Merge remote-tracking branch 'upstream/main' into users/scotttodd/req…

…uirements-matrix
nod-ai · Jan 10, 2025 · d4f56bf · d4f56bf
2 parents fa7ab97 + 2edde8d
commit d4f56bf
Show file tree

Hide file tree

Showing 70 changed files with 2,211 additions and 4,317 deletions.
diff --git a/.github/workflows/ci-libshortfin.yml b/.github/workflows/ci-libshortfin.yml
@@ -131,7 +131,7 @@ jobs:
       working-directory: ${{ env.LIBSHORTFIN_DIR }}
       run: |
         ctest --timeout 30 --output-on-failure --test-dir build
-        pytest -s --durations=10
+        pytest -s --durations=10 --timeout=30
 
   # Depends on all other jobs to provide an aggregate job status.
   ci_libshortfin_summary:

diff --git a/.github/workflows/ci-llama-quick-tests.yaml b/.github/workflows/ci-llama-quick-tests.yaml
@@ -55,18 +55,13 @@ jobs:
           python -m pip install --no-compile --upgrade pip
 
           # Note: We install in three steps in order to satisfy requirements
-          # from non default locations first. Installing the PyTorch CPU
-          # wheels saves multiple minutes and a lot of bandwidth on runner setup.
+          # from non default locations first.
           pip install --no-compile -r pytorch-cpu-requirements.txt
+          pip install -r requirements-iree-pinned.txt
+          pip install --no-compile \
+            -r sharktank/requirements-tests.txt \
+            -e sharktank/
 
-          # Install nightly IREE packages.
-          # We could also pin to a known working or stable version.
-          pip install -f https://iree.dev/pip-release-links.html --pre --upgrade \
-            iree-base-compiler \
-            iree-base-runtime \
-            iree-turbine
-
-          pip install --no-compile -r requirements.txt -r sharktank/requirements-tests.txt -e sharktank/
           pip freeze
 
       - name: Run llama 8b f16 decomposed test

diff --git a/.github/workflows/ci-sglang-benchmark.yml b/.github/workflows/ci-sglang-benchmark.yml
@@ -69,7 +69,7 @@ jobs:
           pip install --no-compile -r requirements.txt -e sharktank/ shortfin/
 
           # Pin to known-working versions.
-          pip install -f https://iree.dev/pip-release-links.html --pre --upgrade \
+          pip install -f https://iree.dev/pip-release-links.html \
             iree-base-compiler==3.1.0rc20241220 \
             iree-base-runtime==3.1.0rc20241220 \
             "numpy<2.0"

diff --git a/.github/workflows/ci-shark-ai.yml b/.github/workflows/ci-shark-ai.yml
diff --git a/.github/workflows/ci-sharktank.yml b/.github/workflows/ci-sharktank.yml
@@ -72,18 +72,15 @@ jobs:
           python -m pip install --no-compile --upgrade pip
 
           # Note: We install in three steps in order to satisfy requirements
-          # from non default locations first. Installing the PyTorch CPU
-          # wheels saves multiple minutes and a lot of bandwidth on runner setup.
-          pip install --no-compile --index-url https://download.pytorch.org/whl/cpu torch==${{matrix.torch-version}}+cpu
+          # from non default locations first.
+          pip install --no-compile \
+            --index-url https://download.pytorch.org/whl/cpu torch==${{matrix.torch-version}}+cpu
+          pip install -r requirements-iree-pinned.txt
+          pip install --no-compile \
+            -r sharktank/requirements-tests.txt \
+            -e sharktank/
 
-          # Install nightly IREE packages.
-          # We could also pin to a known working or stable version.
-          pip install -f https://iree.dev/pip-release-links.html --pre \
-            iree-base-compiler \
-            iree-base-runtime \
-            iree-turbine
-
-          pip install --no-compile -r requirements.txt -r sharktank/requirements-tests.txt -e sharktank/
+          pip freeze
 
       - name: Run sharktank tests
         if: ${{ !cancelled() }}
@@ -121,19 +118,16 @@ jobs:
         run: |
           source ${VENV_DIR}/bin/activate
           python -m pip install --no-compile --upgrade pip
+
           # Note: We install in three steps in order to satisfy requirements
-          # from non default locations first. Installing the PyTorch CPU
-          # wheels saves multiple minutes and a lot of bandwidth on runner setup.
+          # from non default locations first.
           pip install --no-compile -r pytorch-cpu-requirements.txt
+          pip install -r requirements-iree-pinned.txt
+          pip install --no-compile \
+            -r sharktank/requirements-tests.txt \
+            -e sharktank/
 
-          # Install nightly IREE packages.
-          # We could also pin to a known working or stable version.
-          pip install -f https://iree.dev/pip-release-links.html --pre --upgrade \
-            iree-base-compiler \
-            iree-base-runtime \
-            iree-turbine
-
-          pip install --no-compile -r requirements.txt -r sharktank/requirements-tests.txt -e sharktank/
+          pip freeze
 
       - name: Run tests
       # TODO: unify with-*-data flags into a single flag and make it possible to run
@@ -180,18 +174,14 @@ jobs:
           python -m pip install --no-compile --upgrade pip
 
           # Note: We install in three steps in order to satisfy requirements
-          # from non default locations first. Installing the PyTorch CPU
-          # wheels saves multiple minutes and a lot of bandwidth on runner setup.
+          # from non default locations first.
           pip install --no-compile -r pytorch-cpu-requirements.txt
+          pip install -r requirements-iree-pinned.txt
+          pip install --no-compile \
+            -r sharktank/requirements-tests.txt \
+            -e sharktank/
 
-          # Install nightly IREE packages.
-          # We could also pin to a known working or stable version.
-          pip install -f https://iree.dev/pip-release-links.html --pre \
-            iree-base-compiler \
-            iree-base-runtime \
-            iree-turbine
-
-          pip install --no-compile -r requirements.txt -r sharktank/requirements-tests.txt -e sharktank/
+          pip freeze
 
       - name: Run punet tests
         run: |

diff --git a/.github/workflows/ci_eval.yaml b/.github/workflows/ci_eval.yaml
@@ -63,7 +63,7 @@ jobs:
       - name: Run perplexity test with IREE
         run: |
           source ${VENV_DIR}/bin/activate
-          pytest -n 8 -v -s sharktank/tests/evaluate/perplexity_iree_test.py --run-nightly-llama-tests --bs=100 --iree-device=hip://0 --iree-hip-target=gfx942 --iree-hal-target-backends=rocm --llama3-8b-f16-model-path=/data/llama3.1/8b/llama8b_f16.irpa --llama3-8b-tokenizer-path=/data/llama3.1/8b/tokenizer_config.json --html=out/llm/llama/perplexity/iree_perplexity/index.html
+          pytest -n 8 -v -s sharktank/tests/evaluate/perplexity_iree_test.py --run-nightly-llama-tests --bs=100 --iree-device=hip://0 --iree-hip-target=gfx942 --iree-hal-target-device=hip --llama3-8b-f16-model-path=/data/llama3.1/8b/llama8b_f16.irpa --llama3-8b-tokenizer-path=/data/llama3.1/8b/tokenizer_config.json --html=out/llm/llama/perplexity/iree_perplexity/index.html
 
       - name: Deploy to GitHub Pages
         uses: peaceiris/actions-gh-pages@4f9cc6602d3f66b9c108549d475ec49e8ef4d45e # v4.0.0

diff --git a/.github/workflows/ci_eval_short.yaml b/.github/workflows/ci_eval_short.yaml
@@ -52,21 +52,16 @@ jobs:
           python -m pip install --no-compile --upgrade pip
 
           # Note: We install in three steps in order to satisfy requirements
-          # from non default locations first. Installing the PyTorch CPU
-          # wheels saves multiple minutes and a lot of bandwidth on runner setup.
+          # from non default locations first.
           pip install --no-compile -r pytorch-cpu-requirements.txt
+          pip install -r requirements-iree-pinned.txt
+          pip install --no-compile \
+            -r sharktank/requirements-tests.txt \
+            -e sharktank/
 
-          # Install nightly IREE packages.
-          # We could also pin to a known working or stable version.
-          pip install -f https://iree.dev/pip-release-links.html --pre --upgrade \
-            iree-base-compiler \
-            iree-base-runtime \
-            iree-turbine
-
-          pip install --no-compile -r requirements.txt -r sharktank/requirements-tests.txt -e sharktank/
           pip freeze
 
       - name: Run perplexity test with vmfb
         run: |
           source ${VENV_DIR}/bin/activate
-          pytest -n 8 -v -s sharktank/tests/evaluate/perplexity_iree_test.py --bs=5 --iree-device=hip://0 --iree-hip-target=gfx942 --iree-hal-target-backends=rocm --llama3-8b-f16-model-path=/data/llama3.1/8b/llama8b_f16.irpa --llama3-8b-tokenizer-path=/data/llama3.1/8b/tokenizer_config.json
+          pytest -n 8 -v -s sharktank/tests/evaluate/perplexity_iree_test.py --bs=5 --iree-device=hip://0 --iree-hip-target=gfx942 --iree-hal-target-device=hip --llama3-8b-f16-model-path=/data/llama3.1/8b/llama8b_f16.irpa --llama3-8b-tokenizer-path=/data/llama3.1/8b/tokenizer_config.json
diff --git a/.github/workflows/ci_linux_x64_asan-libshortfin.yml b/.github/workflows/ci_linux_x64_asan-libshortfin.yml
@@ -32,8 +32,8 @@ concurrency:
 
