# PyTorch Install with Python3 Broken
**URL:** https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146
**Category:** Jetson TX2
**Created:** [October 13, 2018, 8:44pm UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146 "2018-10-13T20:44:33Z")
**Posts on this page:** 20
**Page:** 1
### Author:  [@reverts](https://forums.developer.nvidia.com/u/reverts)
#### Post date: [October 13, 2018, 8:44pm UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/1 "2018-10-13T20:44:33Z")
I originally had a huge setup, and just decided to wipe the Jetson TX2, reinstall Jetpack, and then use Dusty’s Jetson Reinforcement script. It works ok, but only compiles for Python 2.7, can’t import it into Python 3.
> **[GitHub - dusty-nv/jetson-reinforcement: Deep reinforcement learning GPU libraries for...](https://github.com/dusty-nv/jetson-reinforcement)**
>
> Deep reinforcement learning GPU libraries for NVIDIA Jetson TX1/TX2 with PyTorch, OpenAI Gym, and Gazebo robotics simulator.
So, that’s not going to work.
I’ll leave a note here when/if I get it working for Python3 on my own fork.
---
### Author:  [@AastaLLL](https://forums.developer.nvidia.com/u/AastaLLL)
#### Post date: [October 15, 2018, 9:44am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/2 "2018-10-15T09:44:06Z")
Hi,
You can setup Python3 environment by modifying this script:
[url][https://github.com/dusty-nv/jetson-scripts/blob/master/pytorch\_jetson\_install.sh[/url]](https://github.com/dusty-nv/jetson-scripts/blob/master/pytorch_jetson_install.sh%5B/url%5D)
Just replace the pip related commands with pip3.
Thanks.
---
### Author:  [@reverts](https://forums.developer.nvidia.com/u/reverts)
#### Post date: [October 15, 2018, 3:05pm UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/3 "2018-10-15T15:05:57Z")
I did run that script with the pip changes, and also tried it with both
```auto
sudo python ...
```
to
```auto
sudo python3 ...
```
.
If I don’t use
```auto
sudo python3 ...
```
, it tells me it can’t find any of the libraries I installed under pip3, which makes sense since if I’m invoking pip, a ‘sudo python’ command will look for the Python 2 instance.
Then, when using
```auto
sudo python3 ...
```
and pip3, got this error:
Errors are:
```auto
...about 100 NVLink errors, listing the last few below along with final error log.
nvlink error : entry function '_Z28ncclAllReduceLLKernel_sum_i88ncclColl' with max regcount of 80 calls function '_Z25ncclReduceScatter_max_u64P14CollectiveArgs' with regcount of 96
nvlink error : entry function '_Z29ncclAllReduceLLKernel_sum_i328ncclColl' with max regcount of 80 calls function '_Z25ncclReduceScatter_max_u64P14CollectiveArgs' with regcount of 96
nvlink error : entry function '_Z29ncclAllReduceLLKernel_sum_f168ncclColl' with max regcount of 80 calls function '_Z25ncclReduceScatter_max_u64P14CollectiveArgs' with regcount of 96
nvlink error : entry function '_Z29ncclAllReduceLLKernel_sum_u328ncclColl' with max regcount of 80 calls function '_Z25ncclReduceScatter_max_u64P14CollectiveArgs' with regcount of 96
nvlink error : entry function '_Z29ncclAllReduceLLKernel_sum_f328ncclColl' with max regcount of 80 calls function '_Z25ncclReduceScatter_max_u64P14CollectiveArgs' with regcount of 96
nvlink error : entry function '_Z29ncclAllReduceLLKernel_sum_u648ncclColl' with max regcount of 80 calls function '_Z25ncclReduceScatter_max_u64P14CollectiveArgs' with regcount of 96
nvlink error : entry function '_Z28ncclAllReduceLLKernel_sum_u88ncclColl' with max regcount of 80 calls function '_Z25ncclReduceScatter_max_u64P14CollectiveArgs' with regcount of 96
Makefile:83: recipe for target '/home/nvidia/jetson-reinforcement/build/pytorch/third_party/build/nccl/obj/collectives/device/devlink.o' failed
make[5]: *** [/home/nvidia/jetson-reinforcement/build/pytorch/third_party/build/nccl/obj/collectives/device/devlink.o] Error 255
Makefile:45: recipe for target 'devicelib' failed
make[4]: *** [devicelib] Error 2
Makefile:24: recipe for target 'src.build' failed
make[3]: *** [src.build] Error 2
CMakeFiles/nccl.dir/build.make:60: recipe for target 'lib/libnccl.so' failed
make[2]: *** [lib/libnccl.so] Error 2
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/nccl.dir/all' failed
make[1]: *** [CMakeFiles/nccl.dir/all] Error 2
Makefile:127: recipe for target 'all' failed
make: *** [all] Error 2
Failed to run 'bash ../tools/build_pytorch_libs.sh --use-cuda --use-nnpack nccl caffe2 libshm gloo c10d THD'
```
I’m not sure why the Jetson isn’t liking the setup. What are the regcount errors here?
