modifications to https://github.com/Theano/Theano/wiki/Windowsinstallation

.theanorc

[global]
device = gpu
floatX = float32

[nvcc]
compiler_bindir=C:\programs86\Microsoft Visual Studio 12.0\VC\bin
flags=-m32 # we have this hard coded for now

[blas]
ldflags =
# ldflags = -lopenblas # placeholder for openblas support

gistfile1.txt

After installing mingw & 32b python 2.7.8:
Add C:\python27\Lib\distutils\distutils.cfg file containing:
[build]
compiler=mingw32

Install pip using http://stackoverflow.com/a/12476379/435460
python get-pip.py

Get devdriver here:
https://developer.nvidia.com/cuda-downloads

My VC\bin path was under 12.0 since I'm using VS 2013.

Set CUDA_ROOT == value at CUDA_PATH (dunno if it did anything)

Add the following to your path (cuda uses cl.exe when compiling stuff) http://stackoverflow.com/a/9000354/435460
C:\Program Files (x86)\Microsoft Visual Studio 12.0\VC\bin\amd64;

nvcc_compiler.py

# ~317

# force use of m32
#if local_bitwidth() == 64:
#    cmd.append('-m64')
#else:
cmd.append('-m32')

python27\Lib\distutils\cygwinccompiler.py

# ~ 327        
        self.set_executables(compiler='gcc%s -O -Wall' % no_cygwin,
                             compiler_so='gcc%s -mdll -O -Wall' % no_cygwin,
                             compiler_cxx='g++%s -O -Wall' % no_cygwin,
                             linker_exe='gcc%s' % no_cygwin,
                             linker_so='%s%s %s %s'
                                    % (self.linker_dll, no_cygwin,
                                       shared_option, entry_point))

siteconf.py

BOOST_INC_DIR = []
BOOST_LIB_DIR = []
BOOST_COMPILER = 'gcc43'
USE_SHIPPED_BOOST = True
BOOST_PYTHON_LIBNAME = ['boost_python']
BOOST_THREAD_LIBNAME = ['boost_thread']
CUDA_TRACE = False
CUDA_ROOT = 'C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v6.5'
CUDA_ENABLE_GL = False
CUDA_ENABLE_CURAND = True
CUDADRV_LIB_DIR = ['${CUDA_ROOT}/lib/Win32']
CUDADRV_LIBNAME = ['cuda']
CUDART_LIB_DIR = ['${CUDA_ROOT}/lib/Win32']
CUDART_LIBNAME = ['cudart']
CURAND_LIB_DIR = ['${CUDA_ROOT}/lib/Win32']
CURAND_LIBNAME = ['curand']
CXXFLAGS = ['']
LDFLAGS = ['']