Direct Driver HAL (#816)

.github/CODEOWNERS

@@ -15,8 +15,7 @@
 /compiler/ @MaheshRavishankar @nirvedhmeshram @yzhang93 @Abhishek-Varma @jtuyls
 
 # Runtime
-/runtime/ @nirvedhmeshram
-/runtime/src/iree-amd-aie/aie_runtime @makslevental
+/runtime/ @makslevental
 
 # AIE Passes
 /compiler/plugins/target/AMD-AIE/aie @makslevental

.github/workflows/ci-linux.yml

@@ -49,13 +49,10 @@ jobs:
           git remote add origin $REPO_ADDRESS
           git -c protocol.version=2 fetch --depth 1 origin $BRANCH_NAME
           git reset --hard FETCH_HEAD
-          git -c submodule."third_party/torch-mlir".update=none -c submodule."third_party/stablehlo".update=none -c submodule."src/runtime_src/core/common/aiebu".update=none submodule update --init --recursive --depth 1 --single-branch -j 10
-          
-      - name: Install deps
-        run: |
-          dnf install -y almalinux-release-devel epel-release
-          yum remove -y openssl-devel zlib-devel || true
-          yum install -y protobuf-devel protobuf-compiler tmate
+          git -c submodule."third_party/torch-mlir".update=none \
+              -c submodule."third_party/stablehlo".update=none \
+              -c submodule."third_party/XRT".update=none \
+              submodule update --init --recursive --depth 1 --single-branch -j 10
 
       - name: Python deps
         run: |
@@ -69,6 +66,11 @@ jobs:
           key:  ${{ env.CACHE_KEY }}
           restore-keys: linux-build-test-cpp-
 
+      - name: Peano dep
+        run: |
+          bash build_tools/download_peano.sh
+          echo "PEANO_INSTALL_DIR=$PWD/llvm-aie" >> $GITHUB_ENV
+
       - name: Build packages
         run: |
           export cache_dir="${{ env.CACHE_DIR }}"
@@ -147,60 +149,54 @@ jobs:
           source .venv/bin/activate
           pip install -r tests/requirements.txt
 
+      - name: Query device info
+        run: |
+          source .venv/bin/activate
+          echo "aie-metadata"
+          python build_tools/ci/amdxdna_driver_utils/amdxdna_ioctl.py --aie-metadata
+          echo "aie-version"
+          python build_tools/ci/amdxdna_driver_utils/amdxdna_ioctl.py --aie-version
+          echo "XRT_LITE_N_CORE_ROWS=$(python build_tools/ci/amdxdna_driver_utils/amdxdna_ioctl.py --num-rows)" >> $GITHUB_ENV
+          echo "XRT_LITE_N_CORE_COLS=$(python build_tools/ci/amdxdna_driver_utils/amdxdna_ioctl.py --num-cols)" >> $GITHUB_ENV
+
       - name : E2E comparison of AIE to llvm-cpu
         run: |
           source .venv/bin/activate
-          source /opt/xilinx/xrt/setup.sh
           python build_tools/ci/cpu_comparison/run.py \
             test_aie_vs_cpu \
             $PWD/iree-install \
             $PWD/llvm-aie \
-            --xrt-dir /opt/xilinx/xrt \
             --vitis-dir /opt/Xilinx/Vitis/2024.2 \
-            --reset-npu-between-runs -v
+            --reset-npu-between-runs -v \
+            --xrt_lite_n_core_rows=$XRT_LITE_N_CORE_ROWS \
+            --xrt_lite_n_core_cols=$XRT_LITE_N_CORE_COLS
 
       - name: E2E correctness matmul test
         run: |
-          # Without this additional line an error like
-          #
-          # [XRT] ERROR: Failed to allocate host memory buffer (mmap(len=10616832, prot=3, flags=8193, offset=4294967296)
-          # failed (err=11): Resource temporarily unavailable), make sure host bank is enabled (see xbutil configure --host-mem)
-          # iree-amd-aie/runtime/src/iree-amd-aie/driver/xrt/direct_allocator.cc:179: RESOURCE_EXHAUSTED; could not allocate
-          # memory for buffer; while invoking C++ function matmul_test.generate_random_matrix; while calling import;
-          #
-          # might be observed when too much memory is allocated. This
-          # error was seen when running a bf16->f32 matmul with m=n=k=2304.
-          #
-          # This line was suggested at https://github.com/Xilinx/mlir-air/issues/566
-          #
-          # Note that this is only half of the fix. It is also necessary that
-          # the machine that CI is running on has permission to run this line.
-          #
-          # This permission can be adding by adding the line
-          # ```
-          # %github ALL=(ALL) NOPASSWD: /usr/bin/prlimit *
-          # ```
-          #
-          # to the file /etc/sudoers.d/github, which can be done by running
-          # ```
-          # sudo visudo -f /etc/sudoers.d/github
-          # ```
-          # on the github CI machine.
+          # https://stackoverflow.com/a/17567422
+          # shim_xdna::bo::map_drm_bo does an mmap with MAP_LOCKED
+          # which can fail if limit is to low
           sudo prlimit -lunlimited --pid $$
           source .venv/bin/activate
-          source /opt/xilinx/xrt/setup.sh
           bash build_tools/ci/run_matmul_test.sh \
             test_matmuls \
             iree-install \
             $PWD/llvm-aie \
-            /opt/xilinx/xrt \
             /opt/Xilinx/Vitis/2024.2
 
       - name: Python tests
         run: |
           source .venv/bin/activate
-          source /opt/xilinx/xrt/setup.sh
           pytest -v tests \
             --capture=tee-sys \
             --iree-install-dir=$PWD/iree-install \
-            --peano-install-dir=$PWD/llvm-aie
+            --peano-install-dir=$PWD/llvm-aie \
+            --xrt_lite_n_core_rows=$XRT_LITE_N_CORE_ROWS \
+            --xrt_lite_n_core_cols=$XRT_LITE_N_CORE_COLS
+
+      - name: XRT-LITE tests
+        run: |
+          DEVICE_TEST_DIR="$PWD/iree-install/device_tests"
+          for t in $(ls $DEVICE_TEST_DIR); do
+            $DEVICE_TEST_DIR/$t --xrt_lite_n_core_rows=$XRT_LITE_N_CORE_ROWS --xrt_lite_n_core_cols=$XRT_LITE_N_CORE_COLS
+          done

.github/workflows/ci-macos.yml

@@ -62,7 +62,10 @@ jobs:
           git remote add origin $REPO_ADDRESS
           git -c protocol.version=2 fetch --depth 1 origin $BRANCH_NAME
           git reset --hard FETCH_HEAD
-          git -c submodule."third_party/torch-mlir".update=none -c submodule."third_party/stablehlo".update=none -c submodule."src/runtime_src/core/common/aiebu".update=none submodule update --init --recursive --depth 1 --single-branch -j 10
+          git -c submodule."third_party/torch-mlir".update=none \
+              -c submodule."third_party/stablehlo".update=none \
+              -c submodule."third_party/XRT".update=none \
+              submodule update --init --recursive --depth 1 --single-branch -j 10
 
       - uses: actions/setup-python@v4
         with:

.github/workflows/ci-windows.yml

@@ -60,7 +60,10 @@ jobs:
           git remote add origin $REPO_ADDRESS
           git -c protocol.version=2 fetch --depth 1 origin $BRANCH_NAME
           git reset --hard FETCH_HEAD
-          git -c submodule."third_party/torch-mlir".update=none -c submodule."third_party/stablehlo".update=none -c submodule."src/runtime_src/core/common/aiebu".update=none submodule update --init --recursive --depth 1 --single-branch -j 10
+          git -c submodule."third_party/torch-mlir".update=none \
+              -c submodule."third_party/stablehlo".update=none \
+              -c submodule."src/runtime_src/core/common/aiebu".update=none \
+              submodule update --init --recursive --depth 1 --single-branch -j 10
 
       - name: Setup Cpp
         uses: aminya/setup-cpp@v1
@@ -87,14 +90,18 @@ jobs:
           key: ${{ env.CACHE_KEY }}
           restore-keys: windows-build-test-cpp-
 
+      - name: Peano dep
+        run: |
+          .\build_tools\download_peano.ps1
+          Add-Content -Path $env:GITHUB_ENV -Value "PEANO_INSTALL_DIR=$PWD\llvm-aie"
+
       - name: Build packages
         run: |
           $env:cache_dir = "${{ env.CACHE_DIR }}"
           $env:CCACHE_COMPILERCHECK = "string:$(clang-cl.exe --version)"
           .\build_tools\build_llvm.ps1
           # Remove-Item -Path "$pwd\llvm-build" -Force
           $env:llvm_install_dir = "$pwd\llvm-install"
-          echo $env:llvm_install_dir
           .\build_tools.\build_test_cpp.ps1
 
       - name: Create artifacts
@@ -170,6 +177,7 @@ jobs:
         shell: bash
         run: |
           source .venv/Scripts/activate
+          export DEVICE_HAL=xrt
           bash build_tools/ci/run_matmul_test.sh \
             /c/test_matmuls \
             $PWD/iree-install \
@@ -182,7 +190,8 @@ jobs:
           python build_tools/ci/cpu_comparison/run.py \
             /c/test_aie_vs_cpu \
             $PWD/iree-install \
-            $PWD/llvm-aie -v
+            $PWD/llvm-aie -v \
+            --device-hal=xrt
 
       - name: Python tests
         run: |
@@ -191,5 +200,6 @@ jobs:
           mkdir temp
           pytest tests -sv `
             --basetemp=$PWD\temp `
-            --iree-install-dir="$PWD/iree-install" `
-            --peano-install-dir="$PWD/llvm-aie"
+            --iree-install-dir="$PWD\iree-install" `
+            --peano-install-dir="$PWD\llvm-aie" `
+            --device-hal=xrt