---
### Author:  [@AastaLLL](https://forums.developer.nvidia.com/u/AastaLLL)
#### Post date: [October 18, 2018, 9:42am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/4 "2018-10-18T09:42:26Z")
Hi,
Please turn-off NCCL support since it is only available on the desktop GPU.
```auto
diff --git a/CMakeLists.txt b/CMakeLists.txt
index f7b24b728..f75f610ed 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -93,7 +93,7 @@ option(USE_LMDB "Use LMDB" ON)
option(USE_METAL "Use Metal for iOS build" ON)
option(USE_MOBILE_OPENGL "Use OpenGL for mobile code" ON)
option(USE_NATIVE_ARCH "Use -march=native" OFF)
-option(USE_NCCL "Use NCCL" ON)
+option(USE_NCCL "Use NCCL" OFF)
option(USE_SYSTEM_NCCL "Use system-wide NCCL" OFF)
option(USE_NNAPI "Use NNAPI" OFF)
option(USE_NNPACK "Use NNPACK" ON)
diff --git a/setup.py b/setup.py
index 99817f346..e39042b83 100644
--- a/setup.py
+++ b/setup.py
@@ -195,6 +195,7 @@ IS_LINUX = (platform.system() == 'Linux')
BUILD_PYTORCH = check_env_flag('BUILD_PYTORCH')
USE_CUDA_STATIC_LINK = check_env_flag('USE_CUDA_STATIC_LINK')
RERUN_CMAKE = True
+USE_NCCL = False
NUM_JOBS = multiprocessing.cpu_count()
max_jobs = os.getenv("MAX_JOBS")
```
Thanks.
---
### Author:  [@reverts](https://forums.developer.nvidia.com/u/reverts)
#### Post date: [October 18, 2018, 4:46pm UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/5 "2018-10-18T16:46:32Z")
Ok, so now it gets further in the compile, which is awesome!
I start to get a bunch of errors while it’s processing Onnyx and TensorRT, which are both needed to run any models on the Jetson.
Those are Error 1’s on a few onnx-tensorrt/\* files.
Finally, I get to the breaking code:
```auto
[26%] Building CXX object third_party/onnx/CMakeFiles/onnx.dir/onnx/defs/logical/old.cc.o
[26%] Building CXX object third_party/onnx/CMakeFiles/onnx.dir/onnx/defs/traditionalml/defs.cc.o
[26%] Linking CXX static library ../../lib/libonnx.a
[26%] Built target onnx
Makefile:160: recipe for target 'all' failed
make: *** [all] Error 2
Failed to run 'bash ../tools/build_pytorch_libs.sh --use-cuda --use-nnpack caffe2 libshm gloo c10d THD'
```
It says it builds that target, but then just fails. Does that mean it’s Onnx, or whatever is next in the build?
---
### Author:  [@AastaLLL](https://forums.developer.nvidia.com/u/AastaLLL)
#### Post date: [October 19, 2018, 3:31am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/6 "2018-10-19T03:31:21Z")
Hi,
Could you try to manually run these commands in the _pyTorch_ folder:
```auto
sudo pip install -U setuptools
sudo pip install -r requirements.txt
```
We can successfully build _pyTorch_ with the change shared in the comment#4 by executing the command manually.
It looks some package cannot be well-installed if running inside the script.
To type it directly on the console should be okay.
Thanks and please let us know the result.
---
### Author:  [@reverts](https://forums.developer.nvidia.com/u/reverts)
#### Post date: [October 20, 2018, 1:42am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/7 "2018-10-20T01:42:01Z")
For the above code, I would need to use:
```auto
sudo pip3 install -U setuptools
sudo pip3 install -r requirements.txt
```
Unfortunately, that doesn’t get me any further in the process. Everything breaks down around the Onnx and onnxTensorRT installs. I’m heading over to the PyTorch forums to see if they can help figure anything out.
Again, I can get this to work for Python 2, but not Python3.
