diff --git a/.github/workflows/compatiblity_test_on_dispatch.yml b/.github/workflows/compatiblity_test_on_dispatch.yml
index 3dcc4dfd1..1778d64ee 100644
--- a/.github/workflows/compatiblity_test_on_dispatch.yml
+++ b/.github/workflows/compatiblity_test_on_dispatch.yml
@@ -72,7 +72,7 @@ jobs:
           ssh-key: ${{ secrets.SSH_KEY_FOR_CI }}
       - name: Download cub for CUDA 10.2
         run: |
-          CUDA_VERSION=$(cat $CUDA_HOME/version.txt | grep "CUDA Version" | awk '{print $NF}' | cut -d. -f1,2)
+          CUDA_VERSION=$(nvcc -V | awk -F ',| ' '/release/{print $6}')
 
           # check if it is CUDA 10.2
           # download cub
diff --git a/.github/workflows/compatiblity_test_on_pr.yml b/.github/workflows/compatiblity_test_on_pr.yml
index 5098b8e36..c0f45c65a 100644
--- a/.github/workflows/compatiblity_test_on_pr.yml
+++ b/.github/workflows/compatiblity_test_on_pr.yml
@@ -66,7 +66,7 @@ jobs:
           ssh-key: ${{ secrets.SSH_KEY_FOR_CI }}
       - name: Download cub for CUDA 10.2
         run: |
-          CUDA_VERSION=$(cat $CUDA_HOME/version.txt | grep "CUDA Version" | awk '{print $NF}' | cut -d. -f1,2)
+          CUDA_VERSION=$(nvcc -V | awk -F ',| ' '/release/{print $6}')
 
           # check if it is CUDA 10.2
           # download cub
diff --git a/.github/workflows/compatiblity_test_on_schedule.yml b/.github/workflows/compatiblity_test_on_schedule.yml
index 9802795fa..15ac4f1a9 100644
--- a/.github/workflows/compatiblity_test_on_schedule.yml
+++ b/.github/workflows/compatiblity_test_on_schedule.yml
@@ -61,6 +61,18 @@ jobs:
         with:
           ssh-key: ${{ secrets.SSH_KEY_FOR_CI }}
 
+      - name: Download cub for CUDA 10.2
+        run: |
+          CUDA_VERSION=$(nvcc -V | awk -F ',| ' '/release/{print $6}')
+
+          # check if it is CUDA 10.2
+          # download cub
+          if [ "$CUDA_VERSION" = "10.2" ]; then
+            wget https://github.com/NVIDIA/cub/archive/refs/tags/1.8.0.zip
+            unzip 1.8.0.zip
+            cp -r cub-1.8.0/cub/ colossalai/kernel/cuda_native/csrc/kernels/include/
+          fi
+
       - name: Install Colossal-AI
         run: |
           pip install -v --no-cache-dir .
diff --git a/.github/workflows/cuda_ext_check_before_merge.yml b/.github/workflows/cuda_ext_check_before_merge.yml
index eba5bb98e..686f0f395 100644
--- a/.github/workflows/cuda_ext_check_before_merge.yml
+++ b/.github/workflows/cuda_ext_check_before_merge.yml
@@ -37,6 +37,18 @@ jobs:
       - name: Install PyTorch
         run: eval ${{ matrix.build.torch_command }}
 
+      - name: Download cub for CUDA 10.2
+        run: |
+          CUDA_VERSION=$(nvcc -V | awk -F ',| ' '/release/{print $6}')
+
+          # check if it is CUDA 10.2
+          # download cub
+          if [ "$CUDA_VERSION" = "10.2" ]; then
+            wget https://github.com/NVIDIA/cub/archive/refs/tags/1.8.0.zip
+            unzip 1.8.0.zip
+            cp -r cub-1.8.0/cub/ colossalai/kernel/cuda_native/csrc/kernels/include/
+          fi
+
       - name: Build
         run: |
           CUDA_EXT=1 pip install -v .
diff --git a/pytest.ini b/pytest.ini
index 01e5cd217..e99fe3f08 100644
--- a/pytest.ini
+++ b/pytest.ini
@@ -4,3 +4,4 @@ markers =
     gpu: tests which requires a single GPU
     dist: tests which are run in a multi-GPU or multi-machine environment
     experiment: tests for experimental features
+addopts = --ignore=tests/test_analyzer --ignore=tests/test_auto_parallel --ignore=tests/test_autochunk
diff --git a/version.txt b/version.txt
index 0d91a54c7..9e11b32fc 100644
--- a/version.txt
+++ b/version.txt
@@ -1 +1 @@
-0.3.0
+0.3.1