open-standards/.github/workflows/wai-quantum-opencl.yml
Transaction Science 6517401f01 CI: the gate that decides whether a workflow runs was queued behind the work it gates
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]>
2026-11-03 15:19:13 -05:00

309 lines
18 KiB
YAML

name: wai-quantum OpenCL member
# The fabric's OpenCL member: the gate kernel in OpenCL C, built at run time by
# each device's own compiler and certified bit-exact against the CPU
# reference's pinned corpus digests.
#
# - opencl-hw, on each host with OpenCL devices (jetson-hub: the vendors'
# runtimes for its CPU, its discrete GPU and its integrated GPU; thor: PoCL
# for its CPU and, through CUDA, its GPU): the devices its matrix entry
# declares, each by a key that carries the device's vendor id
# (opencl-<kind>-<vendor id, hex>). The device
# tests must find exactly the declared devices, agree with the CPU on the
# arithmetic probe and on random states (out-of-range runs included),
# reproduce all four pinned certificates on each device (two gatesets,
# random-v1 and fabric-v1), and catch every kernel mutant on each; and the
# dense f64 state vector, run resident (the state on the device for the
# whole run), must come back with the CPU's bits on each, on its pinned
# corpus and on random circuits, its wider kernels certified on their own
# corpus or said not to be (a member whose wider kernels did not certify
# runs gates on up to three qubits at once).
# - opencl-amd (framework-128gb): the same on its integrated GPU, through the
# GPU vendor's OpenCL platform, the state vector included.
# - opencl-macos (the macOS runner): the same on the platform's GPU, through its
# OpenCL 1.2; that GPU has no binary64, so the state vector is refused there,
# with the reason.
# - opencl-absent (a container with no OpenCL): the same build finds no member
# and says why, rather than failing; the device tests run over no member with
# the declaration `none`.
concurrency:
group: ${{ github.workflow }}-${{ github.ref }}
cancel-in-progress: true
on:
pull_request:
paths:
- 'wai/wai-quantum/**'
- 'rust-toolchain.toml'
- '.github/workflows/wai-quantum-opencl.yml'
schedule:
- cron: '35 5 * * *'
jobs:
# Does this pull request's own diff touch the paths below? This forge matches a
# workflow's `paths:` against the push, so a merge of main into a branch fires the
# workflow whatever the pull request changes; every job here waits on this answer.
# A push to main, a schedule, or anything the script cannot determine answers
# `true` (scripts/pr-touches-paths.py).
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-opencl.yml \
"${{ github.event.pull_request.base.sha }}" \
"${{ github.event.pull_request.head.sha }}")
echo "this change touches wai-quantum-opencl: $r"
echo "run=$r" >> "$GITHUB_OUTPUT"
opencl-hw:
name: member · OpenCL GPU and CPU (${{ matrix.host }})
needs: relevant
if: needs.relevant.outputs.run == 'true'
runs-on: ${{ matrix.runner }}
timeout-minutes: 90
strategy:
fail-fast: false
matrix:
include:
# Each host's devices, sorted, as discovery names them. A key carries
# the device's vendor id (opencl-<kind>-<vendor id, hex>); a device
# with no binary64 has its state vector refused, and says so.
- host: jetson-hub
runner: linux-hw
opencl: opencl-cpu-8086,opencl-gpu-10de,opencl-gpu-8086
pocl_arch: ''
# PoCL's CUDA device compiles to PTX for sm_90 (its LLVM 18 has no
# sm_110), which the driver compiles for the GPU.
- host: thor
runner: linux-arm64-hw
opencl: opencl-cpu-13b5,opencl-gpu-10de
pocl_arch: sm_90
env:
WAI_QUANTUM_HW_OPENCL_KEYS: ${{ matrix.opencl }}
POCL_CUDA_GPU_ARCH: ${{ matrix.pocl_arch }}
CARGO_PROFILE_DEV_DEBUG: '0'
CARGO_BUILD_JOBS: '8'
steps:
- name: Fetch the source
shell: bash
run: |
set -euo pipefail
want="${{ github.event.pull_request.head.sha || github.sha }}"
rm -rf src && mkdir src && cd src
git init -q
# Fetch the commit itself, not the ref: a later commit on the ref that this
# workflow's paths do not match would otherwise be fetched instead.
git fetch -q --depth 1 http://100.64.224.0:3400/Transaction-Science/open-standards.git "$want"
git checkout -q FETCH_HEAD
got="$(git rev-parse HEAD)"
[ "$got" = "$want" ] || { echo "::error::fetched $got, not $want"; exit 1; }
- name: The member's unit tests
shell: bash
working-directory: src/wai/wai-quantum
run: |
set -euo pipefail
export PATH="$HOME/.cargo/bin:$PATH"
CARGO_TARGET_DIR="$HOME/hw-target/wai-quantum-opencl" nice -n 10 cargo test --locked --features quantum_fabric_opencl --lib -- quantum_fabric_opencl:: quantum_fabric::opencl:: 2>&1 | tee "$RUNNER_TEMP/unit.log"
grep -qE '^test result: ok\. [1-9][0-9]* passed; 0 failed' "$RUNNER_TEMP/unit.log" || { echo "::error::no unit test ran"; exit 1; }
- name: The member's device tests
shell: bash
working-directory: src/wai/wai-quantum
run: |
set -euo pipefail
export PATH="$HOME/.cargo/bin:$PATH"
# The CPU platform's runtime loads its threading and memory libraries
# from the toolkit's own directories; without them the ICD loader
# skips the platform silently.
oneapi=/opt/intel/oneapi
export LD_LIBRARY_PATH="$oneapi/compiler/latest/lib:$oneapi/tbb/latest/lib/intel64/gcc4.8:$oneapi/tbb/latest/lib:$oneapi/umf/latest/lib:$oneapi/tcm/latest/lib${LD_LIBRARY_PATH:+:$LD_LIBRARY_PATH}"
export WAI_QUANTUM_OPENCL_EXPECT="${WAI_QUANTUM_HW_OPENCL_KEYS:-opencl-cpu-8086,opencl-gpu-10de}"
echo "devices declared: $WAI_QUANTUM_OPENCL_EXPECT"
report="$RUNNER_TEMP/opencl-report.txt"; : > "$report"
WAI_QUANTUM_OPENCL_REPORT="$report" CARGO_TARGET_DIR="$HOME/hw-target/wai-quantum-opencl" nice -n 10 \
cargo test --locked --release --features quantum_fabric_opencl --test opencl_member -- --test-threads=1 2>&1 | tee "$RUNNER_TEMP/device.log"
grep -qE '^test result: ok\. 9 passed; 0 failed' "$RUNNER_TEMP/device.log" || { echo "::error::not every device test ran and passed"; exit 1; }
cat "$report"
grep -qx "opencl devices: as declared ($WAI_QUANTUM_OPENCL_EXPECT)" "$report" || { echo "::error::the devices found are not the declared ones"; exit 1; }
for key in ${WAI_QUANTUM_OPENCL_EXPECT//,/ }; do
n="$(grep -c "^opencl certified: key=$key .* parity=bit-exact " "$report")"
[ "$n" = 4 ] || { echo "::error::$key: $n of 4 certificates were earned"; exit 1; }
grep -qx "opencl mutants: key=$key killed 6 of 6" "$report" || { echo "::error::$key: a kernel mutant survived"; exit 1; }
done
- name: The dense f64 state vector, bit for bit
shell: bash
working-directory: src/wai/wai-quantum
run: |
set -euo pipefail
export PATH="$HOME/.cargo/bin:$PATH"
oneapi=/opt/intel/oneapi
export LD_LIBRARY_PATH="$oneapi/compiler/latest/lib:$oneapi/tbb/latest/lib/intel64/gcc4.8:$oneapi/tbb/latest/lib:$oneapi/umf/latest/lib:$oneapi/tcm/latest/lib${LD_LIBRARY_PATH:+:$LD_LIBRARY_PATH}"
export WAI_QUANTUM_OPENCL_EXPECT="${WAI_QUANTUM_HW_OPENCL_KEYS:-opencl-cpu-8086,opencl-gpu-10de}"
report="$RUNNER_TEMP/opencl-sv-report.txt"; : > "$report"
WAI_QUANTUM_OPENCL_REPORT="$report" CARGO_TARGET_DIR="$HOME/hw-target/wai-quantum-opencl" nice -n 10 \
cargo test --locked --release --features quantum_fabric_opencl,quantum_sv --test opencl_sv -- --test-threads=1 2>&1 | tee "$RUNNER_TEMP/sv.log"
grep -qE '^test result: ok\. 4 passed; 0 failed' "$RUNNER_TEMP/sv.log" || { echo "::error::not every state-vector test ran and passed"; exit 1; }
cat "$report"
for key in ${WAI_QUANTUM_OPENCL_EXPECT//,/ }; do
if grep -qx "opencl sv refused: key=$key the device has no binary64" "$report"; then continue; fi
grep -q "^opencl sv certified: key=$key .* parity=bit-exact cases=48 digest=b7d129d0593adf70dc164477664970c5c0b8c09f79f1a0573036bd60888dbcd1$" "$report" || { echo "::error::$key: the state vector is not the CPU's bit for bit"; exit 1; }
grep -qx "opencl sv mutants: key=$key killed 4 of 4" "$report" || { echo "::error::$key: a state-vector mutant survived"; exit 1; }
# The wider kernels certify on their own corpus, or the member says
# they did not (it then runs gates on up to three qubits at once).
grep -qE "^opencl sv wide certified: key=$key .* kernel=wai.quantum.sv/run-wide .* parity=(bit-exact cases=120 digest=093614f2386ec5120ab6ccde5187a544518432751591535a6f58c7a44061ebaf|not-bit-exact cases=120 digest=[0-9a-f]{64})$" "$report" || { echo "::error::$key: the wider kernels' certificate was not established"; exit 1; }
grep -qE "^opencl sv random: key=$key [0-9]+ resident runs on (every kernel|the narrow kernels) match the CPU bit for bit$" "$report" || { echo "::error::$key: a random circuit's resident run is not the CPU's"; exit 1; }
done
opencl-amd:
name: member · OpenCL GPU (framework-128gb)
needs: relevant
if: needs.relevant.outputs.run == 'true'
runs-on: linux-amd
timeout-minutes: 90
env:
# The devices are declared by the host, in its runner's environment
# (WAI_QUANTUM_AMD_OPENCL_KEYS); with no declaration, its integrated GPU.
CARGO_PROFILE_DEV_DEBUG: '0'
CARGO_BUILD_JOBS: '8'
steps:
- name: Fetch the source
shell: bash
run: |
set -euo pipefail
want="${{ github.event.pull_request.head.sha || github.sha }}"
