dcharlot
bbc5e531ce
Some checks failed
repository checks / lockfiles, advisories, tests in CI, workflow hygiene, spec and section references, public text (push) Has been cancelled
305 lines
17 KiB
YAML
305 lines
17 KiB
YAML
name: wai-quantum CPU vector units
|
|
|
|
# The fabric's CPU vector-unit members: the gate's fixed-point arithmetic in
|
|
# the processor's own vector extension, certified bit-exact against the CPU
|
|
# reference's pinned corpus digests.
|
|
#
|
|
# - avx2 on silicon: the x86-64 processor of whichever `linux` runner takes
|
|
# the leg. jetson-hub's has avx2 alone; framework-128gb-linux's has avx512
|
|
# too, and there the leg declares both, as the kernel reports them, and
|
|
# checks the avx2 member as on any other;
|
|
# - avx2 and avx512 on silicon together: framework-128gb's processor, which has
|
|
# both, one member each;
|
|
# - sve and rvv under a user-mode emulator the job fetches and runs itself
|
|
# (Debian's static QEMU, pinned), never the host's binfmt handler, at
|
|
# several vector lengths each: the scalable kernels set their length per
|
|
# pass, and every leg checks the length the member read from the unit is the
|
|
# one the emulator was given, so a leg cannot pass at the wrong length;
|
|
# - the macOS runner, whose processor has no vector extension these kernels
|
|
# use outside streaming mode, finding no member.
|
|
#
|
|
# Silicon for sve and rvv: a processor with them runs the same tests with no
|
|
# emulator declared.
|
|
|
|
concurrency:
|
|
group: ${{ github.workflow }}-${{ github.ref }}
|
|
cancel-in-progress: true
|
|
|
|
on:
|
|
pull_request:
|
|
paths:
|
|
- 'wai/wai-quantum/**'
|
|
- 'rust-toolchain.toml'
|
|
- '.github/workflows/wai-quantum-vector.yml'
|
|
schedule:
|
|
- cron: '53 6 * * *'
|
|
|
|
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: gate
|
|
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-vector.yml \
|
|
"${{ github.event.pull_request.base.sha }}" \
|
|
"${{ github.event.pull_request.head.sha }}")
|
|
echo "this change touches wai-quantum-vector: $r"
|
|
echo "run=$r" >> "$GITHUB_OUTPUT"
|
|
|
|
vector:
|
|
name: member · vector unit, ${{ matrix.name }}
|
|
needs: relevant
|
|
if: needs.relevant.outputs.run == 'true'
|
|
runs-on: linux
|
|
timeout-minutes: 120
|
|
strategy:
|
|
fail-fast: false
|
|
max-parallel: 1
|
|
matrix:
|
|
include:
|
|
- name: avx2 (x86_64 silicon)
|
|
target: x86_64-unknown-linux-gnu
|
|
image: ghcr.io/cross-rs/aarch64-unknown-linux-gnu@sha256:7f8308a8734d9fcd2ebbe9a3e4bdea74af293f0799d80c3cc341e340cda49a4c
|
|
key: cpu-vector-avx2
|
|
bits: 256
|
|
emulator: ''
|
|
qemu: ''
|
|
qemu_cpu: ''
|
|
- name: sve 128-bit (aarch64, emulated)
|
|
target: aarch64-unknown-linux-gnu
|
|
image: ghcr.io/cross-rs/aarch64-unknown-linux-gnu@sha256:7f8308a8734d9fcd2ebbe9a3e4bdea74af293f0799d80c3cc341e340cda49a4c
|
|
key: cpu-vector-sve
|
|
bits: 128
|
|
emulator: qemu-user 10.0.13 (Debian's static build, run by the job)
|
|
qemu: aarch64
|
|
qemu_cpu: max,sve-default-vector-length=16
