: ${{ secrets.JOB_SECRET != '' }}
EXTERNAL_CONTRIBUTION_RUNNER: "large_ubuntu_64-22.04"
on:
# Allows you to run this workflow manually from the Actions tab as an alternative.
workflow_dispatch:
pull_request:
permissions:
contents: read
# zizmor: ignore[concurrency-limits] concurrency is managed after instance setup to ensure safe provisioning
jobs:
should-run:
name: aws_tfhe_param_prod_tests/should-run
runs-on: ubuntu-latest
permissions:
pull-requests: read # Needed to check for file change
outputs:
csprng_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.csprng_any_changed }}
zk_pok_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.zk_pok_any_changed }}
versionable_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.versionable_any_changed }}
core_crypto_test: ${{ env.IS_PULL_REQUEST == 'false' ||
steps.changed-files.outputs.core_crypto_any_changed ||
steps.changed-files.outputs.dependencies_any_changed }}
boolean_test: ${{ env.IS_PULL_REQUEST == 'false' ||
steps.changed-files.outputs.boolean_any_changed ||
steps.changed-files.outputs.dependencies_any_changed }}
shortint_test: ${{ env.IS_PULL_REQUEST == 'false' ||
steps.changed-files.outputs.shortint_any_changed ||
steps.changed-files.outputs.dependencies_any_changed }}
integer_test: ${{ env.IS_PULL_REQUEST == 'false' ||
steps.changed-files.outputs.integer_any_changed ||
steps.changed-files.outputs.dependencies_any_changed }}
wasm_test: ${{ env.IS_PULL_REQUEST == 'false' ||
steps.changed-files.outputs.wasm_any_changed ||
steps.changed-files.outputs.dependencies_any_changed }}
high_level_api_test: ${{ env.IS_PULL_REQUEST == 'false' ||
steps.changed-files.outputs.high_level_api_any_changed ||
steps.changed-files.outputs.dependencies_any_changed }}
user_docs_test: ${{ env.IS_PULL_REQUEST == 'false' ||
steps.changed-files.outputs.user_docs_any_changed ||
steps.changed-files.outputs.dependencies_any_changed }}
any_file_changed: ${{ env.IS_PULL_REQUEST == 'false' || steps.aggregated-changes.outputs.any_changed }}
steps:
- name: Checkout tfhe-rs
uses: actions/checkout@1af3b93b6815bc44a9784bd300feb67ff0d1eeb3
with:
fetch-depth: 0
persist-credentials: 'false'
token: ${{ env.CHECKOUT_TOKEN }}
- name: Check for file changes
id: changed-files
uses: tj-actions/changed-files@24d32ffd492484c1d75e0c0b894501ddb9d30d62 # v47.0.0
with:
files_yaml: |
dependencies:
- tfhe/Cargo.toml
- tfhe-csprng/**
- tfhe-fft/**
- tfhe-zk-pok/**
- utils/tfhe-versionable/**
- utils/tfhe-versionable-derive/**
csprng:
- tfhe-csprng/**
zk_pok:
- tfhe-zk-pok/**
versionable:
- utils/tfhe-versionable/**
- utils/tfhe-versionable-derive/**
core_crypto:
- tfhe/src/core_crypto/**
boolean:
- tfhe/src/core_crypto/**
- tfhe/src/boolean/**
shortint:
- tfhe/src/core_crypto/**
- tfhe/src/shortint/**
integer:
- tfhe/src/core_crypto/**
- tfhe/src/shortint/**
- tfhe/src/integer/**
wasm:
- tfhe/src/**
- tfhe/js_on_wasm_tests/**
- tfhe/web_wasm_parallel_tests/**
- '!tfhe/src/c_api/**'
- '!tfhe/src/boolean/**'
high_level_api:
- tfhe/src/**
- '!tfhe/src/c_api/**'
- '!tfhe/src/boolean/**'
- '!tfhe/src/c_api/**'
- '!tfhe/src/js_on_wasm_api/**'
user_docs:
- tfhe/src/**
- '!tfhe/src/c_api/**'
- 'tfhe/docs/**/**.md'
- README.md
- name: Aggregate file changes
id: aggregated-changes
if: ( steps.changed-files.outputs.dependencies_any_changed == 'true' ||
steps.changed-files.outputs.csprng_any_changed == 'true' ||
steps.changed-files.outputs.zk_pok_any_changed == 'true' ||
steps.changed-files.outputs.versionable_any_changed == 'true' ||
steps.changed-files.outputs.core_crypto_any_changed == 'true' ||
steps.changed-files.outputs.boolean_any_changed == 'true' ||
steps.changed-files.outputs.shortint_any_changed == 'true' ||
