From ff781ed05945ea8760aebe4c1b368f54ed5a29e8 Mon Sep 17 00:00:00 2001
From: Soumith Chintala <soumith@gmail.com>
Date: Tue, 18 Apr 2017 15:39:26 -0400
Subject: [PATCH] Update CONTRIBUTING.md

---
 CONTRIBUTING.md | 44 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 44 insertions(+)

diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index d5a34877194..4ddf731120a 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -70,5 +70,49 @@ For example:
 
 You do not need to repeatedly install after modifying python files.
 
+When you are developing on the C++ side of things, the environment variables `DEBUG` and `NOCUDA` are helpful.
+
+- `DEBUG=1` will enable debug builds (-g -O0)
+- `NOCUDA=1` will disable compiling CUDA (in case you are developing on something not CUDA related), to save compile time.
+
+For example:
+```
+NO_CUDA=1 DEBUG=1 python setup.py build develop
+```
+
+Also, if you are developing a lot, using ccache is a real time-saver. By default, ccache does not properly support CUDA stuff, so here are the instructions for installing a custom `ccache` fork that has CUDA support:
+```
+# install and export ccache
+if ! ls ~/ccache/bin/ccache
+then
+    sudo apt-get update
+    sudo apt-get install -y automake autoconf
+    sudo apt-get install -y asciidoc
+    mkdir -p ~/ccache
+    pushd /tmp
+    rm -rf ccache
+    git clone https://github.com/colesbury/ccache -b ccbin
+    pushd ccache
+    ./autogen.sh
+    ./configure
+    make install prefix=~/ccache
+    popd
+    popd
+
+    mkdir -p ~/ccache/lib
+    mkdir -p ~/ccache/cuda
+    ln -s ~/ccache/bin/ccache ~/ccache/lib/cc
+    ln -s ~/ccache/bin/ccache ~/ccache/lib/c++
+    ln -s ~/ccache/bin/ccache ~/ccache/lib/gcc
+    ln -s ~/ccache/bin/ccache ~/ccache/lib/g++
+    ln -s ~/ccache/bin/ccache ~/ccache/cuda/nvcc
+
+    ~/ccache/bin/ccache -M 25Gi
+fi
+
+export PATH=~/ccache/lib:$PATH
+export CUDA_NVCC_EXECUTABLE=~/ccache/cuda/nvcc
+```
+
 
 Hope this helps, and thanks for considering to contribute.