mawei/StarPilot
1
0
Fork 0
mirror of https://github.com/firestar5683/StarPilot.git synced 2026-09-10 18:23:44 +08:00
Code Issues Packages Projects Releases Wiki Activity

The smallest promotion

Browse Source
This commit is contained in:
firestar5683
2026-08-10 22:53:46 -05:00
parent d938c9daa6
commit 6bd48c28e8
582 changed files with 64651 additions and 17100 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tinygrad_repo/.github/workflows/autogen.yml
Vendored
+1 -1
View File
@@ -42,7 +42,7 @@ jobs:
run: |
find tinygrad/runtime/autogen -type f -name "*.py" -not -path "*/amd/*" -not -name "__init__.py" -not -name "comgr.py" -not -name "metal.py" -not -name "iokit.py" -not -name "corefoundation.py" -not -name "libclang.py" -delete
python3 -c "from tinygrad.runtime.autogen import opencl"
python3 -c "from tinygrad.runtime.autogen import cuda, nvrtc, nvjitlink, nv_570, nv_580, nv"
python3 -c "from tinygrad.runtime.autogen import cuda, nvrtc, nvjitlink, nv_570, nv_580, nv_610, nv"
python3 -c "from tinygrad.runtime.autogen import comgr_3, hsa, hip, amd_gpu, sqtt, rocprof, amdgpu_kd, amdgpu_drm"
python3 -c "from tinygrad.runtime.autogen.am import *"
python3 -c "from tinygrad.runtime.autogen.nv_regs import *"
tinygrad_repo/.github/workflows/benchmark.yml
Vendored
+352 -586
View File
File diff suppressed because it is too large Load Diff
tinygrad_repo/.github/workflows/mlperf.yml
Vendored
+2 -2
View File
@@ -1,8 +1,8 @@
name: Run MLPerf Training
on:
schedule:
- cron: '5 8 * * *' # Runs at 08:05 UTC (12:05 AM Pacific Time)
#schedule:
# - cron: '5 8 * * *' # Runs at 08:05 UTC (12:05 AM Pacific Time)
push:
branches:
- update_mlperf
tinygrad_repo/.github/workflows/platform.yml
Vendored
+213
View File
@@ -0,0 +1,213 @@
name: Platform Tests
env:
# increment this when downloads substantially change to avoid the internet
CACHE_VERSION: '19'
CAPTURE_PROCESS_REPLAY: ${{ github.event_name == 'pull_request' && contains(github.event.pull_request.title, '[pr]') && '1' || '0' }}
GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
PYTHONPATH: ${{ github.workspace }}
CHECK_OOB: 1
on:
push:
branches:
- master
pull_request:
workflow_dispatch:
concurrency:
group: platform-${{ github.event_name }}-${{ github.event_name == 'pull_request' && github.event.pull_request.number || github.run_id }}
cancel-in-progress: ${{ github.event_name == 'pull_request' }}
jobs:
# ****** OSX Tests ******
unittestmacos:
name: MacOS (unit)
runs-on: macos-26
timeout-minutes: 20
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: unittest-macos
deps: testing_unit
- name: Run unit tests
run: DEV=METAL python -m pytest -n=auto test/unit/ --durations=20
- name: Test tensor core ops (fake)
run: DEV=METAL DEBUG=3 TC=2 python test/backend/test_ops.py TestOps.test_gemm
- name: Test tensor core ops (real)
run: DEV=METAL DEBUG=3 python test/backend/test_ops.py TestOps.test_big_gemm
- name: Test Beam Search
run: DEV=METAL IGNORE_BEAM_CACHE=1 python3 -m pytest extra/optimization/test_beam_search.py
- name: Test Device Specific
run: DEV=METAL python3 -m pytest test/device/test_metal.py
#- name: Fuzz Test linearizer
# run: DEV=METAL DEPTH=4 FUZZ_N=50 FUZZ_MAX_SIZE=1000000 python test/external/fuzz_linearizer.py
- name: Run process replay tests
uses: ./.github/actions/process-replay
unittestmacosmock:
name: MacOS (unit, mock)
runs-on: macos-26
timeout-minutes: 20
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: unittest-macos-mock
deps: testing_unit
amd: 'true'
ocelot: 'true'
- name: Run NULL backend tests
run: SPEC=2 DEV=NULL python -m pytest -n=auto test/null/ --durations=20
- name: Run pytest (amd)
env:
DEV: MOCKKFD+AMD
FORWARD_ONLY: 1
run: |
python3 -m pytest -n=auto test/device/test_hcq.py test/test_tiny.py --durations=20
- name: Run pytest (ptx)
env:
DEV: "MOCK+NV:PTX"
FORWARD_ONLY: 1
