diff --git a/cmake/cmake_options.yml b/cmake/cmake_options.yml
index 81138241dfc6..6df148323df8 100644
--- a/cmake/cmake_options.yml
+++ b/cmake/cmake_options.yml
@@ -46,7 +46,7 @@ USE_SIGNAL_HANDLER: "ON" # Print stack traces on segfaults.
 USE_TENSORRT: "OFF" # Enable infeference optimization with TensorRT.
 USE_ASAN: "OFF" # Enable Clang/GCC ASAN sanitizers.
 ENABLE_TESTCOVERAGE: "OFF" # Enable compilation with test coverage metric output
-CMAKE_BUILD_TYPE: "Debug"
+CMAKE_BUILD_TYPE: "Release"
 CMAKE_CUDA_COMPILER_LAUNCHER: "ccache"
 CMAKE_C_COMPILER_LAUNCHER: "ccache"
 CMAKE_CXX_COMPILER_LAUNCHER: "ccache"
diff --git a/docs/static_site/src/pages/get_started/build_from_source.md b/docs/static_site/src/pages/get_started/build_from_source.md
index bf57c28add81..e8f7d468b399 100644
--- a/docs/static_site/src/pages/get_started/build_from_source.md
+++ b/docs/static_site/src/pages/get_started/build_from_source.md
@@ -241,7 +241,7 @@ For example, you can specify using all cores on Linux as follows:
 
 ```bash
 mkdir build && cd build
-cmake -GNinja ..
+cmake -DCMAKE_BUILD_TYPE=Release -GNinja ..
 ninja -v
 ```
 
@@ -251,7 +251,7 @@ ninja -v
 
 ```bash
 mkdir build && cd build
-cmake -DUSE_CUDA=1 -DUSE_CUDA_PATH=/usr/local/cuda -DUSE_CUDNN=1 -DUSE_MKLDNN=1 -GNinja ..
+cmake -DUSE_CUDA=1 -DUSE_CUDA_PATH=/usr/local/cuda -DUSE_CUDNN=1 -DUSE_MKLDNN=1 -DCMAKE_BUILD_TYPE=Release -GNinja ..
 ninja -v
 ```
 
@@ -260,7 +260,7 @@ ninja -v
 
 ```bash
 mkdir build && cd build
-cmake -DBLAS=open -DUSE_CUDA=1 -DUSE_CUDA_PATH=/usr/local/cuda -DUSE_CUDNN=1 -GNinja ..
+cmake -DBLAS=open -DUSE_CUDA=1 -DUSE_CUDA_PATH=/usr/local/cuda -DUSE_CUDNN=1 -DCMAKE_BUILD_TYPE=Release -GNinja ..
 ninja -v
 ```
 
@@ -269,7 +269,7 @@ ninja -v
 
 ```bash
 mkdir build && cd build
-cmake -DUSE_CUDA=0 -DUSE_MKLDNN=1 -GNinja ..
+cmake -DUSE_CUDA=0 -DUSE_MKLDNN=1 -DCMAKE_BUILD_TYPE=Release -GNinja ..
 ninja -v
 ```
 
@@ -278,7 +278,7 @@ ninja -v
 
 ```bash
 mkdir build && cd build
-cmake -DUSE_CUDA=0 -DBLAS=open -GNinja ..
+cmake -DUSE_CUDA=0 -DBLAS=open -DCMAKE_BUILD_TYPE=Release -GNinja ..
 ninja -v
 ```
 
@@ -288,7 +288,7 @@ ninja -v
 
 ```bash
 mkdir build && cd build
-cmake -DUSE_OPENCV=0 -GNinja ..
+cmake -DUSE_OPENCV=0 -DCMAKE_BUILD_TYPE=Release -GNinja ..
 ninja -v
 ```
 
@@ -296,7 +296,7 @@ ninja -v
 
 ```bash
 mkdir build && cd build
-cmake -DBLAS=apple -DUSE_OPENCV=0 -DUSE_OPENMP=0 -GNinja ..
+cmake -DBLAS=apple -DUSE_OPENCV=0 -DUSE_OPENMP=0 -DCMAKE_BUILD_TYPE=Release -GNinja ..
 ninja -v
 ```
 
@@ -305,7 +305,7 @@ ninja -v
 ```bash
 brew install llvm
 mkdir build && cd build
-cmake -DBLAS=apple -DUSE_OPENMP=1 -GNinja ..
+cmake -DBLAS=apple -DUSE_OPENMP=1 -DCMAKE_BUILD_TYPE=Release -GNinja ..
 ninja -v
 ```
 
diff --git a/docs/static_site/src/pages/get_started/c_plus_plus.md b/docs/static_site/src/pages/get_started/c_plus_plus.md
index a32496dd68a7..9f4800cd99bf 100644
--- a/docs/static_site/src/pages/get_started/c_plus_plus.md
+++ b/docs/static_site/src/pages/get_started/c_plus_plus.md
@@ -31,7 +31,7 @@ To enable C++ package, just add `USE_CPP_PACKAGE=1` in the [build from source](b
 For example to build MXNet with GPU support and the C++ package, OpenCV, and OpenBLAS, from the project root you would run:
 
 ```bash
-cmake -DUSE_CUDA=1 -DUSE_CUDA_PATH=/usr/local/cuda -DUSE_CUDNN=1 -DUSE_MKLDNN=1 -DUSE_CPP_PACKAGE=1 -GNinja ..
+cmake -DUSE_CUDA=1 -DUSE_CUDA_PATH=/usr/local/cuda -DUSE_CUDNN=1 -DUSE_MKLDNN=1 -DUSE_CPP_PACKAGE=1 -DCMAKE_BUILD_TYPE=Release -GNinja ..
 ninja -v
 ```
 
diff --git a/docs/static_site/src/pages/get_started/ubuntu_setup.md b/docs/static_site/src/pages/get_started/ubuntu_setup.md
index 2d6867d0eadd..e7879eaa1b9e 100644
--- a/docs/static_site/src/pages/get_started/ubuntu_setup.md
+++ b/docs/static_site/src/pages/get_started/ubuntu_setup.md
@@ -213,6 +213,7 @@ Build with CMake and ninja, without GPU and without MKL.
         -DCMAKE_CUDA_COMPILER_LAUNCHER=ccache \
         -DCMAKE_C_COMPILER_LAUNCHER=ccache \
         -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
+        -DCMAKE_BUILD_TYPE=Release \
     ..
     ninja
 ```
@@ -228,6 +229,7 @@ If building on CPU and using MKL and MKL-DNN (make sure MKL is installed accordi
         -DCMAKE_CUDA_COMPILER_LAUNCHER=ccache \
         -DCMAKE_C_COMPILER_LAUNCHER=ccache \
         -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
+        -DCMAKE_BUILD_TYPE=Release \
     ..
     ninja
 ```
@@ -244,6 +246,7 @@ Cuda 10.1 in Ubuntu 18.04 builds fine but is not currently tested in CI.
         -DCMAKE_CUDA_COMPILER_LAUNCHER=ccache \
         -DCMAKE_C_COMPILER_LAUNCHER=ccache \
         -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
+        -DCMAKE_BUILD_TYPE=Release \
     ..
     ninja
 ```