Installing theano 0.9 with GPU support on Ubuntu 14.04, TitanX pascal, Cuda8.0 - theano

Question

It's driving me crazy. I'm following this guide to install the required libgpuarray. But I keep on getting errors.
In this guide:
cd <dir>
rm -rf build Build
mkdir Build
cd Build
cmake .. -DCMAKE_INSTALL_PREFIX=~/.local -DCMAKE_BUILD_TYPE=Release
make
make install
DEVICE="<test device>" make test
cd ..
# Run the following export and add them in your ~/.bashrc file
export CPATH=$CPATH:~/.local/include
export LIBRARY_PATH=$LIBRARY_PATH:~/.local/lib
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:~/.local/lib
python setup.py build
python setup.py install --user
cd
DEVICE="<test device>" python -c "import pygpu;pygpu.test()"
I'm supposed to reference my gpu in DEVICE="<test device>". However I tried everything here: gpu, gpu0, etc
but it still cannot find any GPU...
I get the following errors:
$python setup.py build
Traceback (most recent call last):
File "setup.py", line 15, in <module>
raise Exception('cython is too old or not installed '
Exception: cython is too old or not installed (at least 0.25 required)
which I just ignore since my pip is not able to install a lower version than the one that is globally installed, although I'm using virtualenv. And finally I get this error:
$DEVICE="gpu0" python -c "import pygpu;pygpu.test()"
ERROR: Failure: ValueError (Unknown device format:gpu)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/usr/local/lib/python2.7/dist-packages/nose/loader.py", line 418, in loadTestsFromName
addr.filename, addr.module)
File "/usr/local/lib/python2.7/dist-packages/nose/importer.py", line 47, in importFromPath
return self.importFromDir(dir_path, fqname)
File "/usr/local/lib/python2.7/dist-packages/nose/importer.py", line 94, in importFromDir
mod = load_module(part_fqname, fh, filename, desc)
File "/home/myUser/.local/lib/python2.7/site-packages/pygpu-0.6.5-py2.7-linux-x86_64.egg/pygpu/tests/test_tools.py", line 5, in <module>
from .support import (guard_devsup, rand, check_flags, check_meta, check_all,
File "/home/myUser/.local/lib/python2.7/site-packages/pygpu-0.6.5-py2.7-linux-x86_64.egg/pygpu/tests/support.py", line 32, in <module>
context = gpuarray.init(get_env_dev())
File "pygpu/gpuarray.pyx", line 634, in pygpu.gpuarray.init (pygpu/gpuarray.c:9407)
File "pygpu/gpuarray.pyx", line 583, in pygpu.gpuarray.pygpu_init (pygpu/gpuarray.c:9073)
ValueError: Unknown device format:gpu
----------------------------------------------------------------------
Ran 7 tests in 0.002s
FAILED (errors=7)
Edit:
Ok, so it compiles now when using "cuda0"
DEVICE="cuda0" python -c "import pygpu;pygpu.test()"
but still the computation seems to fail:
======================================================================
FAIL: pygpu.tests.test_blas.test_rgemmBatch_3d(16, 16, 9, 16, 'float32', ('f', 'f', 'c'), (False, False), False, 1, True, True, 0.6, 0.6)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/usr/local/lib/python2.7/dist-packages/nose/case.py", line 197, in runTest
self.test(*self.arg)
File "/home/myUser/.local/lib/python2.7/site-packages/pygpu-0.6.5-py2.7-linux-x86_64.egg/pygpu/tests/support.py", line 39, in f
func(*args, **kwargs)
File "/home/myUser/.local/lib/python2.7/site-packages/pygpu-0.6.5-py2.7-linux-x86_64.egg/pygpu/tests/test_blas.py", line 225, in rgemmBatch_3d
numpy.testing.assert_allclose(cr, numpy.asarray(gr), rtol=1e-5)
File "/usr/local/lib/python2.7/dist-packages/numpy/testing/utils.py", line 1392, in assert_allclose
verbose=verbose, header=header)
File "/usr/local/lib/python2.7/dist-packages/numpy/testing/utils.py", line 739, in assert_array_compare
raise AssertionError(msg)
AssertionError:
Not equal to tolerance rtol=1e-05, atol=0
(mismatch 99.9565972222%)
x: array([[[ 364.203369, 258.57843 , 282.774231, ..., 322.785706,
185.089233, 300.292389],
[ 203.502457, 226.207962, 202.701843, ..., 213.377808,...
y: array([[[ 239.617569, 217.584824, 167.131165, ..., 211.116486,
313.409271, 244.467926],
[ 248.086868, 237.813416, 164.860977, ..., 216.871902,...
----------------------------------------------------------------------
Ran 7292 tests in 177.637s
FAILED (failures=137)