# TODO: failing due to library loading error
CAPTURE_PROCESS_REPLAY: 0
run: |
python3 -m pytest -n=auto test/device/test_hcq.py test/test_tiny.py \
test/testextra/test_hevc.py::TestHevc::test_hevc_decode_compile --durations=20
- name: Run process replay tests
uses: ./.github/actions/process-replay
testmetal:
strategy:
fail-fast: false
matrix:
group: [1, 2]
name: MacOS (DEV=METAL) (${{ matrix.group }})
runs-on: macos-26
timeout-minutes: 20
env:
DEV: METAL
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: macos-metal
deps: testing_unit
- name: Check Device.DEFAULT and print some source
run: |
python -c "from tinygrad import Device; assert Device.DEFAULT == 'METAL'"
DEBUG=4 python test/test_tiny.py TestTiny.test_plus
- name: Run backend tests
run: python -m pytest -n=auto test/backend --durations=20 --splits 2 --group ${{ matrix.group }}
- name: Run process replay tests
uses: ./.github/actions/process-replay
testmacos:
strategy:
fail-fast: false
matrix:
dev:
- 'CPU:CLANG'
- 'CPU:LLVM'
- 'CPU:LVP'
- 'WEBGPU'
name: MacOS (DEV=${{ matrix.dev }})
runs-on: macos-26
timeout-minutes: 20
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: macos-${{ matrix.dev }}
deps: "testing_unit${{ contains(matrix.dev, 'LVP') && ' mesa' || '' }}"
llvm: ${{ contains(matrix.dev, 'LLVM') || contains(matrix.dev, 'LVP') }}
webgpu: ${{ matrix.dev == 'WEBGPU' }}
- name: Set env
run: printf "DEV=${{ matrix.dev }}${{ matrix.dev == 'CPU:CLANG' && '\nCPU_COUNT=2' || '' }}" >> $GITHUB_ENV
- name: Check Device.DEFAULT and print some source
run: |
python -c "from tinygrad import Device; from tinygrad.helpers import Target; assert Device.DEFAULT == Target.parse('${{ matrix.dev }}').device"
DEBUG=4 python test/test_tiny.py TestTiny.test_plus
- name: Run test_tiny
run: python -m pytest -n=auto test/test_tiny.py --durations=20
- name: Run process replay tests
uses: ./.github/actions/process-replay
# ****** Windows Tests ******
testwindows:
strategy:
fail-fast: false
matrix:
dev:
- 'CPU:CLANG'
- 'CPU:LLVM'
- 'CPU:X86'
- 'WEBGPU'
name: Windows (DEV=${{ matrix.dev }})
runs-on: windows-2025
timeout-minutes: 15
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: windows-${{ matrix.dev }}-minimal
deps: testing_unit
pydeps: ${{ matrix.dev == 'WEBGPU' && 'dawn-python' || '' }}
- name: Set env
shell: bash
run: printf "DEV=${{ matrix.dev }}${{ matrix.dev == 'CPU:CLANG' && '\nCPU_COUNT=2' || '' }}" >> $GITHUB_ENV
- name: Check Device.DEFAULT and print some source
shell: bash
run: |
python -c "from tinygrad import Device; from tinygrad.helpers import Target; assert Device.DEFAULT == Target.parse('${{ matrix.dev }}').device"
DEBUG=4 python test/test_tiny.py TestTiny.test_plus
- name: Run test_tiny
shell: bash
run: python -m pytest -n=auto test/test_tiny.py --durations=20
qcomclcompiletests:
name: Compile-only (QCOM CL)
runs-on: ubuntu-24.04-arm
timeout-minutes: 15
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: compile-qcomcl
deps: testing_unit
tinydreno: 'true'
- name: Set env
shell: bash
run: printf "DEV=NULL:QCOMCL:a630\nNULL_ALLOW_COPYOUT=1" >> $GITHUB_ENV
- name: Run test_ops
shell: bash
run: |
python -c "from tinygrad import Device; assert Device.DEFAULT == 'NULL'"
DEBUG=4 python3 test/backend/test_ops.py TestOps.test_add
python -m pytest -n=auto test/backend/test_ops.py --durations=20
- name: Run test_ops (IMAGE)
shell: bash
env:
IMAGE: 1
DEV: "NULL:QCOMCL:a630,IMAGE_PITCH_ALIGNMENT=64"
run: |
DEBUG=4 python test/backend/test_ops.py TestOps.test_gemm | grep read_imagef
python -m pytest -n=auto test/backend/test_ops.py --durations=20
tinygrad_repo/.github/workflows/szdiff.yml
Vendored
+7 -1
View File
@@ -14,12 +14,15 @@ jobs:
outputs:
branchstat: ${{ steps.brstat.outputs.stat}}
steps:
- name: Check code from PR branch
- name: Check code from PR branch
uses: actions/checkout@v6
with:
repository: ${{ github.event.pull_request.head.repo.full_name }}
ref: ${{ github.event.pull_request.head.sha }}
fetch-depth: 0
# PR code is only inspected with git rev-list, never executed
allow-unsafe-pr-checkout: true
persist-credentials: false
- name: Check whether branch is up-to-date
id: brstat
run: |
@@ -51,6 +54,9 @@ jobs:
repository: ${{ github.event.pull_request.head.repo.full_name }}
ref: ${{ github.event.pull_request.head.sha }}
path: pr
# PR code is only line-counted by master's sz.py, never executed
allow-unsafe-pr-checkout: true
persist-credentials: false
# the base default to tinygrad master and cannot be other fork branch for security purpose
- name: Checkout code from tinygrad master
uses: actions/checkout@v6
tinygrad_repo/.github/workflows/test.yml
Vendored
+71 -300
View File
@@ -77,45 +77,14 @@ jobs:
run: |
sudo apt update || true
sudo apt install -y --no-install-recommends ninja-build
- name: Test one op
run: FORWARD_ONLY=1 TINY_BACKEND=1 python3 test/test_tiny.py TestTiny.test_plus
- name: Test ResNet-18
run: DEBUG=2 python3 extra/torch_backend/example.py
- name: custom tests
run: python3 -m pytest -n auto extra/torch_backend/test.py --durations=20
- name: Test one op in torch tests
run: DEBUG=2 python3 extra/torch_backend/torch_tests.py TestTinyBackendPRIVATEUSE1.test_unary_log_tiny_float32
- name: Test Ops with TINY_BACKEND
run: DEV=CPU:LLVM LLVMOPT=0 TINY_BACKEND=1 python3 -m pytest -n auto test/backend/test_ops.py --durations=20
- name: Test in-place operations on views
run: TORCH_DEBUG=1 python3 extra/torch_backend/test_inplace.py
- name: Test multi-gpu
run: DEV=CPU:LLVM GPUS=4 TORCH_DEBUG=1 python3 extra/torch_backend/test_multigpu.py
- name: Test kernel fusion
run: python3 extra/torch_backend/test_kernel_fusion.py
torchbackendmore:
name: Torch Backend Tests More
runs-on: *linux
timeout-minutes: 15
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: torch-backend-pillow-torchvision-et-pt
deps: testing_unit
llvm: 'true'
- name: Install ninja
run: |
sudo apt update || true
sudo apt install -y --no-install-recommends ninja-build
- name: Test beautiful_mnist in torch with TINY_BACKEND
run: STEPS=20 DEV=CPU TARGET_EVAL_ACC_PCT=90.0 MAX_BUFFER_SIZE=0 TINY_BACKEND=1 python3 examples/other_mnist/beautiful_mnist_torch.py
- name: Test some torch tests (expect failure)
run: python3 -m pytest extra/torch_backend/torch_tests.py -v --tb=no || true
- name: Custom tests
run: DEV=CPU:LLVM GPUS=4 TINY_BACKEND=1 python3 -m pytest -nauto extra/torch_backend/test.py extra/torch_backend/test_inplace.py extra/torch_backend/test_multigpu.py extra/torch_backend/test_kernel_fusion.py --durations=20
bepython:
name: Python Backend
@@ -133,46 +102,26 @@ jobs:
run: SKIP_SLOW_TEST=1 DEV=PYTHON python3 -m pytest -n=auto test/backend/test_dtype.py test/backend/test_dtype_alu.py test/backend/test_ops.py test/backend/test_uops.py test/backend/test_symbolic_ops.py test/backend/test_renderer_failures.py::TestRendererFailures --durations=20
- name: Test IMAGE support
run: IMAGE=1 DEV=PYTHON python3 test/backend/test_ops.py TestOps.test_gemm TestOps.test_simple_conv2d
- name: Test emulated METAL tensor cores
- name: Test emulated tensor cores
env:
DEV: 'PYTHON::METAL'
DEBUG: 2
N: 64
CNT: 1
SHOULD_USE_TC: 1
run: |
DEBUG=2 python3 test/backend/test_ops.py TestOps.test_big_gemm
python3 -m pytest -nauto test/opt/test_tensor_cores.py
- name: Test emulated AMD tensor cores
env:
DEV: 'PYTHON::gfx1100'
parallel -k --link --tagstring '[{1}]' '{2} python3 ./extra/gemm/simple_matmul.py' \
::: metal gfx950 gfx1100 gfx1100_acchalf gfx1201 gfx1201_acchalf sm_75 sm_80_half sm_80_tf32 \
::: 'DEV=PYTHON::METAL' 'DEV=PYTHON::gfx950 HALF=1 ACC_HALF=0' \
'DEV=PYTHON::gfx1100 HALF=1 ACC_HALF=0' 'DEV=PYTHON::gfx1100 HALF=1 ACC_HALF=1 ATOL=1e-3' \
