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.

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

OS X makefile changes #5

Closed
wants to merge 1 commit into from
Closed

OS X makefile changes #5

wants to merge 1 commit into from

Conversation

sergeyk
Copy link
Contributor

@sergeyk sergeyk commented Nov 26, 2013

No description provided.

@@ -85,9 +86,11 @@ LIBRARIES := cudart cublas curand protobuf opencv_core opencv_highgui \
PYTHON_LIBRARIES := boost_python python2.7
WARNINGS := -Wall

# NOTE: on OS X, use clang++ to have NVCC play nice.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could you move these 2 likes above - near the CUDA_DIR definition and stuff, since we may need to change it on different platforms?

(We should probably figure out how to write separate config files instead of modifying Makefile for paths... Any advice?)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

One option to split Makefile into Makefile and Makefile.config. We'd place
variables like *_DIR, CXX, etc. which are likely to change on different
computers in Makefile.config, and then include Makefile.config at the top
of Makefile. That way, hopefully only Makefile.config needs to be edited.
Another option is to figure out how to use automake/autoconf or cmake
properly :/. These tools always seem like complicated overkill, but maybe
they're the right way to go.

Ross

On Tue, Nov 26, 2013 at 7:54 AM, Yangqing Jia notifications@github.comwrote:

In Makefile:

@@ -85,9 +86,11 @@ LIBRARIES := cudart cublas curand protobuf opencv_core opencv_highgui
PYTHON_LIBRARIES := boost_python python2.7
WARNINGS := -Wall

+# NOTE: on OS X, use clang++ to have NVCC play nice.

Could you move these 2 likes above - near the CUDA_DIR definition and
stuff, since we may need to change it on different platforms?

(We should probably figure out how to write separate config files instead
of modifying Makefile for paths... Any advice?)


Reply to this email directly or view it on GitHubhttps://github.com/Yangqing/caffe/pull/5/files#r7928356
.

http://www.cs.berkeley.edu/~rbg/

@Yangqing
Copy link
Member

Sergey - I incorporated your changes and also Ross's comment in the current Yangqing/caffe head. I'll close the pull request then :) Doing a git pull on your side should suffice (I assume...)

@Yangqing Yangqing closed this Nov 26, 2013
@sguada
Copy link
Contributor

sguada commented Nov 26, 2013

I get a lot of warnings like
ld: warning: directory not found for option '-L/Developer/NVIDIA/CUDA-5.5/lib64'
ld: warning: directory not found for option '-L/opt/intel/mkl/lib/intel64'

This is due to defining
CUDA_LIB_DIR := $(CUDA_DIR)/lib $(CUDA_DIR)/lib64
MKL_LIB_DIR := $(MKL_DIR)/lib $(MKL_DIR)/lib/intel64

And not having a li64 or intel64 folders, I assume they are 64bits by default.

@sguada sguada reopened this Nov 26, 2013
@Yangqing Yangqing closed this Nov 26, 2013
@sergeyk
Copy link
Contributor Author

sergeyk commented Nov 26, 2013

Correct, but on Linux you need those folders. One more thing that would be
fixed by a proper config system...

On Tue, Nov 26, 2013 at 12:19 PM, Sergio Guadarrama <
notifications@github.com> wrote:

I get a lot of warnings like
ld: warning: directory not found for option
'-L/Developer/NVIDIA/CUDA-5.5/lib64'
ld: warning: directory not found for option '-L/opt/intel/mkl/lib/intel64'

This is due to defining
CUDA_LIB_DIR := $(CUDA_DIR)/lib $(CUDA_DIR)/lib64
MKL_LIB_DIR := $(MKL_DIR)/lib $(MKL_DIR)/lib/intel64

And not having a li64 or intel64 folders, I assume they are 64bits by
default.


Reply to this email directly or view it on GitHubhttps://github.com/Yangqing/caffe/pull/5#issuecomment-29329996
.

@Yangqing
Copy link
Member

They are there for different platforms (e.g. on Linux there is actually a
lib64 folder), so simply ignore the warnings.