rm -rf src && mkdir src && cd src
git init -q
# Fetch the commit itself, not the ref: a later commit on the ref that this
# workflow's paths do not match would otherwise be fetched instead. The
# host names its own route to the forge (WAI_FORGE_URL).
git fetch -q --depth 1 "${WAI_FORGE_URL:-http://100.64.224.0:3400}/Transaction-Science/open-standards.git" "$want"
git checkout -q FETCH_HEAD
got="$(git rev-parse HEAD)"
[ "$got" = "$want" ] || { echo "::error::fetched $got, not $want"; exit 1; }
- name: The member's unit tests
shell: bash
working-directory: src/wai/wai-quantum
run: |
set -euo pipefail
export PATH="$HOME/.cargo/bin:$PATH"
CARGO_TARGET_DIR="$HOME/hw-target/wai-quantum-opencl" nice -n 10 cargo test --locked --features quantum_fabric_opencl --lib -- quantum_fabric_opencl:: quantum_fabric::opencl:: 2>&1 | tee "$RUNNER_TEMP/unit.log"
grep -qE '^test result: ok\. [1-9][0-9]* passed; 0 failed' "$RUNNER_TEMP/unit.log" || { echo "::error::no unit test ran"; exit 1; }
- name: The member's device tests
shell: bash
working-directory: src/wai/wai-quantum
run: |
set -euo pipefail
export PATH="$HOME/.cargo/bin:$PATH"
export WAI_QUANTUM_OPENCL_EXPECT="${WAI_QUANTUM_AMD_OPENCL_KEYS:-opencl-gpu-1002}"
echo "devices declared: $WAI_QUANTUM_OPENCL_EXPECT"
report="$RUNNER_TEMP/opencl-report.txt"; : > "$report"
WAI_QUANTUM_OPENCL_REPORT="$report" CARGO_TARGET_DIR="$HOME/hw-target/wai-quantum-opencl" nice -n 10 \
cargo test --locked --release --features quantum_fabric_opencl --test opencl_member -- --test-threads=1 2>&1 | tee "$RUNNER_TEMP/device.log"
grep -qE '^test result: ok\. 9 passed; 0 failed' "$RUNNER_TEMP/device.log" || { echo "::error::not every device test ran and passed"; exit 1; }
cat "$report"
grep -qx "opencl devices: as declared ($WAI_QUANTUM_OPENCL_EXPECT)" "$report" || { echo "::error::the devices found are not the declared ones"; exit 1; }
for key in ${WAI_QUANTUM_OPENCL_EXPECT//,/ }; do
n="$(grep -c "^opencl certified: key=$key .* parity=bit-exact " "$report")"
[ "$n" = 4 ] || { echo "::error::$key: $n of 4 certificates were earned"; exit 1; }
grep -qx "opencl mutants: key=$key killed 6 of 6" "$report" || { echo "::error::$key: a kernel mutant survived"; exit 1; }
done
- name: The dense f64 state vector, bit for bit
shell: bash
working-directory: src/wai/wai-quantum
run: |
set -euo pipefail
export PATH="$HOME/.cargo/bin:$PATH"
export WAI_QUANTUM_OPENCL_EXPECT="${WAI_QUANTUM_AMD_OPENCL_KEYS:-opencl-gpu-1002}"
report="$RUNNER_TEMP/opencl-sv-report.txt"; : > "$report"
WAI_QUANTUM_OPENCL_REPORT="$report" CARGO_TARGET_DIR="$HOME/hw-target/wai-quantum-opencl" nice -n 10 \
cargo test --locked --release --features quantum_fabric_opencl,quantum_sv --test opencl_sv -- --test-threads=1 2>&1 | tee "$RUNNER_TEMP/sv.log"
grep -qE '^test result: ok\. 4 passed; 0 failed' "$RUNNER_TEMP/sv.log" || { echo "::error::not every state-vector test ran and passed"; exit 1; }
cat "$report"
for key in ${WAI_QUANTUM_OPENCL_EXPECT//,/ }; do
if grep -qx "opencl sv refused: key=$key the device has no binary64" "$report"; then continue; fi
grep -q "^opencl sv certified: key=$key .* parity=bit-exact cases=48 digest=b7d129d0593adf70dc164477664970c5c0b8c09f79f1a0573036bd60888dbcd1$" "$report" || { echo "::error::$key: the state vector is not the CPU's bit for bit"; exit 1; }
grep -qx "opencl sv mutants: key=$key killed 4 of 4" "$report" || { echo "::error::$key: a state-vector mutant survived"; exit 1; }
# The wider kernels certify on their own corpus, or the member says
# they did not (it then runs gates on up to three qubits at once).
grep -qE "^opencl sv wide certified: key=$key .* kernel=wai.quantum.sv/run-wide .* parity=(bit-exact cases=120 digest=093614f2386ec5120ab6ccde5187a544518432751591535a6f58c7a44061ebaf|not-bit-exact cases=120 digest=[0-9a-f]{64})$" "$report" || { echo "::error::$key: the wider kernels' certificate was not established"; exit 1; }
grep -qE "^opencl sv random: key=$key [0-9]+ resident runs on (every kernel|the narrow kernels) match the CPU bit for bit$" "$report" || { echo "::error::$key: a random circuit's resident run is not the CPU's"; exit 1; }
done
opencl-macos:
name: member · OpenCL GPU (macos)
needs: relevant
if: needs.relevant.outputs.run == 'true'
runs-on: macos
timeout-minutes: 90
env:
WAI_QUANTUM_OPENCL_EXPECT: opencl-gpu-1027f00
steps:
- uses: actions/checkout@v5
# launchd's PATH omits cargo on the macOS host runner.
- name: Put cargo on PATH
shell: bash
run: echo "$HOME/.cargo/bin" >> "$GITHUB_PATH"
- name: The member's device tests
shell: bash
working-directory: wai/wai-quantum
run: |
set -euo pipefail
report="$RUNNER_TEMP/opencl-report.txt"; : > "$report"
WAI_QUANTUM_OPENCL_REPORT="$report" nice -n 10 \
cargo test --locked --release --features quantum_fabric_opencl --test opencl_member -- --test-threads=1 2>&1 | tee "$RUNNER_TEMP/device.log"
grep -qE '^test result: ok\. 9 passed; 0 failed' "$RUNNER_TEMP/device.log" || { echo "::error::not every device test ran and passed"; exit 1; }
cat "$report"
key="$WAI_QUANTUM_OPENCL_EXPECT"
grep -qx "opencl devices: as declared ($key)" "$report" || { echo "::error::the devices found are not the declared ones"; exit 1; }
n="$(grep -c "^opencl certified: key=$key .* parity=bit-exact " "$report")"
[ "$n" = 4 ] || { echo "::error::$n of 4 certificates were earned"; exit 1; }
grep -qx "opencl mutants: key=$key killed 6 of 6" "$report" || { echo "::error::a kernel mutant survived"; exit 1; }
- name: The state vector refuses a device without binary64, and says so
shell: bash
working-directory: wai/wai-quantum
run: |
set -euo pipefail
report="$RUNNER_TEMP/opencl-sv-report.txt"; : > "$report"
WAI_QUANTUM_OPENCL_REPORT="$report" nice -n 10 \
cargo test --locked --release --features quantum_fabric_opencl,quantum_sv --test opencl_sv -- --test-threads=1 2>&1 | tee "$RUNNER_TEMP/sv.log"
grep -qE '^test result: ok\. 4 passed; 0 failed' "$RUNNER_TEMP/sv.log" || { echo "::error::not every state-vector test ran and passed"; exit 1; }
grep -qx "opencl sv refused: key=$WAI_QUANTUM_OPENCL_EXPECT the device has no binary64" "$report" || { echo "::error::the refusal was not established"; exit 1; }
opencl-absent:
name: member · OpenCL absent (linux)
needs: relevant
if: needs.relevant.outputs.run == 'true'
runs-on: linux
timeout-minutes: 60
env:
WAI_QUANTUM_OPENCL_EXPECT: none
steps:
- uses: actions/checkout@v5
- name: Install Rust 1.98.0
shell: bash
run: |
export PATH="$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" >> "$GITHUB_PATH"
- name: No OpenCL here, and the member says so
shell: bash
working-directory: wai/wai-quantum
run: |
set -euo pipefail
report="$RUNNER_TEMP/opencl-report.txt"; : > "$report"
WAI_QUANTUM_OPENCL_REPORT="$report" cargo test --locked --features quantum_fabric_opencl --test opencl_member -- --test-threads=1 2>&1 | tee "$RUNNER_TEMP/device.log"
grep -qE '^test result: ok\. 9 passed; 0 failed' "$RUNNER_TEMP/device.log" || { echo "::error::not every device test ran and passed"; exit 1; }
grep -qx 'opencl devices: as declared (none)' "$report" || { echo "::error::the absent case was not established"; exit 1; }
cargo test --locked --features quantum_fabric_opencl,quantum_sv --test opencl_sv -- --test-threads=1 2>&1 | tee "$RUNNER_TEMP/sv.log"
grep -qE '^test result: ok\. 4 passed; 0 failed' "$RUNNER_TEMP/sv.log" || { echo "::error::the state-vector tests did not run over no device"; exit 1; }