|
|
- name: sve 512-bit (aarch64, emulated)
|
|
target: aarch64-unknown-linux-gnu
|
|
image: ghcr.io/cross-rs/aarch64-unknown-linux-gnu@sha256:7f8308a8734d9fcd2ebbe9a3e4bdea74af293f0799d80c3cc341e340cda49a4c
|
|
key: cpu-vector-sve
|
|
bits: 512
|
|
emulator: qemu-user 10.0.13 (Debian's static build, run by the job)
|
|
qemu: aarch64
|
|
qemu_cpu: max,sve-default-vector-length=64
|
|
- name: sve 2048-bit (aarch64, emulated)
|
|
target: aarch64-unknown-linux-gnu
|
|
image: ghcr.io/cross-rs/aarch64-unknown-linux-gnu@sha256:7f8308a8734d9fcd2ebbe9a3e4bdea74af293f0799d80c3cc341e340cda49a4c
|
|
key: cpu-vector-sve
|
|
bits: 2048
|
|
emulator: qemu-user 10.0.13 (Debian's static build, run by the job)
|
|
qemu: aarch64
|
|
qemu_cpu: max,sve-default-vector-length=256
|
|
- name: rvv 128-bit (riscv64, emulated)
|
|
target: riscv64gc-unknown-linux-gnu
|
|
image: ghcr.io/cross-rs/riscv64gc-unknown-linux-gnu@sha256:67ee97dd4719390d299edfcf39f226360ccfb1539e08b078ec187d4848415794
|
|
key: cpu-vector-rvv
|
|
bits: 128
|
|
emulator: qemu-user 10.0.13 (Debian's static build, run by the job)
|
|
qemu: riscv64
|
|
qemu_cpu: rv64,v=true,vlen=128,elen=64
|
|
- name: rvv 1024-bit (riscv64, emulated)
|
|
target: riscv64gc-unknown-linux-gnu
|
|
image: ghcr.io/cross-rs/riscv64gc-unknown-linux-gnu@sha256:67ee97dd4719390d299edfcf39f226360ccfb1539e08b078ec187d4848415794
|
|
key: cpu-vector-rvv
|
|
bits: 1024
|
|
emulator: qemu-user 10.0.13 (Debian's static build, run by the job)
|
|
qemu: riscv64
|
|
qemu_cpu: rv64,v=true,vlen=1024,elen=64
|
|
container:
|
|
image: ${{ matrix.image }}
|
|
options: --memory=6g --memory-swap=6g
|
|
env:
|
|
WAI_QUANTUM_VECTOR_EXPECT: ${{ matrix.key }}
|
|
WAI_QUANTUM_VECTOR_EMULATOR: ${{ matrix.emulator }}
|
|
# The emulator every foreign test binary runs under: Debian's static QEMU,
|
|
# SHA-256 pinned, fetched by the job and named as cargo's runner. Not the
|
|
# image's own (5.1 for aarch64, 6.1 for riscv64), and not the host's binfmt
|
|
# handler, which some `linux` runners lack and others run at another
|
|
# version. (Not QEMU_*: qemu-user reads every QEMU_<option> variable as that
|
|
# option.)
|
|
USER_EMULATOR_VERSION: 10.0.13
|
|
USER_EMULATOR_DEB: qemu-user_10.0.13+ds-0+deb13u1_amd64.deb
|
|
USER_EMULATOR_SHA256: ca6ede739327a20ae5a3498230c8a3a9a072c586e131bc6bcc1201eb1725e336
|
|
USER_EMULATOR_SHA1: dec1373851d29f91642c197f4a339d0d2ad25ed0
|
|
steps:
|
|
# The image has no Node for a checkout action: the forge's archive of the
|
|
# exact commit, its embedded commit id checked.
|
|
- name: Fetch the source
|
|
shell: bash
|
|
run: |
|
|
set -euo pipefail
|
|
want="${{ github.event.pull_request.head.sha || github.sha }}"
|
|
rm -rf src src.tar.gz src.tar && mkdir src
|
|
curl -sSf -o src.tar.gz "http://100.64.224.0:3400/Transaction-Science/open-standards/archive/$want.tar.gz"
|
|
gzip -dc src.tar.gz > src.tar
|
|
got="$(git get-tar-commit-id < src.tar)"
|
|
[ "$got" = "$want" ] || { echo "::error::the archive is of $got, not $want"; exit 1; }
