Build tf_core_kernels on With VS2015 stops with "fatal error C1060: compiler is out of heap space"

[stefanopiovesan]

System information

• Build tf_core_kernels.vcxproj
• Windows 10 Enterprise Build 1607
• TensorFlow installed from source
• TensorFlow version Release 1.4.0 from github
• VS2015 Enterprise 14.0.25431 Update 3

Cannot build tf_core_kernels.vcxproj (generated with CMake) because of a fatal error C1060: compiler is out of heap space.

Source code / logs

1>------ Build started: Project: tf_core_kernels, Configuration: RelWithDebInfo x64 ------
1> training_ops.cc
1>e:\software\tensorflow\tensorflow\contrib\cmake\build\external\eigen_archive\unsupported\eigen\cxx11\src/Tensor/TensorBase.h(245): fatal error C1060: compiler is out of heap space
1> INTERNAL COMPILER ERROR in 'C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\bin\x86_amd64\CL.exe'

[tensorflowbutler]

Thank you for your post. We noticed you have not filled out the following field in the issue template. Could you update them if they are relevant in your case, or leave them as N/A? Thanks.
Have I written custom code
OS Platform and Distribution
Bazel version
CUDA/cuDNN version
GPU model and memory
Exact command to reproduce

[spiovesan]

Have I written custom code: No
OS Platform and Distribution: Windows 10 Enterprise Build 1607
Bazel version: N/A
CUDA/cuDNN version: N/A
GPU model and memory: N/A
Exact command to reproduce:
Followed instructions here:
https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/cmake
and then:
MSBuild /p:Configuration=RelWithDebInfo ALL_BUILD.vcxproj

[rongjiecomputer]

I hit this out of heap space error before as well, this prones to happen when you don't have enough RAM.

A workaround is to remove all /MPcompile flag to make cl.exe create less processes and hence use less memory (works for me).

• https://github.com/tensorflow/tensorflow/search?l=CMake&q=%2FMP&type=&utf8=%E2%9C%93
• https://msdn.microsoft.com/en-us/library/bb385193.aspx

[snnn]

If may because of two things:

1. you don't have enough memory.

2. you are using 32 bits toolchains.
   Please upgrade cmake to 3.10.x and add "-T host=x64" to cmake command line args.

[tensorflowbutler]

It has been 14 days with no activity and the awaiting response label was assigned. Is this still an issue? Please update the label and/or status accordingly.

[spiovesan]

I compiled manually one file at a time and then it worked. But compiling the whole project is still an issue.

[tensorflowbutler]

Nagging Awaiting Response: It has been 14 days with no activityand the awaiting response label was assigned. Is this still an issue?

[tensorflowbutler]

Nagging Awaiting Response: It has been 14 days with no activityand the awaiting response label was assigned. Is this still an issue?

[tensorflowbutler]

Nagging Awaiting Response: It has been 14 days with no activityand the awaiting response label was assigned. Is this still an issue?

[habom2310]

In the install instruction from tensorflow, it's already been mentioned.
Building TensorFlow from source can use a lot of RAM. If your system is memory-constrained, limit Bazel's RAM usage with: --local_resources 2048,.5,1.0.