Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Support tf version >= 1.10 #135

Open
wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Support tf version >= 1.10 #135

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
5229151
Support tf version >= 1.10
Apr 1, 2019
08017b9
adapt to cuda runtime version
HawkAaron Sep 23, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions src/reduce.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,7 +41,11 @@ struct CTAReduce {

T shuff;
for (int offset = warp_size / 2; offset > 0; offset /= 2) {
#if CUDART_VERSION < 9000
shuff = __shfl_down(x, offset);
#else
shuff = __shfl_down_sync(0xFFFFFFFF, x, offset);
#endif
if (tid + offset < count && tid < offset)
x = g(x, shuff);
}
Expand Down
21 changes: 9 additions & 12 deletions tensorflow_binding/setup.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,7 @@
import unittest
import warnings
from setuptools.command.build_ext import build_ext as orig_build_ext
from distutils.version import LooseVersion

# We need to import tensorflow to find where its include directory is.
try:
Expand All @@ -27,13 +28,6 @@
else:
enable_gpu = True


if "TENSORFLOW_SRC_PATH" not in os.environ:
print("Please define the TENSORFLOW_SRC_PATH environment variable.\n"
"This should be a path to the Tensorflow source directory.",
file=sys.stderr)
sys.exit(1)

if platform.system() == 'Darwin':
lib_ext = ".dylib"
else:
Expand All @@ -52,13 +46,16 @@
root_path = os.path.realpath(os.path.dirname(__file__))

tf_include = tf.sysconfig.get_include()
tf_src_dir = os.environ["TENSORFLOW_SRC_PATH"]
tf_src_dir = tf.sysconfig.get_lib() # os.environ["TENSORFLOW_SRC_PATH"]
tf_includes = [tf_include, tf_src_dir]
warp_ctc_includes = [os.path.join(root_path, '../include')]
include_dirs = tf_includes + warp_ctc_includes

if tf.__version__ >= '1.4':
include_dirs += [tf_include + '/../../external/nsync/public']
if LooseVersion(tf.__version__) >= LooseVersion('1.4'):
nsync_dir = '../../external/nsync/public'
if LooseVersion(tf.__version__) >= LooseVersion('1.10'):
nsync_dir = 'external/nsync/public'
include_dirs += [os.path.join(tf_include, nsync_dir)]

if os.getenv("TF_CXX11_ABI") is not None:
TF_CXX11_ABI = os.getenv("TF_CXX11_ABI")
Expand All @@ -78,9 +75,9 @@
extra_compile_args += ['-Wno-return-type']

extra_link_args = []
if tf.__version__ >= '1.4':
if LooseVersion(tf.__version__) >= LooseVersion('1.4'):
if os.path.exists(os.path.join(tf_src_dir, 'libtensorflow_framework.so')):
extra_link_args = ['-L' + tf.sysconfig.get_lib(), '-ltensorflow_framework']
extra_link_args = ['-L' + tf_src_dir, '-ltensorflow_framework']

if (enable_gpu):
extra_compile_args += ['-DWARPCTC_ENABLE_GPU']
Expand Down