'DEV=PYTHON::gfx1201 HALF=1 ACC_HALF=0' 'DEV=PYTHON::gfx1201 HALF=1 ACC_HALF=1 ATOL=1e-3' \
'DEV=PYTHON::sm_75 HALF=1' 'DEV=PYTHON::sm_80 HALF=1' 'DEV=PYTHON::sm_80 ALLOW_TF32=1'
- name: Run additional tensor core tests
run: |
DEBUG=2 N=16 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
DEBUG=2 N=64 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
DEBUG=2 N=16 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
DEBUG=2 N=64 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
python3 -m pytest -nauto test/opt/test_tensor_cores.py
- name: Test emulated AMD MFMA tensor cores
env:
DEV: 'PYTHON::gfx950'
run: |
DEBUG=2 N=64 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
python3 -m pytest -nauto test/opt/test_tensor_cores.py
- name: Test emulated AMD RDNA4 tensor cores
env:
DEV: 'PYTHON::gfx1201'
run: |
DEBUG=2 N=16 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
DEBUG=2 N=64 HALF=1 ACC_HALF=0 python3 ./extra/gemm/simple_matmul.py
DEBUG=2 N=16 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
DEBUG=2 N=64 HALF=1 ACC_HALF=1 ATOL=1e-3 python3 ./extra/gemm/simple_matmul.py
python3 -m pytest -nauto test/opt/test_tensor_cores.py
- name: Test emulated CUDA tensor cores
run: |
DEBUG=2 DEV=PYTHON::sm_80 python3 test/backend/test_ops.py TestOps.test_gemm_fp16
DEBUG=2 ALLOW_TF32=1 DEV=PYTHON::sm_80 python3 test/backend/test_ops.py TestOps.test_gemm
DEBUG=2 DEV=PYTHON::sm_75 python3 test/backend/test_ops.py TestOps.test_gemm_fp16
DEV=PYTHON::METAL python3 -m pytest -nauto test/opt/test_tensor_cores.py test/null/test_uops_stats.py::TestUOpsStatsMatmulHalf
DEV=PYTHON::gfx1100 python3 -m pytest -nauto test/opt/test_tensor_cores.py test/null/test_uops_stats.py::TestUOpsStatsMatmulHalf
DEV=PYTHON::gfx950 python3 -m pytest -nauto test/opt/test_tensor_cores.py
DEV=PYTHON::gfx1201 python3 -m pytest -nauto test/opt/test_tensor_cores.py
ALLOW_TF32=1 DEV=PYTHON::sm_89 python3 -m pytest -nauto test/opt/test_tensor_cores.py
- name: Test device flop counts
run: |
DEBUG=2 DEV=PYTHON::METAL python3 ./test/null/test_uops_stats.py TestUOpsStatsMatmulHalf
DEBUG=2 DEV=PYTHON::gfx1100 python3 ./test/null/test_uops_stats.py TestUOpsStatsMatmulHalf
DEBUG=2 DEV=PYTHON::sm_80 python3 ./test/null/test_uops_stats.py TestUOpsStatsMatmulHalf
linter:
@@ -218,21 +167,24 @@ jobs:
uses: ./.github/actions/setup-tinygrad
with:
key: unittest-13
pydeps: "pillow ftfy regex pre-commit"
python-version: '3.11'
deps: testing_unit
llvm: 'true'
amd: 'true'
- name: Run NULL backend tests
run: DEV=NULL python -m pytest -n=auto test/null/ --durations=20
run: SPEC=2 DEV=NULL python -m pytest -n=auto test/null/ --durations=20
- name: Run targeted tests on NULL backend
run: DEV=NULL python3 -m unittest test.backend.test_multitensor.TestMultiTensor.test_data_parallel_resnet_train_step
run: |
DEV=NULL python3 -m unittest test.backend.test_multitensor.TestMultiTensor.test_data_parallel_resnet_train_step
DEV=NULL VIZ=1 python3 -m pytest -n=auto test/null/test_viz.py
DEBUG=7 python -m tinygrad.viz.cli --json | jq empty
# TODO: too slow
# - name: Run SDXL on NULL backend
# run: DEV=NULL DEBUG=1 python3 examples/sdxl.py --seed 0 --noshow --timing --fakeweights
- name: Run Clip tests for SD MLPerf on NULL backend
run: DEV=NULL python -m pytest -n=auto test/external/mlperf_stable_diffusion/external_test_models.py::TestOpenClip --durations=20
- name: Run AMD emulated BERT training on NULL backend
run: DEV=NULL::gfx1201 NULL_ALLOW_COPYOUT=1 CAPTURE_PROCESS_REPLAY=0 DEFAULT_FLOAT=HALF BENCHMARK=10 BS=66 GPUS=1 BERT_LAYERS=2 MODEL=bert python3 examples/mlperf/model_train.py
run: DEV=NULL::gfx1201 NULL_ALLOW_COPYOUT=1 DEFAULT_FLOAT=HALF BENCHMARK=10 BS=66 GPUS=1 BERT_LAYERS=2 MODEL=bert python3 examples/mlperf/model_train.py
# TODO: support fake weights
#- name: Run LLaMA 7B on 4 fake devices