steps.changed-files.outputs.integer_any_changed == 'true' ||
steps.changed-files.outputs.wasm_any_changed == 'true' ||
steps.changed-files.outputs.high_level_api_any_changed == 'true' ||
steps.changed-files.outputs.user_docs_any_changed == 'true')
run: |
echo "any_changed=true" >> "$GITHUB_OUTPUT"
setup-instance:
name: aws_tfhe_param_prod_tests/setup-instance
if: github.event_name == 'workflow_dispatch' ||
(github.event_name != 'workflow_dispatch' && needs.should-run.outputs.any_file_changed == 'true')
needs: should-run
runs-on: ubuntu-latest
outputs:
runner-name: ${{ steps.start-remote-instance.outputs.label || steps.start-github-instance.outputs.runner_group }}
steps:
- name: Start remote instance
id: start-remote-instance
if: env.SECRETS_AVAILABLE == 'true'
uses: zama-ai/slab-github-runner@973c1d22702de8d0acd2b34e83404c96ed92c264 # v1.4.2
with:
mode: start
github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
slab-url: ${{ secrets.SLAB_BASE_URL }}
job-secret: ${{ secrets.JOB_SECRET }}
backend: aws
profile: cpu-big
# This instance will be spawned especially for pull-request from forked repository
- name: Start GitHub instance
id: start-github-instance
if: env.SECRETS_AVAILABLE == 'false'
run: |
echo "runner_group=${EXTERNAL_CONTRIBUTION_RUNNER}" >> "$GITHUB_OUTPUT"
param-prod-tests:
name: Fast CPU tests
needs: [ should-run, setup-instance ]
concurrency:
group: ${{ github.workflow_ref }}
cancel-in-progress: true
runs-on: ${{ needs.setup-instance.outputs.runner-name }}
steps:
- name: Checkout tfhe-rs
uses: actions/checkout@1af3b93b6815bc44a9784bd300feb67ff0d1eeb3
with:
persist-credentials: 'false'
token: ${{ env.CHECKOUT_TOKEN }}
- name: Install latest stable
uses: dtolnay/rust-toolchain@e97e2d8cc328f1b50210efc529dca0028893a2d9 # zizmor: ignore[stale-action-refs] this action doesn't create releases
with:
toolchain: stable
- name: Run user docs tests
if: needs.should-run.outputs.user_docs_test == 'true'
run: |
make test_user_doc
- name: Get Node version
run: |
echo "NODE_VERSION=$(make node_version)" >> "${GITHUB_ENV}"
- name: Node cache restoration
id: node-cache
uses: actions/cache/restore@0057852bfaa89a56745cba8c7296529d2fc39830 #v4.3.0
with:
path: |
~/.nvm
~/.npm
key: node-${{ env.NODE_VERSION }}
- name: Install Node
if: steps.node-cache.outputs.cache-hit != 'true'
run: |
make install_node
- name: Node cache save
uses: actions/cache/save@0057852bfaa89a56745cba8c7296529d2fc39830 #v4.3.0
if: steps.node-cache.outputs.cache-hit != 'true'
with:
path: |
~/.nvm
~/.npm
key: node-${{ env.NODE_VERSION }}
- name: Run js on wasm API tests
if: needs.should-run.outputs.wasm_test == 'true'
run: |
make test_nodejs_wasm_api_ci
- name: Gen Keys if required
if: needs.should-run.outputs.shortint_test == 'true' ||
needs.should-run.outputs.integer_test == 'true'
run: |
make gen_key_cache
- name: Run shortint tests
if: needs.should-run.outputs.shortint_test == 'true'
run: |
BIG_TESTS_INSTANCE=TRUE FAST_TESTS=FALSE make test_shortint_ci
- name: Run integer tests
if: needs.should-run.outputs.integer_test == 'true'
run: |
BIG_TESTS_INSTANCE=TRUE FAST_TESTS=FALSE make test_param_prod_integer_ci
- name: Run high-level API tests
if: needs.should-run.outputs.high_level_api_test == 'true'
run: |
make test_high_level_api
- name: Set pull-request URL
if: ${{ failure() && github.event_name == 'pull_request' }}
run: |
echo "PULL_REQUEST_MD_LINK=[pull-request](${PR_BASE_URL}${PR_NUMBER}), " >> "${GITHUB_ENV}"
env:
PR_BASE_URL: ${{ vars.PR_BASE_URL }}
PR_NUMBER: ${{ github.event.pull_request.number }}
- name: Slack Notification
if: ${{ failure() && env.SECRETS_AVAILABLE == 'true' }}
continue-on-error: true
