Every path-gated workflow begins with one small job -- `does this change touch
it`, or `which workspaces this change touches` for the standards build -- that
all of its real jobs declare in `needs`. All 55 of those gate jobs ran on
`linux-arm64`.
That is the one label the fleet cannot spare. Sampled while 267 runs were
queued, every waiting run had the same shape:
[waiting] linux-arm64 does this change touch it
[blocked] linux member · dataflow device, simulated
[blocked] macos member · dataflow device absent
237 runs, each holding a gate in the longest queue, and every real job blocked
behind it. Not only `linux` jobs: `macos`, `linux-hw`, `linux-arm64-hw` and
`macos-m3-ultra` jobs were blocked too. The measured effect was `linux-arm64`
263 deep against `linux` **zero**, with jetson-hub's 20 cores taking no job for
a full hour -- starved not for want of work but because the work could not be
unblocked. A gate waiting on the resource it is rationing is head-of-line
blocking, and it makes every other runner in the fleet idle on its behalf.
The gates move to `linux`. Nothing else changes: the diff is 55 `runs-on` lines
and no others, and every touched file still parses.
Why `linux` and not elsewhere. The gate is not free -- it compares
`BASE...HEAD`, which needs the merge base, so `fetch-depth: 0` is a full clone
rather than a ten-second script. `macos` is served by hosts on a Mac Studio
sitting at load 185; `linux` is a separate 20-core box that was measurably
empty. This moves a serialisation point to the least contended label; it does
not remove it, and the gate is still a single point of failure for its
workflow. The structural fix is an early-exit first step inside each job, which
costs one checkout per leg and is a larger change.
Follow-up, deliberately not bundled: jetson-hub runs at capacity 2 with 15 GB of
RAM, so 55 clones still serialise two at a time. Raising capacity helps gates
and hurts the heavy emulated-architecture jobs that share the label, so it wants
measuring on its own.
Co-Authored-By: Claude Opus 5 <[email protected]>
137 lines
7.4 KiB
YAML
137 lines
7.4 KiB
YAML
name: wai-quantum variational and algorithm library
|
|
|
|
# The variational and algorithm library (wai/wai-quantum): a job's params
|
|
# bound exactly (engine::catalog::bind_params, a sweep fixing some of them),
|
|
# every gate's derivative by each parameter carried by the reader
|
|
# (qasm::tangents, refused with Q129 where it cannot follow one), exact
|
|
# gradients by adjoint differentiation and the parameter-shift rule,
|
|
# deterministic optimisers whose every step is a link of a digest chain, the
|
|
# variational eigensolver beside exact diagonalisation, QAOA for Max-Cut
|
|
# beside brute force (engine::variational), and the algorithm builders, each
|
|
# program checked against a reference built from its definition by the
|
|
# equivalence ladder's matrix and truth-table rungs (engine::algorithms,
|
|
# engine::equiv).
|
|
#
|
|
# Ground truth first: every gradient is held to one worked out by hand within
|
|
# its stated rounding bound, to central finite differences and to the other
|
|
# method; the spectrum to closed forms and to the trace and squared-norm
|
|
# invariants; QAOA at p = 1 to its closed form on an edge and a ring; phase
|
|
# and amplitude estimation to the Fejér distribution; the adder to a truth
|
|
# table written apart; QSVT to the Chebyshev polynomials of a signal and of a
|
|
# matrix. The answers' bytes are pinned, and the pins run on x86_64 and arm64
|
|
# here (and on every architecture wai-energy-receipts.yml's matrix runs); the
|
|
# browser engine holds the same bytes in wasm32 (wai-web-wasm.yml). The
|
|
# surfaces' tests run in wai-quantum-cli.yml, wai-quantum-service.yml and the
|
|
# contract (wai-quantum-contract.yml).
|
|
|
|
concurrency:
|
|
group: ${{ github.workflow }}-${{ github.ref }}
|
|
cancel-in-progress: true
|
|
|
|
on:
|
|
push:
|
|
branches: [main]
|
|
paths:
|
|
- 'wai/wai-quantum/**'
|
|
- 'rust-toolchain.toml'
|
|
- '.github/workflows/wai-quantum-variational.yml'
|
|
pull_request:
|
|
paths:
|
|
- 'wai/wai-quantum/**'
|
|
- 'rust-toolchain.toml'
|
|
- '.github/workflows/wai-quantum-variational.yml'
|
|
|
|
jobs:
|
|
relevant:
|
|
name: does this change touch it
|
|
runs-on: linux
|
|
outputs:
|
|
run: ${{ steps.gate.outputs.run }}
|
|
steps:
|
|
- uses: actions/checkout@v5
|
|
with:
|
|
fetch-depth: 0
|
|
- id: gate
|
|
run: |
|
|
set -euo pipefail
|
|
r=$(python3 scripts/pr-touches-paths.py .github/workflows/wai-quantum-variational.yml \
|
|
"${{ github.event.pull_request.base.sha }}" \
|
|
"${{ github.event.pull_request.head.sha }}")
|
|
echo "this change touches the variational and algorithm library: $r"
|
|
echo "run=$r" >> "$GITHUB_OUTPUT"
|
|
|
|
native:
|
|
name: parameters, gradients, optimisers, VQE, QAOA and builders · ${{ matrix.os }}, release
|
|
needs: relevant
|
|
if: needs.relevant.outputs.run == 'true'
|
|
strategy:
|
|
fail-fast: false
|
|
matrix:
|
|
os: [linux, linux-arm64]
|
|
runs-on: ${{ matrix.os }}
|
|
timeout-minutes: 60
|
|
steps:
|
|
- uses: actions/checkout@v5
|
|
- name: Install Rust 1.98.0
|
|
shell: bash
|
|
run: |
|
|
export PATH="$HOME/.cargo/bin:$PATH"
|
|
command -v rustup >/dev/null 2>&1 || \
|
|
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --profile minimal --default-toolchain 1.98.0
|
|
rustup toolchain install 1.98.0 --profile minimal
|
|
rustup default 1.98.0
|
|
echo "$HOME/.cargo/bin" >> "$GITHUB_PATH"