|
|
# -m: every file gets the time it was extracted, not its commit's. The build
|
|
# reuses a target directory kept between runs, and cargo takes a crate as
|
|
# built when no source file is newer than its last build: with the commit's
|
|
# times, an older commit ran the binary another commit had built there.
|
|
tar -xmf src.tar -C src --strip-components=1
|
|
rm src.tar.gz src.tar
|
|
- name: The member's tests
|
|
shell: bash
|
|
working-directory: src/wai/wai-quantum
|
|
run: |
|
|
set -euo pipefail
|
|
cache=/ci-cache/wai-quantum-vector
|
|
mkdir -p "$cache"
|
|
exec 9>"$cache/.lock"
|
|
flock 9
|
|
# Every source file newer than every build in the cache, now that the lock is
|
|
# held: another run may have built here while this one waited for it, after
|
|
# this one's extraction. (../.. is the extracted tree.)
|
|
find ../.. -type f -exec touch {} +
|
|
export RUSTUP_HOME="$cache/rustup" CARGO_HOME="$cache/cargo" CARGO_TARGET_DIR="$cache/target-${{ matrix.target }}"
|
|
export PATH="$CARGO_HOME/bin:$PATH"
|
|
if ! command -v rustup >/dev/null; then
|
|
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --no-modify-path --profile minimal --default-toolchain none
|
|
fi
|
|
rustup toolchain install
|
|
rustup target add ${{ matrix.target }}
|
|
# The emulated processor, for the test binaries only.
|
|
if [ -n "${{ matrix.qemu_cpu }}" ]; then export QEMU_CPU="${{ matrix.qemu_cpu }}"; fi
|
|
# A trivial program for the preflight below: it prints the machine it runs on.
|
|
cat > "$RUNNER_TEMP/probe.c" <<'EOF'
|
|
#include <stdio.h>
|
|
#include <sys/utsname.h>
|
|
int main(void) { struct utsname u; if (uname(&u) != 0) return 1; printf("%s\n", u.machine); return 0; }
|
|
EOF
|
|
if [ -n "${{ matrix.qemu }}" ]; then
|
|
# The emulator, fetched once per host into the cache (under the lock above)
|
|
# and checked against its pin, then named as cargo's runner, so this leg
|
|
# passes or fails the same on every `linux` runner.
|
|
q="$cache/qemu-user-$USER_EMULATOR_VERSION"
|
|
if [ ! -x "$q/qemu-aarch64" ] || [ ! -x "$q/qemu-riscv64" ]; then
|
|
d="$cache/qemu-user-fetch"; rm -rf "$d" && mkdir "$d"
|
|
curl -sSfL -o "$d/$USER_EMULATOR_DEB" "https://deb.debian.org/debian/pool/main/q/qemu/$USER_EMULATOR_DEB" \
|
|
|| curl -sSfL -o "$d/$USER_EMULATOR_DEB" "https://snapshot.debian.org/file/$USER_EMULATOR_SHA1"
|
|
echo "$USER_EMULATOR_SHA256 $d/$USER_EMULATOR_DEB" | sha256sum -c -
|
|
dpkg-deb --fsys-tarfile "$d/$USER_EMULATOR_DEB" | tar -x -C "$d" ./usr/bin/qemu-aarch64 ./usr/bin/qemu-riscv64
|
|
rm -rf "$q.tmp" && mkdir "$q.tmp" && mv "$d/usr/bin/qemu-aarch64" "$d/usr/bin/qemu-riscv64" "$q.tmp/"
|
|
rm -rf "$q" && mv "$q.tmp" "$q" && rm -rf "$d"
|
|
fi
|
|
runner="$q/qemu-${{ matrix.qemu }} -L /usr/${{ matrix.qemu }}-linux-gnu"
|
|
export "CARGO_TARGET_$(printf %s "${{ matrix.target }}" | tr 'a-z-' 'A-Z_')_RUNNER=$runner"