# run: DEV=NULL python3 examples/llama.py --gen 1 --size 7B --shard 4 --prompt "Hello." --count 3 --temperature 0 --timing
@@ -249,7 +201,8 @@ jobs:
uses: ./.github/actions/setup-tinygrad
with:
key: unittest-13
pydeps: "pillow ftfy regex pre-commit"
python-version: '3.11'
pydeps: "pre-commit"
deps: testing_unit
llvm: 'true'
- name: Run pre-commit test hooks
@@ -266,15 +219,8 @@ jobs:
run: python3 test/external/external_benchmark_schedule.py
- name: Run process replay tests
uses: ./.github/actions/process-replay
- name: Regen dataset on test_tiny
run: |
test/external/process_replay/reset.py
CAPTURE_PROCESS_REPLAY=1 python test/test_tiny.py TestTiny.test_plus
python extra/optimization/extract_dataset.py
gzip -c /tmp/sops > extra/datasets/sops.gz
#DEBUG=1 MIN_ASTS=1 python extra/optimization/get_action_space.py
- name: Repo line count < 25000 lines
run: MAX_LINE_COUNT=25000 python sz.py
- name: Repo line count <= 26000 lines
run: MAX_LINE_COUNT=26000 python sz.py
spec:
strategy:
@@ -294,7 +240,7 @@ jobs:
deps: testing_unit
llvm: 'true'
- name: Test SPEC=2
run: SPEC=2 pytest --maxfail=10 -n auto --durations=30 test/unit test/backend test/opt --ignore test/backend/test_custom_kernel.py --ignore test/unit/test_hashing.py --timeout 60 -k "not test_setitem_big" -k "not test_conv2d_ceildiv_edge_case" --splits 2 --group ${{ matrix.group }}
run: SPEC=2 pytest --maxfail=10 -n auto --durations=30 test/unit test/backend test/opt --ignore test/backend/test_custom_kernel.py --ignore test/unit/test_hashing.py -k "not test_setitem_big" -k "not test_conv2d_ceildiv_edge_case" --splits 2 --group ${{ matrix.group }}
fuzzing:
name: Fuzzing
@@ -308,14 +254,9 @@ jobs:
with:
key: fuzzing-unit
deps: testing_unit
- name: Fuzz Test symbolic
run: python test/external/fuzz_symbolic.py
- name: Fuzz Test symbolic (symbolic divisors)
run: python test/external/fuzz_symbolic_symbolic_div.py
- name: Fuzz Test fast idiv
run: python test/external/fuzz_fast_idiv.py
- name: Fuzz Test shape ops
run: python test/external/fuzz_shape_ops.py
- name: Fuzz Tests
run: |
parallel --tagstring '[{}]' 'python test/external/fuzz_{}.py' ::: symbolic symbolic_div fast_idiv shape_ops
testopenclimage:
name: CL IMAGE Tests
@@ -337,31 +278,6 @@ jobs:
- name: Run process replay tests
uses: ./.github/actions/process-replay
testgpumisc:
name: CL Misc tests
runs-on: *linux
timeout-minutes: 10
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: gen-dataset
deps: testing
opencl: 'true'
- name: Generate Dataset
run: DEV=CL extra/optimization/generate_dataset.sh
- name: Run Kernel Count Test
run: DEV=CL python -m pytest -n=auto test/external/external_test_opt.py
- name: Run fused optimizer tests
run: DEV=CL FUSE_OPTIM=1 python -m pytest -n=auto test/models/test_mnist.py test/backend/test_optim.py -k "not muon"
- name: Upload artifact
uses: actions/upload-artifact@v7
with:
name: sops.gz
path: /tmp/sops.gz
testopenpilot:
name: openpilot Compile Tests
runs-on: *linux
@@ -378,7 +294,8 @@ jobs:
llvm: 'true'
- name: Test openpilot model kernel count and gate usage
run: |
ALLOWED_KERNEL_COUNT=123 ALLOWED_READ_IMAGE=1468 ALLOWED_GATED_READ_IMAGE=18 FLOAT16=1 DEV=CL IMAGE=1 python examples/openpilot/compile3.py https://gitlab.com/commaai/openpilot-lfs.git/gitlab-lfs/objects/cf6376aa9a090f0da26c280ef69eabf9bbdd51d1faac9ed392919c3db69be916
ALLOWED_KERNEL_COUNT=123 ALLOWED_READ_IMAGE=1361 ALLOWED_GATED_READ_IMAGE=38 FLOAT16=1 DEV="CL::IMAGE_PITCH_ALIGNMENT=64" IMAGE=1 python examples/openpilot/compile3.py https://gitlab.com/commaai/openpilot-lfs.git/gitlab-lfs/objects/cf6376aa9a090f0da26c280ef69eabf9bbdd51d1faac9ed392919c3db69be916
# IMAGE_PITCH_ALIGNMENT=64 matches adreno 630
- name: Test openpilot CL compile fp32 (test correctness)
run: |
DEV=CL IMAGE=1 SELFTEST=1 python examples/openpilot/compile3.py https://github.com/haraschax/filedump/raw/refs/heads/master/driving_vision_fp32.onnx