uses: rtCamp/action-slack-notify@e31e87e03dd19038e411e38ae27cbad084a90661
env:
SLACK_COLOR: ${{ job.status }}
SLACK_MESSAGE: "Fast AWS tests finished with status: ${{ job.status }}. (${{ env.PULL_REQUEST_MD_LINK }}[action run](${{ env.ACTION_RUN_URL }}))"
teardown-instance:
name: aws_tfhe_param_prod_tests/teardown-instance
if: ${{ always() && needs.setup-instance.result == 'success' }}
needs: [ setup-instance, param-prod-tests ]
runs-on: ubuntu-latest
steps:
- name: Stop remote instance
id: stop-instance
if: env.SECRETS_AVAILABLE == 'true'
uses: zama-ai/slab-github-runner@973c1d22702de8d0acd2b34e83404c96ed92c264 # v1.4.2
with:
mode: stop
github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
slab-url: ${{ secrets.SLAB_BASE_URL }}
job-secret: ${{ secrets.JOB_SECRET }}
label: ${{ needs.setup-instance.outputs.runner-name }}
- name: Slack Notification
if: ${{ failure() || (cancelled() && github.event_name != 'pull_request') }}
uses: rtCamp/action-slack-notify@e31e87e03dd19038e411e38ae27cbad084a90661
env:
SLACK_COLOR: ${{ job.status }}
SLACK_MESSAGE: "Instance teardown (param-prod-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

View File

@@ -934,6 +934,15 @@ test_integer_ci: install_cargo_nextest
--cargo-profile "$(CARGO_PROFILE)" --avx512-support "$(AVX512_SUPPORT)" \
--tfhe-package "tfhe"
.PHONY: test_param_prod_integer_ci # Run the tests for integer ci
test_param_prod_integer_ci: install_cargo_nextest
BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
FAST_TESTS="$(FAST_TESTS)" \
NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
./scripts/integer-tests.sh \
--cargo-profile "$(CARGO_PROFILE)" --avx512-support "$(AVX512_SUPPORT)" \
--run-prod-only --tfhe-package "tfhe"
.PHONY: test_unsigned_integer_ci # Run the tests for unsigned integer ci
test_unsigned_integer_ci: install_cargo_nextest
BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \

View File

@@ -6,10 +6,12 @@ function usage() {
echo "$0: integer test runner"
echo
echo "--help Print this message"
echo "--no-run Does not run the tests, but prints which tests would be ran (except doctests)"
echo "--rust-toolchain The toolchain to run the tests with default: stable"
echo "--multi-bit Run multi-bit tests only: default off"
echo "--unsigned-only Run only unsigned integer tests, by default both signed and unsigned tests are run"
echo "--signed-only Run only signed integer tests, by default both signed and unsigned tests are run"
echo "--run-prod-only Run only the tests using the prod parameters"
echo "--nightly-tests Run integer tests configured for nightly runs (3_3 params)"
echo "--fast-tests Run integer set but skip a subset of longer tests"
echo "--long-tests Run only long run integer tests"
@@ -33,6 +35,8 @@ backend="cpu"
gpu_feature=""
avx512_feature=""
tfhe_package="tfhe"
prod_param_argument=
no_run=false
while [ -n "$1" ]
do
@@ -42,6 +46,14 @@ do
exit 0
;;
"--no-run" )
no_run=true
;;
"--run-prod-only" )
prod_param_argument="--run-prod-only"
;;
"--rust-toolchain" )
shift
RUST_TOOLCHAIN="$1"
@@ -150,7 +162,7 @@ if [[ "${backend}" == "gpu" ]]; then
fi
fi
filter_expression=$(/usr/bin/python3 scripts/test_filtering.py --layer integer --backend "${backend}" ${fast_tests_argument:+$fast_tests_argument} ${long_tests_argument:+$long_tests_argument} ${nightly_tests_argument:+$nightly_tests_argument} ${no_big_params_argument_gpu:+$no_big_params_argument_gpu} ${multi_bit_argument:+$multi_bit_argument} ${sign_argument:+$sign_argument} ${no_big_params_argument:+$no_big_params_argument})