|
|
# Preflight: the trivial program, built for the target and run as cargo will
|
|
# run the tests. An emulator that cannot run it fails the leg here, saying
|
|
# so, rather than as a test that never ran.
|
|
"${{ matrix.qemu }}-linux-gnu-gcc" -o "$RUNNER_TEMP/probe" "$RUNNER_TEMP/probe.c"
|
|
got="$($runner "$RUNNER_TEMP/probe")" || { echo "::error::the emulator cannot run a program for ${{ matrix.target }}: $runner"; exit 1; }
|
|
[ "$got" = "${{ matrix.qemu }}" ] || { echo "::error::the preflight program ran as $got, not ${{ matrix.qemu }}"; exit 1; }
|
|
echo "emulator: $("$q/qemu-${{ matrix.qemu }}" --version | head -1), cpu ${QEMU_CPU:-default}; a program for ${{ matrix.target }} runs under it"
|
|
fi
|
|
# Every unit this processor has must be declared, so on silicon the
|
|
# declaration comes from the kernel's own flags, not the probe under test.
|
|
if [ "${{ matrix.key }}" = cpu-vector-avx2 ] && grep -qw avx512f /proc/cpuinfo; then
|
|
export WAI_QUANTUM_VECTOR_EXPECT=cpu-vector-avx2,cpu-vector-avx512
|
|
fi
|
|
t="--target ${{ matrix.target }} --features quantum_fabric_vector"
|
|
cargo test --locked --release -j 6 $t --lib -- quantum_fabric_vector:: 2>&1 | tee "$cache/unit.log"
|
|
# The first build after the refresh compiles wai-quantum; one that compiled
|
|
# none ran a binary another commit left in the cache.
|
|
grep -qE '^ *Compiling wai-quantum v[0-9]' "$cache/unit.log" || { echo "::error::the build compiled no wai-quantum: its tests ran a binary already in the cache"; exit 1; }
|
|
grep -qE '^test result: ok\. [1-9][0-9]* passed; 0 failed' "$cache/unit.log" || { echo "::error::no unit test ran"; exit 1; }
|
|
report="$(mktemp)"
|
|
WAI_QUANTUM_VECTOR_REPORT="$report" cargo test --locked --release -j 6 $t --test vector_member -- --test-threads=1 2>&1 | tee "$cache/device.log"
|
|
grep -qE '^test result: ok\. 9 passed; 0 failed' "$cache/device.log" || { echo "::error::not every device test ran and passed"; exit 1; }
|
|
cat "$report"
|
|
key="${{ matrix.key }}"
|
|
grep -qx "vector devices: as declared ($WAI_QUANTUM_VECTOR_EXPECT)" "$report" || { echo "::error::the units found are not the declared ones"; exit 1; }
|
|
grep -q "^vector member: key=$key .* bits=${{ matrix.bits }} " "$report" || { echo "::error::the unit's vector length is not ${{ matrix.bits }} bits"; exit 1; }
|
|
if [ -n "$WAI_QUANTUM_VECTOR_EMULATOR" ]; then
|
|
grep -q "^vector member: key=$key .*emulator:qemu-user" "$report" || { echo "::error::the member does not say it ran under the emulator"; exit 1; }
|
|
else
|
|
grep -q "^vector member: key=$key .*; silicon;" "$report" || { echo "::error::the member does not say it ran on silicon"; exit 1; }
|
|
fi
|
|
n="$(grep -c "^vector certified: key=$key .* parity=bit-exact " "$report")"
|
|
[ "$n" = 4 ] || { echo "::error::$n of 4 certificates were earned"; exit 1; }
|
|
grep -qx "vector mutants: key=$key killed 3 of 3" "$report" || { echo "::error::a kernel mutant survived"; exit 1; }
|
|
|
|
vector-amd:
|
|
name: member · vector units, avx2 and avx512 (framework-128gb silicon)
|
|
needs: relevant
|
|
if: needs.relevant.outputs.run == 'true'
|
|
runs-on: linux-amd
|
|
timeout-minutes: 60
|
|
env:
|
|
CARGO_PROFILE_DEV_DEBUG: '0'
|
|
CARGO_BUILD_JOBS: '12'
|
|
steps:
|
|
- name: Fetch the source
|
|
shell: bash
|
|
run: |
|
|
set -euo pipefail
|
|
want="${{ 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. 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 $want"; exit 1; }
|
|
- name: Both units on silicon
|
|
shell: bash
|
|
working-directory: src/wai/wai-quantum
|
|
run: |
|
|
set -euo pipefail
|
|
export PATH="$HOME/.cargo/bin:$PATH"
|
|
export CARGO_TARGET_DIR="$HOME/hw-target/wai-quantum-vector"
|
|
# The units this host's processor has, declared here (or by the runner).
|
|
export WAI_QUANTUM_VECTOR_EXPECT="${WAI_QUANTUM_AMD_VECTOR_EXPECT:-cpu-vector-avx2,cpu-vector-avx512}"
|
|
nice -n 10 cargo test --locked --release --features quantum_fabric_vector --lib -- quantum_fabric_vector:: 2>&1 | tee "$RUNNER_TEMP/unit.log"
|
|
grep -qE '^test result: ok\. [1-9][0-9]* passed; 0 failed' "$RUNNER_TEMP/unit.log" || { echo "::error::no unit test ran"; exit 1; }
|
|
report="$RUNNER_TEMP/vector-report.txt"; : > "$report"
|
|
WAI_QUANTUM_VECTOR_REPORT="$report" nice -n 10 cargo test --locked --release --features quantum_fabric_vector --test vector_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 "$report"
|
|
grep -qx "vector devices: as declared ($WAI_QUANTUM_VECTOR_EXPECT)" "$report" || { echo "::error::the units found are not the declared ones"; exit 1; }
|
|
for key in ${WAI_QUANTUM_VECTOR_EXPECT//,/ }; do
|
|
case "$key" in cpu-vector-avx2) bits=256 ;; cpu-vector-avx512) bits=512 ;; *) echo "::error::no vector length known for $key"; exit 1 ;; esac
|
|
grep -q "^vector member: key=$key .* bits=$bits .*; silicon;" "$report" || { echo "::error::$key: not $bits bits on silicon"; exit 1; }
|
|
n="$(grep -c "^vector certified: key=$key .* parity=bit-exact " "$report")"
|
|
[ "$n" = 4 ] || { echo "::error::$key: $n of 4 certificates were earned"; exit 1; }
|
|
grep -qx "vector mutants: key=$key killed 3 of 3" "$report" || { echo "::error::$key: a kernel mutant survived"; exit 1; }
|
|
done
|
|
|
|
vector-absent:
|
|
name: member · vector unit absent (macos)
|
|
needs: relevant
|
|
if: needs.relevant.outputs.run == 'true'
|
|
runs-on: macos
|
|
timeout-minutes: 60
|
|
env:
|
|
WAI_QUANTUM_VECTOR_EXPECT: none
|
|
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: No vector unit these kernels use, and the member says so
|
|
shell: bash
|
|
working-directory: wai/wai-quantum
|
|
run: |
|
|
set -euo pipefail
|
|
report="$RUNNER_TEMP/vector-report.txt"; : > "$report"
|
|
nice -n 10 cargo test --locked --features quantum_fabric_vector --lib -- quantum_fabric_vector:: 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; }
|
|
WAI_QUANTUM_VECTOR_REPORT="$report" nice -n 10 cargo test --locked --features quantum_fabric_vector --test vector_member -- --test-threads=1 2>&1 | tee "$RUNNER_TEMP/device.log"
|
|
grep -qE '^test result: ok\. 9 passed; 0 failed' "$RUNNER_TEMP/device.log" || { echo "::error::not every device test ran and passed"; exit 1; }
|
|
grep -qx 'vector devices: as declared (none)' "$report" || { echo "::error::the absent case was not established"; exit 1; }
|