Merge branch 'main' into zhewen_bdchain

.github/workflows/black.yml

@@ -0,0 +1,12 @@
+name: Lint
+
+on: [push, pull_request]
+
+jobs:
+  lint:
+    runs-on: nod-ai-shared-cpubuilder-manylinux-x86_64
+    steps:
+      - uses: actions/checkout@v4
+      - uses: psf/black@stable
+
+# see https://black.readthedocs.io/en/stable/integrations/github_actions.html

.github/workflows/ci-linux.yml

@@ -160,59 +160,51 @@ jobs:
           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 correctness matmul test
+        run: |
+          # 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 $$
+          source .venv/bin/activate
+          bash build_tools/ci/run_matmul_test.sh \
+            test_matmuls \
+            iree-install \
+            $PWD/llvm-aie
+
+
       - name : E2E comparison of AIE to llvm-cpu
         run: |
+          sudo prlimit -lunlimited --pid $$
           source .venv/bin/activate
           python build_tools/ci/cpu_comparison/run.py \
             test_aie_vs_cpu \
             $PWD/iree-install \
-            $PWD/llvm-aie \
-            --vitis-dir /opt/Xilinx/Vitis/2024.2 \
-            --target_device "npu1_4col" \
-            --reset-npu-between-runs -v \
+            --peano_dir=$PWD/llvm-aie \
+            --vitis_dir=/opt/Xilinx/Vitis/2024.2 \
+            --target_device="npu1_4col" \
+            --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 : Performance benchmarks
-
         run: |
           source .venv/bin/activate
           python build_tools/ci/cpu_comparison/run.py \
             test_aie_vs_cpu \
             $PWD/iree-install \
-            $PWD/llvm-aie \
-            --vitis-dir /opt/Xilinx/Vitis/2024.2 \
-            --target_device "npu1_4col" \
-            --reset-npu-between-runs -v \
+            --peano_dir=$PWD/llvm-aie \
+            --vitis_dir=/opt/Xilinx/Vitis/2024.2 \
+            --target_device="npu1_4col" \
+            --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 \
             --tests=Performance > performance.log
 
           python build_tools/ci/cpu_comparison/performance_summarizer.py \
             performance.log
 
-      - name: E2E correctness matmul test
-        run: |
-          # 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
-          bash build_tools/ci/run_matmul_test.sh \
-            test_matmuls \
-            iree-install \
-            $PWD/llvm-aie \
-            /opt/Xilinx/Vitis/2024.2
-
-      - name: Python tests
-        run: |
-          source .venv/bin/activate
-          pytest -rv tests \
-            --capture=tee-sys \
-            --iree-install-dir=$PWD/iree-install \
-            --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: |
@@ -274,11 +266,9 @@ jobs:
           python build_tools/ci/cpu_comparison/run.py \
             test_aie_vs_cpu \
             $PWD/iree-install \
-            /opt/xilinx/Vitis/2024.2 \
-            --vitis-dir /opt/xilinx/Vitis/2024.2 \
-            --target_device "npu4" \
-            --use_chess "1" \
-            --reset-npu-between-runs \
+            --vitis_dir=/opt/xilinx/Vitis/2024.2 \
+            --target_device="npu4" \
+            --reset_npu_between_runs \
             --xrt_lite_n_core_rows=$XRT_LITE_N_CORE_ROWS \
             --xrt_lite_n_core_cols=$XRT_LITE_N_CORE_COLS \
             -v

.github/workflows/ci-windows.yml

@@ -190,17 +190,7 @@ jobs:
           python build_tools/ci/cpu_comparison/run.py \
             /c/test_aie_vs_cpu \
             $PWD/iree-install \
-            $PWD/llvm-aie -v \
+            --peano_dir=$PWD/llvm-aie -v \
             --target_device "npu1_4col" \
-            --device-hal=xrt
-
-      - name: Python tests
-        run: |
-          ls $env:XILINX_XRT
-          .\.venv\Scripts\Activate.ps1
-          mkdir temp
-          pytest tests -sv `
-            --basetemp=$PWD\temp `
-            --iree-install-dir="$PWD\iree-install" `
-            --peano-install-dir="$PWD\llvm-aie" `
-            --device-hal=xrt
+            --device_hal=xrt \
+            --skip_tests "Performance"

build_tools/build_test_cpp.sh

@@ -81,7 +81,6 @@ CMAKE_ARGS=(
   -DIREE_INPUT_TORCH=OFF
   -DCMAKE_OBJECT_PATH_MAX=4096
   -DIREE_CMAKE_PLUGIN_PATHS="$repo_root"
-  -DIREE_AMDAIE_TIME_KERNEL=1
 )
 
 PEANO_INSTALL_DIR=${PEANO_INSTALL_DIR:-""}

build_tools/ci/cpu_comparison/performance_summarizer.py

@@ -14,14 +14,18 @@
     with open(path, "r") as f:
         lines = f.readlines()
     print("============================")
-    first_print = True
     for line in lines:
-        if "Run #1" in line:
-            if not first_print:
-                print("\n" + line.split()[-1])
-            else:
-                print(line.split()[-1])
-            first_print = False
-        if "IREE_AMDAIE" in line:
+        if "Performance benchmark:" in line:
+            print("\n")
+            print(line.split()[-1])
+        if "----------------------" in line:
+            print(line.strip())
+        if "Benchmark" in line:
+            print(line.strip())
+        if "real_time_mean" in line:
+            print(line.strip())
+        if "real_time_median" in line:
+            print(line.strip())
+        if "real_time_stddev" in line:
             print(line.strip())
     print("============================")