Improve CI by trying a warmup before unit tests (#14669)

.github/workflows/pr-test.yml

@@ -457,11 +457,6 @@ jobs:
       - name: Checkout code
         uses: actions/checkout@v4
 
-      - name: Cleanup
-        run: |
-          ls -alh sgl-kernel/dist || true
-          rm -rf sgl-kernel/dist/* || true
-
       - name: Download artifacts
         if: needs.check-changes.outputs.sgl_kernel == 'true'
         uses: actions/download-artifact@v4
@@ -504,11 +499,6 @@ jobs:
       - name: Checkout code
         uses: actions/checkout@v4
 
-      - name: Cleanup
-        run: |
-          ls -alh sgl-kernel/dist || true
-          rm -rf sgl-kernel/dist/* || true
-
       - name: Download artifacts
         if: needs.check-changes.outputs.sgl_kernel == 'true'
         uses: actions/download-artifact@v4
@@ -557,8 +547,8 @@ jobs:
 
         - name: Install dependencies
           run: |
-            pip install "bitsandbytes>=0.44.0"
             CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+
         - name: Run test
           timeout-minutes: 30
           run: |
@@ -724,6 +714,12 @@ jobs:
         run: |
           CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
 
+      - name: Warmup Weights and JIT Compilation
+        timeout-minutes: 20
+        run: |
+          # An example command for testing the warmup. TODO: make this more general and move them to python scripts.
+          python3 -m sglang.compile_deep_gemm --model deepseek-ai/DeepSeek-V3-0324 --tp 8 --trust-remote-code
+
       - name: Run test
         timeout-minutes: 20
         run: |