Yangqing

On Tue, Nov 26, 2013 at 12:19 PM, Sergio Guadarrama <
notifications@github.com> wrote:

I get a lot of warnings like
ld: warning: directory not found for option
'-L/Developer/NVIDIA/CUDA-5.5/lib64'
ld: warning: directory not found for option '-L/opt/intel/mkl/lib/intel64'

This is due to defining
CUDA_LIB_DIR := $(CUDA_DIR)/lib $(CUDA_DIR)/lib64
MKL_LIB_DIR := $(MKL_DIR)/lib $(MKL_DIR)/lib/intel64

And not having a li64 or intel64 folders, I assume they are 64bits by
default.


Reply to this email directly or view it on GitHubhttps://github.com/Yangqing/caffe/pull/5#issuecomment-29329996
.

@jermainewang jermainewang mentioned this pull request Jan 20, 2014
puzzledqs pushed a commit to puzzledqs/caffe that referenced this pull request Oct 8, 2014
zoemcc pushed a commit to zoemcc/caffe that referenced this pull request Apr 29, 2015
@cysin cysin mentioned this pull request Jul 8, 2015
sh1r0 pushed a commit to sh1r0/caffe that referenced this pull request Jan 2, 2016
Include CUDA only if using CUDA. Using OpenCL and CPU only didn't need include CUDA dependencies
happynear pushed a commit to happynear/caffe that referenced this pull request Feb 26, 2016
naibaf7 pushed a commit that referenced this pull request Mar 6, 2016
Split legacy and non legacy version of cll_backward kernel to reduce branching
mtourne pushed a commit to mtourne/caffe that referenced this pull request Jun 27, 2016
lukeyeager pushed a commit to lukeyeager/caffe that referenced this pull request Jul 8, 2016
Make FP16 blobs support FP32 inputs.
coder-james pushed a commit to coder-james/caffe that referenced this pull request Nov 28, 2016
mindcont pushed a commit to mindcont/caffe that referenced this pull request Apr 25, 2017
mbassov added a commit to mbassov/caffe that referenced this pull request Aug 28, 2017
DEV-26710: Remove OHEM layer from Curalate Caffe
dillonfzw added a commit to bluemindor/caffe that referenced this pull request Dec 3, 2017
*   3f48aeb merge to new bvlc base c0597b1
|\
| *   fc0a02e Merge pull request BVLC#6 from yahoo/jun_cos_layer
| |\
| | * 807ee66 enhance cos layer
| |/
| *   ce1db4b Merge pull request BVLC#5 from yahoo/new_data_layer
| |\
| | * d4b2bf1 new datalayer changes
| |/
| *   5498759 Merge pull request BVLC#4 from yahoo/clean_permission
| |\
| | * 9b7fff8 Permission change for travis
| | * 8377456 Added dynamic linker resolution
| | * 840d6b5 Synchedmem
| | * 5fc8416 Need set_gpu_data
| | * ea0a8cb Merge branch 'bvlc_master' into clean_permission
| | * fb7e2a9 fix file permissions
| |/
| *   611197a Merge pull request BVLC#3 from yahoo/python_path_patch
| |\
| | * e107fb7 Python path patch
| |/
| *   486f979 Merge pull request BVLC#2 from anfeng/master
| |\
| | * 9cdfeb2 fix field Id for dataframe_format
| |/
| * f4e26f1 move dataframe format setting into prototxt
| *   23b0191 Merge pull request BVLC#1 from yahoo/afeng_df
| |\
| | * 177e0d9 MemoryDataLayer optional fields for DataFrames
| |/
| * 4b677c6 initial commit
*   eb8dfc8 Merge pull request BVLC#6
|\
| * 864aa65 implement CPU-GPU parameter update for data parallelism
* 411aafd Merge pull request BVLC#5
* 0cb7d18 Avoid sending the prefetched batch back to host
@RenatGaliew RenatGaliew mentioned this pull request Apr 27, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants