diff --git a/CMakeLists.txt b/CMakeLists.txt
index 2f80f38..faeb0e7 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -25,7 +25,12 @@ add_subdirectory(third_party/sentencepiece)
 
 if (GGML_CUBLAS)
     add_compile_definitions(GGML_USE_CUBLAS)
-    set(CUDA_ARCHITECTURES "52;61;70;75;80;86" CACHE STRING "chatglm: cuda architectures to compile")
+    enable_language(CUDA)
+    if (CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL "11.8")
+        set(CUDA_ARCHITECTURES "52;61;70;75;80;86;90" CACHE STRING "chatglm: cuda architectures to compile")
+    else ()
+        set(CUDA_ARCHITECTURES "52;61;70;75;80;86" CACHE STRING "chatglm: cuda architectures to compile")
+    endif ()
     set_property(TARGET ggml PROPERTY CUDA_ARCHITECTURES ${CUDA_ARCHITECTURES})
 endif ()
 
diff --git a/README.md b/README.md
index d090989..67addd9 100644
--- a/README.md
+++ b/README.md
@@ -274,7 +274,7 @@ cmake -B build -DGGML_CUBLAS=ON -DCUDA_ARCHITECTURES="80"       # for A100
 cmake -B build -DGGML_CUBLAS=ON -DCUDA_ARCHITECTURES="70;75"    # compatible with both V100 and T4
 ```
 
-To find out the CUDA architecture of your GPU device, see [Matching CUDA arch and CUDA gencode for various NVIDIA architectures](https://arnon.dk/matching-sm-architectures-arch-and-gencode-for-various-nvidia-cards/).
+To find out the CUDA architecture of your GPU device, see [Your GPU Compute Capability](https://developer.nvidia.com/cuda-gpus).
 
 **Metal**