 env:
   PYENV_ROOT: ${{ github.workspace }}/pyenv
-  PYENV_REF: 96b3fb2fc3bee85650cb22e2cb06c83c24509a6d # v2.4.17
-  PYTHON_VER: 3.12.7
+  PYENV_REF: 2aff0df76deacd9b6c49de8870f0255d24d3b31b # v2.5.0
+  PYTHON_VER: 3.12.8
   CACHE_ASAN_VER: 2
   CACHE_DEPS_VER: 1
   LIBSHORTFIN_DIR: ${{ github.workspace }}/shortfin/
@@ -76,7 +76,7 @@ jobs:
       working-directory: ${{ env.PYENV_ROOT }}
       run: |
         src/configure && make -C src
-        export PATH=${{ env.PYENV_ROOT }}/bin:$PATH && eval "$(pyenv init -)"
+        export PATH=${{ env.PYENV_ROOT }}/bin:$PATH && eval "$(pyenv init - bash)"
         CC=clang-18 CXX=clang++-18 LDFLAGS="-lstdc++" PYTHON_CONFIGURE_OPTS="--with-address-sanitizer" pyenv install -v ${{ env.PYTHON_VER }}
         pyenv global ${{ env.PYTHON_VER }}
 
@@ -104,7 +104,7 @@ jobs:
       uses: actions/cache/restore@0c45773b623bea8c8e75f6c82b208c3cf94ea4f9 # v4.0.2
       with:
         path: ${{ env.PYENV_ROOT }}
-        key: ${{ runner.os }}-python-deps-${{ hashFiles('shortfin/requirements-tests.txt') }}-v${{ env.CACHE_DEPS_VER }}
+        key: ${{ runner.os }}-python-deps-${{ hashFiles('shortfin/requirements-tests.txt', 'requirements-iree-pinned.txt') }}-v${{ env.CACHE_DEPS_VER }}
 
     - name: Restore Python ASan cache
       id: cache-python-asan
@@ -122,7 +122,7 @@ jobs:
       if: steps.cache-python-deps-restore.outputs.cache-hit != 'true'
       working-directory: ${{ env.LIBSHORTFIN_DIR }}
       run: |
-        eval "$(pyenv init -)"
+        eval "$(pyenv init - bash)"
         pip install -r requirements-tests.txt
         pip install -r ../requirements-iree-pinned.txt
         pip freeze
@@ -138,7 +138,7 @@ jobs:
     - name: Build shortfin
       working-directory: ${{ env.LIBSHORTFIN_DIR }}
       run: |
-        eval "$(pyenv init -)"
+        eval "$(pyenv init - bash)"
         SHORTFIN_ENABLE_ASAN=ON \
         SHORTFIN_DEV_MODE=ON \
         SHORTFIN_RUN_CTESTS=ON \
@@ -151,5 +151,5 @@ jobs:
         ASAN_OPTIONS: detect_odr_violation=0
       working-directory: ${{ env.LIBSHORTFIN_DIR }}
       run: |
-        eval "$(pyenv init -)"
-        pytest -s
+        eval "$(pyenv init - bash)"
+        pytest -s --durations=10 --timeout=30
diff --git a/.github/workflows/pkgci.yml b/.github/workflows/pkgci.yml
@@ -0,0 +1,39 @@
+# Copyright 2024 Advanced Micro Devices, Inc.
+#
+# Licensed under the Apache License v2.0 with LLVM Exceptions.
+# See https://llvm.org/LICENSE.txt for license information.
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+
+name: PkgCI
+
+on:
+  workflow_dispatch:
+  pull_request:
+  push:
+    branches:
+      - main
+
+permissions:
+  contents: read
+
+concurrency:
+  # A PR number if a pull request and otherwise the commit hash. This cancels
+  # queued and in-progress runs for the same PR (presubmit) or commit
+  # (postsubmit). The workflow name is prepended to avoid conflicts between
+  # different workflows.
+  group: ${{ github.workflow }}-${{ github.event.number || github.sha }}
+  cancel-in-progress: true
+
+jobs:
+  build_packages:
+    name: Build Packages
+    uses: ./.github/workflows/build_packages.yml
+    permissions:
+      contents: write
+    with:
+      build_type: "dev"
+
+  test_shark_ai:
+    name: Test shark-ai
+    needs: [build_packages]
+    uses: ./.github/workflows/pkgci_shark_ai.yml