|
|
# A module filter is satisfied by any one test of the module, so each
|
|
# new test is required by name to have run and passed.
|
|
- name: The library's tests, each required by name
|
|
shell: bash
|
|
working-directory: wai/wai-quantum
|
|
run: |
|
|
set -euo pipefail
|
|
cargo test --locked --release --features full --lib -- \
|
|
engine::variational:: engine::algorithms:: engine::equiv:: \
|
|
qasm::tests::a_gate_s_derivative_by_each_parameter_is_carried_to_it \
|
|
qasm::tests::a_parameter_the_derivative_cannot_follow_is_refused_on_its_line \
|
|
engine::catalog::tests::a_job_binds_its_parameters_exactly \
|
|
engine::sets::tests::a_sweep_binds_the_parameters_params_fix \
|
|
engine::schema::tests::the_variational_and_algorithm_answers_follow_their_schema \
|
|
engine::refusal:: qasm::diag:: 2>&1 | tee "$RUNNER_TEMP/variational.log"
|
|
grep -qE '^test result: ok\. [1-9][0-9]+ passed; 0 failed' "$RUNNER_TEMP/variational.log" || { echo "::error::the library's tests did not pass"; exit 1; }
|
|
for t in engine::variational::tests::each_method_gives_the_gradient_worked_out_by_hand \
|
|
engine::variational::tests::both_methods_agree_with_finite_differences_and_with_each_other \
|
|
engine::variational::tests::a_gradient_is_the_same_bytes_every_time \
|
|
engine::variational::tests::what_a_gradient_cannot_take_is_refused_with_its_code \
|
|
engine::variational::tests::the_ansatz_is_the_variational_ml_module_s_and_so_is_its_gradient \
|
|
engine::variational::tests::the_spectrum_is_the_closed_form_and_keeps_the_invariants \
|
|
engine::variational::tests::each_optimiser_reaches_the_minimum_and_is_deterministic \
|
|
engine::variational::tests::an_optimisation_s_receipt_and_chain_bind_every_step \
|
|
engine::variational::tests::vqe_finds_the_ground_energy_diagonalisation_gives \
|
|
engine::variational::tests::qaoa_is_max_cut_and_finds_an_optimal_cut \
|
|
engine::variational::tests::the_answers_are_the_same_bytes_on_every_platform \
|
|
engine::algorithms::tests::phase_estimation_reads_the_phase_with_the_fejer_distribution \
|
|
engine::algorithms::tests::amplitude_estimation_reads_the_amplitude \
|
|
engine::algorithms::tests::the_adder_is_its_truth_table \
|
|
engine::algorithms::tests::signal_processing_gives_its_polynomial \
|
|
engine::algorithms::tests::the_singular_value_transformation_applies_the_polynomial_to_the_matrix \
|
|
engine::algorithms::tests::a_builder_s_refusals_name_their_code \
|
|
qasm::tests::a_gate_s_derivative_by_each_parameter_is_carried_to_it \
|
|
qasm::tests::a_parameter_the_derivative_cannot_follow_is_refused_on_its_line \
|
|
engine::catalog::tests::a_job_binds_its_parameters_exactly \
|
|
engine::sets::tests::a_sweep_binds_the_parameters_params_fix \
|
|
engine::schema::tests::the_variational_and_algorithm_answers_follow_their_schema \
|
|
engine::refusal::tests::the_documented_table_is_the_table \
|
|
qasm::diag::tests::the_documented_table_is_the_table; do
|
|
grep -q "^test $t \.\.\. ok$" "$RUNNER_TEMP/variational.log" || { echo "::error::$t did not run and pass"; exit 1; }
|
|
done
|
|
# Without warnings: a feature gate that leaves code unreachable or a
|
|
# value unused in one feature set shows here.
|
|
- name: Each feature set the hosts build still builds, without warnings
|
|
shell: bash
|
|
working-directory: wai/wai-quantum
|
|
env:
|
|
RUSTFLAGS: -D warnings
|
|
run: |
|
|
set -euo pipefail
|
|
cargo check --locked --no-default-features --features quantum_engine --lib
|
|
cargo check --locked --no-default-features --features quantum_engine,quantum_receipt --lib
|
|
cargo check --locked --no-default-features --features quantum_engine,quantum_receipt,quantum_qec,quantum_noise --lib
|