@@ -421,7 +338,6 @@ jobs:
with:
key: optim
deps: testing
pydeps: "tensorflow==2.19"
opencl: 'true'
#- name: Test Optimization Helpers
# run: DEBUG=1 python3 extra/optimization/test_helpers.py
@@ -430,13 +346,15 @@ jobs:
- name: Test Beam Search
run: DEV=CL IGNORE_BEAM_CACHE=1 python3 -m pytest extra/optimization/test_beam_search.py
- name: Test MLPerf stuff
run: DEV=CL python -m pytest -n=auto test/external/external_test_optim.py test/external/external_test_losses.py test/external/external_test_metrics.py test/external/external_test_datasets.py --durations=20
run: DEV=CL python -m pytest -n=auto test/external/external_test_lr_schedule.py test/external/external_test_losses.py test/external/external_test_metrics.py test/external/external_test_datasets.py --durations=20
- name: DEV=NULL beautiful_mnist_multigpu
run: DEV=NULL NULL_ALLOW_COPYOUT=1 python examples/beautiful_mnist_multigpu.py
- name: Test Bert training
run: DEV=NULL NULL_ALLOW_COPYOUT=1 DEFAULT_FLOAT=HALF BENCHMARK=10 BS=24 GPUS=4 BERT_LAYERS=2 MODEL=bert python3 examples/mlperf/model_train.py
- name: Test llama 3 training
run: DEV=NULL NULL_ALLOW_COPYOUT=1 SAMPLES=300 BS=8 SEQLEN=512 GRADIENT_ACC_STEPS=1 FAKEDATA=1 DEFAULT_FLOAT=bfloat16 OPTIM_DTYPE=bfloat16 LLAMA3_SIZE=1B MODEL=llama3 python3 examples/mlperf/model_train.py
- name: Test gpt-oss training
run: DEV=NULL NULL_ALLOW_COPYOUT=1 SAMPLES=32 BS=2 SEQLEN=128 GRADIENT_ACC_STEPS=1 FAKEDATA=1 DEFAULT_FLOAT=bfloat16 OPTIM_DTYPE=bfloat16 MXFP8=1 VOCAB_SIZE=32000 LAYERS=2 EXPERTS=4 MODEL=gptoss PYTHONPATH=. python3 examples/mlperf/model_train.py
- name: Run process replay tests
uses: ./.github/actions/process-replay
@@ -468,7 +386,7 @@ jobs:
# ****** Models Tests ******
testmodels:
name: Models (llvm+cpu+gpu)
name: Models
runs-on: *linux
timeout-minutes: 15
steps:
@@ -479,34 +397,12 @@ jobs:
with:
key: models
deps: testing
opencl: 'true'
llvm: 'true'
- name: Test models (llvm)
run: DEV=CPU:LLVM python -m pytest -n=auto test/models --durations=20
- name: Test models (opencl)
run: DEV=CL python -m pytest -n=auto test/models --durations=20
- name: Test models (cpu)
run: DEV=CPU python -m pytest -n=auto test/models --durations=20
- name: Run process replay tests
uses: ./.github/actions/process-replay
testmetalmodels:
name: Models (metal)
runs-on: &macos macos-26
timeout-minutes: 20
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: metal
deps: testing
- name: Test models (Metal)
run: DEV=METAL python -m pytest -n=auto test/models --durations=20
- name: Test LLaMA compile speed
run: DEV=METAL python test/external/external_test_speed_llama.py
# ****** Feature Tests ******
testdsp:
@@ -548,9 +444,8 @@ jobs:
uses: ./.github/actions/setup-tinygrad
with:
key: linux-${{ matrix.dev }}
deps: testing_unit
deps: "testing_unit${{ contains(matrix.dev, 'LVP') && ' mesa' || '' }}"
llvm: ${{ contains(matrix.dev, 'LLVM') || contains(matrix.dev, 'LVP') || contains(matrix.dev, 'CLANG') }}
mesa: ${{ contains(matrix.dev, 'LVP') && 'cpu' || 'false' }}
webgpu: ${{ matrix.dev == 'WEBGPU' }}
opencl: ${{ matrix.dev == 'CL' }}
- name: Set env
@@ -588,7 +483,7 @@ jobs:
wget -qO- https://apt.llvm.org/llvm-snapshot.gpg.key | sudo tee /etc/apt/trusted.gpg.d/apt.llvm.org.asc
echo "deb http://apt.llvm.org/$(lsb_release -cs)/ llvm-toolchain-$(lsb_release -cs)-21 main" | sudo tee /etc/apt/sources.list.d/llvm.list
sudo apt-get update
sudo apt-get install llvm-21 llvm-21-tools cloc
sudo apt-get install -y llvm-21 llvm-21-tools cloc
- name: Install rocprof-trace-decoder
run: sudo PYTHONPATH="." ./extra/sqtt/install_rocprof_decoder.py
- name: Run AMD renderer tests
@@ -602,7 +497,7 @@ jobs:
AMD: 0
run: |
PYTHONPATH=. DEV=NULL:HIP:gfx1100 python extra/mmapeak/mmapeak.py
PYTHONPATH=. DEV=NULL:HIP:gfx950 python3 -m pytest -n=auto test/testextra/test_tk.py test/backend/test_asm_gemm.py