filter_expression=$(/usr/bin/python3 scripts/test_filtering.py --layer integer --backend "${backend}" ${fast_tests_argument:+$fast_tests_argument} ${long_tests_argument:+$long_tests_argument} ${nightly_tests_argument:+$nightly_tests_argument} ${no_big_params_argument_gpu:+$no_big_params_argument_gpu} ${multi_bit_argument:+$multi_bit_argument} ${sign_argument:+$sign_argument} ${no_big_params_argument:+$no_big_params_argument} ${prod_param_argument:+$prod_param_argument})
if [[ "${FAST_TESTS}" == "TRUE" ]]; then
echo "Running 'fast' test set"
@@ -168,24 +180,39 @@ fi
echo "${filter_expression}"
cargo ${RUST_TOOLCHAIN:+"$RUST_TOOLCHAIN"} nextest run \
--tests \
--cargo-profile "${cargo_profile}" \
--package "${tfhe_package}" \
--profile ci \
--no-default-features \
--features=integer,internal-keycache,zk-pok,experimental,"${avx512_feature}","${gpu_feature}" \
--test-threads "${test_threads}" \
-E "$filter_expression"
if $no_run then; then
# This is very close to the run command, but the `--profile ci`
# is not 'supported'
cargo ${RUST_TOOLCHAIN:+"$RUST_TOOLCHAIN"} nextest list \
--tests \
--cargo-profile "${cargo_profile}" \
--package "${tfhe_package}" \
--no-default-features \
--features=integer,internal-keycache,zk-pok,experimental,"${avx512_feature}","${gpu_feature}" \
-E "$filter_expression"
else
cargo ${RUST_TOOLCHAIN:+"$RUST_TOOLCHAIN"} nextest run \
--tests \
--cargo-profile "${cargo_profile}" \
--package "${tfhe_package}" \
--profile ci \
--no-default-features \
--features=integer,internal-keycache,zk-pok,experimental,"${avx512_feature}","${gpu_feature}" \
--test-threads "${test_threads}" \
-E "$filter_expression"
if [[ -z ${multi_bit_argument} && -z ${long_tests_argument} ]]; then
cargo ${RUST_TOOLCHAIN:+"$RUST_TOOLCHAIN"} test \
# Unfortunately, we cannot skip running doctest with `--no-run`
if [[ -z ${multi_bit_argument} && -z ${long_tests_argument} ]]; then
cargo ${RUST_TOOLCHAIN:+"$RUST_TOOLCHAIN"} test \
--profile "${cargo_profile}" \
--package "${tfhe_package}" \
--no-default-features \
--features=integer,internal-keycache,experimental,"${avx512_feature}","${gpu_feature}" \
--doc \
-- --test-threads="${doctest_threads}" integer::"${gpu_feature}"
fi
fi
echo "Test ran in $SECONDS seconds"

View File

@@ -121,18 +121,17 @@ def filter_integer_tests(input_args):
("_multi_bit", "_group_[0-9]") if input_args.multi_bit else ("", "")
)
backend_filter = ""
if input_args.run_prod_only:
filter_expression = [
"test(/^integer::.*_param_prod/)",
]
elif not input_args.long_tests:
if not input_args.long_tests:
if input_args.backend == "gpu":
backend_filter = "gpu::"
if multi_bit_filter:
# For now, GPU only has specific parameters set for multi-bit
multi_bit_filter = "_gpu_multi_bit"
filter_expression = [f"test(/^integer::{backend_filter}.*/)"]
if input_args.run_prod_only:
filter_expression = [f"test(/^integer::{backend_filter}.*_param_prod/)"]
else:
filter_expression = [f"test(/^integer::{backend_filter}.*/)"]
if input_args.multi_bit:
filter_expression.append("test(~_multi_bit)")
@@ -175,6 +174,9 @@ def filter_integer_tests(input_args):
for pattern in excluded_tests:
filter_expression.append(f"not test({pattern})")
if not input_args.run_prod_only:
filter_expression.append("not test(/.*_param_prod_.*/)")
else:
if input_args.backend == "gpu":
filter_expression = [
@@ -184,11 +186,10 @@ def filter_integer_tests(input_args):
filter_expression = [
"test(/^integer::server_key::radix_parallel::tests_long_run.*/)"
]
# Do not run noise check tests by default as they can be very slow
# they will be run e.g. nightly or on demand
filter_expression.append(
f"not test(/^integer::gpu::server_key::radix::tests_noise_distribution::.*::test_gpu_noise_check.*/)"
"not test(/^integer::gpu::server_key::radix::tests_noise_distribution::.*::test_gpu_noise_check.*/)"
)
return " and ".join(filter_expression)