Use to test llm-performance on spr-perf (#9316)

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

* Update action.yml

* Create cpu-perf-test.yaml

* Update action.yml

* Update action.yml

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

* Update llm_performance_tests.yml

.github/workflows/llm_performance_tests.yml

@@ -137,4 +137,51 @@ jobs:
           curl -T ./*.csv ${LLM_FTP_URL}/llm/ggml-actions/perf/
           cp ./*.csv /mnt/disk1/nightly_perf/
           cd ../../../test/benchmark
-          python csv_to_html.py -f /mnt/disk1/nightly_perf/
+          python csv_to_html.py -f ../../dev/benchmark/all-in-one
+          cp ./*.html /mnt/disk1/nightly_perf/
+          
+  llm-performance-test-on-spr:
+    needs: llm-cpp-build
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.9"]
+    runs-on: [self-hosted, llm, spr-perf]
+    env:
+      OMP_NUM_THREADS: 16
+      THREAD_NUM: 16
+      ANALYTICS_ZOO_ROOT: ${{ github.workspace }}
+    steps:
+      - uses: actions/checkout@v3
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v4
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Install dependencies
+        shell: bash
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install --upgrade wheel
+          python -m pip install --upgrade omegaconf
+          python -m pip install --upgrade pandas
+          python -m pip install --upgrade einops
+
+      - name: Download llm binary
+        uses: ./.github/actions/llm/download-llm-binary
+
+      - name: Run LLM install (all) test
+        uses: ./.github/actions/llm/setup-llm-env
+
+      - name: Test on cpu
+        shell: bash
+        run: |
+          mv python/llm/test/benchmark/cpu-perf-test.yaml python/llm/dev/benchmark/all-in-one/config.yaml
+          cd python/llm/dev/benchmark/all-in-one
+          export http_proxy=${HTTP_PROXY}
+          export https_proxy=${HTTPS_PROXY}
+          python run.py
+          cp ./*.csv /mnt/disk1/nightly_perf_cpu/
+          cd ../../../test/benchmark
+          python csv_to_html.py -f /mnt/disk1/nightly_perf_cpu/

python/llm/test/benchmark/cpu-perf-test.yaml

@@ -0,0 +1,18 @@
+repo_id:
+  - 'meta-llama/Llama-2-7b-chat-hf'
+local_model_hub: '/mnt/disk1/models'
+warm_up: 1
+num_trials: 3
+num_beams: 1 # default to greedy search
+low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
+in_out_pairs:
+  - '32-32'
+  - '1024-128'
+test_api:
+  - "transformer_int4"
+  # - "native_int4"
+  # - "optimize_model"
+  # - "pytorch_autocast_bf16"
+  # - "ipex_fp16_gpu" # on Intel GPU
+  # - "transformer_int4_gpu"  # on Intel GPU
+  # - "optimize_model_gpu"  # on Intel GPU