PYTHONPATH=. DEV=NULL:HIP:gfx950 python3 -m pytest -n=auto test/testextra/test_tk.py
- name: Run matmul on MOCKKFD
run: |
PYTHONPATH="." DEV=MOCKKFD+AMD N=256 python3 extra/gemm/amd_asm_matmul.py
@@ -610,6 +505,29 @@ jobs:
- name: Run LLVM test
run: DEV=MOCKKFD+AMD:LLVM python test/device/test_amd_llvm.py
hcq2:
name: hcq2
runs-on: *linux
timeout-minutes: 5
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: hcq2
deps: testing_unit
amd: 'true'
- name: Run HCQ2 tests
run: HCQ_RUNTIME_DEV=PYTHON HCQ2=1 DEV=MOCKKFD+AMD FORWARD_ONLY=1 PYTHONPATH=. python test/test_tiny.py
- name: Run HCQ2 multi-device tests
run: |
HCQ_RUNTIME_DEV=PYTHON HCQ2=1 DEV=MOCKKFD+AMD FORWARD_ONLY=1 PYTHONPATH=. python test/unit/test_multitensor.py \
TestMultiTensor.test_simple_add TestMultiTensor.test_shard_reduce \
TestMultiTensor.test_backward_sum TestMultiTensor.test_matmul_shard_0_0
- name: Run HCQ2 JIT tests
run: HCQ_RUNTIME_DEV=PYTHON HCQ2=1 DEV=MOCKKFD+AMD FORWARD_ONLY=1 PYTHONPATH=. python test/unit/test_jit.py
testmockam:
name: Linux (am)
runs-on: *linux
@@ -662,13 +580,16 @@ jobs:
key: ${{ matrix.backend }}-minimal
deps: testing_unit
amd: 'true'
llvm: ${{ matrix.backend == 'amdllvm' && 'true' }}
llvm: 'true'
- name: Check Device.DEFAULT and print some source
run: |
python3 -c "from tinygrad import Device; assert Device.DEFAULT in ['AMD'], Device.DEFAULT"
DEBUG=5 FORWARD_ONLY=1 python3 test/test_tiny.py TestTiny.test_plus
- name: Run MXFP4 Llama training on NULL backend
if: ${{ matrix.backend == 'amd' && matrix.arch == 'gfx950' }}
run: PYTHONPATH=. DEV=NULL:HIP:gfx950 MXFP4=1 LLAMA_LAYERS=2 BENCHMARK=3 NULL_ALLOW_COPYOUT=1 NO_HIPCC=1 ROCM_PATH=/opt/rocm JITBEAM=0 examples/mlperf/training_submission_v6.0/tinycorp/benchmarks/llama31_8b/implementations/tinybox_8xMI350X/profile.sh
- name: Run pytest (amd)
run: python -m pytest -n=auto test/backend/test_ops.py test/backend/test_dtype.py test/backend/test_dtype_alu.py test/backend/test_linearizer.py test/backend/test_randomness.py test/backend/test_jit.py test/backend/test_graph.py test/backend/test_multitensor.py test/device/test_hcq.py test/external/external_test_am.py test/backend/test_asm_gemm.py::TestAsmGEMM --durations=20
run: python -m pytest -n=auto test/backend/test_ops.py test/backend/test_dtype.py test/backend/test_dtype_alu.py test/backend/test_linearizer.py test/backend/test_randomness.py test/backend/test_jit.py test/backend/test_graph.py test/backend/test_multitensor.py test/device/test_hcq.py test/external/external_test_am.py test/backend/test_asm_gemm.py::TestAsmGEMM test/opt/test_tensor_cores.py --durations=20
- name: Run disk copy tests
run: python -m pytest test/unit/test_disk_tensor.py -k test_copy_from_disk
- name: Run TRANSCENDENTAL math
@@ -711,125 +632,6 @@ jobs:
- name: Run process replay tests
uses: ./.github/actions/process-replay
# ****** OSX Tests ******
unittestmacos:
name: MacOS (unit)
runs-on: *macos
timeout-minutes: 20
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: unittest-macos
deps: testing_unit
amd: 'true'
ocelot: 'true'
- name: Run unit tests
run: DEV=METAL python -m pytest -n=auto test/unit/ --durations=20
- name: Run NULL backend tests
run: DEV=NULL python -m pytest -n=auto test/null/ --durations=20
- name: Test tensor core ops (fake)
run: DEV=METAL DEBUG=3 TC=2 python test/backend/test_ops.py TestOps.test_gemm
- name: Test tensor core ops (real)
run: DEV=METAL DEBUG=3 python test/backend/test_ops.py TestOps.test_big_gemm
- name: Test Beam Search
run: DEV=METAL IGNORE_BEAM_CACHE=1 python3 -m pytest extra/optimization/test_beam_search.py
- name: Test Device Specific
run: DEV=METAL python3 -m pytest test/device/test_metal.py
#- name: Fuzz Test linearizer
# run: DEV=METAL DEPTH=4 FUZZ_N=50 FUZZ_MAX_SIZE=1000000 python test/external/fuzz_linearizer.py
- name: Run pytest (amd)
env:
DEV: MOCKKFD+AMD
FORWARD_ONLY: 1
run: |
python3 -m pytest -n=auto test/device/test_hcq.py test/test_tiny.py --durations=20
- name: Run pytest (ptx)
env:
DEV: "MOCK+NV:PTX"
FORWARD_ONLY: 1
# TODO: failing due to library loading error
CAPTURE_PROCESS_REPLAY: 0
run: |
python3 -m pytest -n=auto test/device/test_hcq.py test/test_tiny.py --durations=20
- name: Run process replay tests
uses: ./.github/actions/process-replay
testmacos:
strategy:
fail-fast: false
matrix:
dev:
- 'CPU:CLANG'
- 'CPU:LLVM'
- 'CPU:LVP'
- 'METAL'
- 'WEBGPU'
name: MacOS (DEV=${{ matrix.dev }})
runs-on: *macos
timeout-minutes: 20
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: macos-${{ matrix.dev }}
deps: testing_unit
llvm: ${{ contains(matrix.dev, 'LLVM') || contains(matrix.dev, 'LVP') }}
mesa: ${{ contains(matrix.dev, 'LVP') && 'cpu' || 'false' }}
webgpu: ${{ matrix.dev == 'WEBGPU' }}
- name: Set env
run: printf "DEV=${{ matrix.dev }}${{ matrix.dev == 'CPU:CLANG' && '\nCPU_COUNT=2' || '' }}" >> $GITHUB_ENV
- name: Check Device.DEFAULT and print some source
run: |
python -c "from tinygrad import Device; from tinygrad.helpers import Target; assert Device.DEFAULT == Target.parse('${{ matrix.dev }}').device"
DEBUG=4 python test/test_tiny.py TestTiny.test_plus
- name: Run backend tests
run: python -m pytest -n=auto test/backend --durations=20
- name: Run process replay tests
uses: ./.github/actions/process-replay
# ****** Windows Tests ******
testwindows:
strategy:
fail-fast: false
matrix:
dev:
- 'CPU:CLANG'
- 'CPU:LLVM'
- 'CPU:X86'
- 'WEBGPU'
name: Windows (DEV=${{ matrix.dev }})
runs-on: windows-2025
timeout-minutes: 15
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: windows-${{ matrix.dev }}-minimal
deps: testing_unit
pydeps: ${{ matrix.dev == 'WEBGPU' && 'dawn-python' || '' }}
- name: Set env
shell: bash
run: printf "DEV=${{ matrix.dev }}${{ matrix.dev == 'CPU:CLANG' && '\nCPU_COUNT=2' || '' }}" >> $GITHUB_ENV
- name: Check Device.DEFAULT and print some source
shell: bash
run: |
python -c "from tinygrad import Device; from tinygrad.helpers import Target; assert Device.DEFAULT == Target.parse('${{ matrix.dev }}').device"
DEBUG=4 python test/test_tiny.py TestTiny.test_plus
- name: Run test_tiny
shell: bash
run: python -m pytest -n=auto test/test_tiny.py --durations=20
# ****** Compile-only Tests ******
compiletests:
@@ -847,8 +649,7 @@ jobs:
uses: ./.github/actions/setup-tinygrad
with:
key: compile-${{ matrix.backend }}
deps: testing_unit
mesa: ${{ (matrix.backend == 'ir3' || matrix.backend == 'nak') && 'true' }}
deps: "testing_unit mesa"
- name: Set env
shell: bash
run: printf "NULL_ALLOW_COPYOUT=1\n${{ matrix.backend == 'ir3' && 'DEV=NULL:IR3:a630' || matrix.backend == 'nak' && 'DEV=NULL:NAK:sm_120' }}" >> $GITHUB_ENV
@@ -867,33 +668,3 @@ jobs:
run: |
DEBUG=4 python3 test/backend/test_ops.py TestOps.test_gemm | grep image_load
python -m pytest -n=auto test/backend/test_ops.py --durations=20
qcomclcompiletests:
name: Compile-only (QCOM CL)
runs-on: ubuntu-24.04-arm
timeout-minutes: 15
steps:
- name: Checkout Code
uses: actions/checkout@v6
- name: Setup Environment
uses: ./.github/actions/setup-tinygrad
with:
key: compile-qcomcl
deps: testing_unit
tinydreno: 'true'
- name: Set env
shell: bash
run: printf "DEV=NULL:QCOMCL:a630\nNULL_ALLOW_COPYOUT=1" >> $GITHUB_ENV
- name: Run test_ops
shell: bash
run: |
python -c "from tinygrad import Device; assert Device.DEFAULT == 'NULL'"
DEBUG=4 python3 test/backend/test_ops.py TestOps.test_add
python -m pytest -n=auto test/backend/test_ops.py --durations=20
- name: Run test_ops (IMAGE)
shell: bash
env:
IMAGE: 1
DEV: "NULL:QCOMCL:a630,IMAGE_PITCH_ALIGNMENT=64"
run: |
DEBUG=4 python test/backend/test_ops.py TestOps.test_gemm | grep read_imagef
python -m pytest -n=auto test/backend/test_ops.py --durations=20