---
### Author:  [@reverts](https://forums.developer.nvidia.com/u/reverts)
#### Post date: [October 22, 2018, 1:04am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/8 "2018-10-22T01:04:01Z")
Ok, so I got it installed, latest version v1.0 release. It does NOT have TensorRT installed on it. I’m still figuring out if this is a problem, or if I should just do the Pytorch-\>Onnx-\>TensorRT on a desktop GPU and then transfer that to the Jetson directly.
Here’s how I did it:
Note: Don’t use Dusty’s long install that checks out v3.0. It’s unnecessary anymore.
```auto
sudo nvpmodel -m 0
sudo pip3 install numpy
git clone --recursive https://github.com/pytorch/pytorch
cd pytorch
sudo pip3 install -U setuptools
sudo pip3 install -r requirements.txt
```
Apply patches as noted above:
CmakeLists.txt : Change NCCL to ‘Off’ on line 97
setup.py: Add USE\_NCCL = False below line 198
```auto
sudo python3 setup.py install
```
This will take awhile, even in mode 0 cranking all CPUs
```auto
cd ..
python3
```
```auto
import torch
torch. __version__
```
Should return 1.0\*
---
### Author:  [@393702504](https://forums.developer.nvidia.com/u/393702504)
#### Post date: [October 30, 2018, 2:04pm UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/9 "2018-10-30T14:04:46Z")
hello, i tried to install the pytorch follow the commet8, however it errored as:
/home/nvidia/Tools/pytorch/torch/csrc/cuda/nccl.h:21:23: error: variable or field ‘throw\_nccl\_error’ declared void
void throw\_nccl\_error(ncclResult\_t status);
I have changed NCCL to 'Off’in cmakelists and added USE\_NCCL = False in setup.py,why it still make the nccl?How can I solve the problem? Thanks,waiting for reply.
---
### Author:  [@brenozanchetta](https://forums.developer.nvidia.com/u/brenozanchetta)
#### Post date: [October 31, 2018, 2:45pm UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/10 "2018-10-31T14:45:42Z")
I`m also having trouble on how to set NCCL off. I can`t understand what AastaLLL did.
Can anyone help to elaborate further please?
---
### Author:  [@brenozanchetta](https://forums.developer.nvidia.com/u/brenozanchetta)
#### Post date: [November 6, 2018, 8:47pm UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/11 "2018-11-06T20:47:03Z")
Finally, I was able to work my way around the problem.
The final tutorial is to be used after a clean JetPack flash.
I strongly suggest either JetPack3.2.1 or JetPack3.3
[https://github.com/Bfzanchetta/DLARM/blob/master/jetsontx2/pytorch\_install.sh](https://github.com/Bfzanchetta/DLARM/blob/master/jetsontx2/pytorch_install.sh)
Have a good one!
---
### Author:  [@ts\_alpaca](https://forums.developer.nvidia.com/u/ts_alpaca)
#### Post date: [April 18, 2019, 8:27am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/12 "2019-04-18T08:27:32Z")
hello, I would like to compile libtorch c++ api with command
mkdir build
cd build && python …/tools/build\_libtorch.py
You said nccl is only available for desktop GPU. So I change USE\_NCCL=“OFF” in build\_pytorch\_libs.py
but it failed in progress 98%, It seems something others has used nccl.h log as follows
In file included from /home/nvidia/pytorch/torch/lib/c10d/ProcessGroupNCCL.cpp:1:0:
/home/nvidia/pytorch/torch/lib/c10d/…/c10d/ProcessGroupNCCL.hpp:17:18: fatal error: nccl.h: No such file or directory
compilation terminated.
caffe2/torch/lib/c10d/CMakeFiles/c10d.dir/build.make:206: recipe for target ‘caffe2/torch/lib/c10d/CMakeFiles/c10d.dir/ProcessGroupNCCL.cpp.o’ failed
make[2]: \*\*\* [caffe2/torch/lib/c10d/CMakeFiles/c10d.dir/ProcessGroupNCCL.cpp.o] Error 1
CMakeFiles/Makefile2:9267: recipe for target ‘caffe2/torch/lib/c10d/CMakeFiles/c10d.dir/all’ failed
make[1]: \*\*\* [caffe2/torch/lib/c10d/CMakeFiles/c10d.dir/all] Error 2
Makefile:140: recipe for target ‘all’ failed
make: \*\*\* [all] Error 2
Traceback (most recent call last):
File “…/tools/build\_libtorch.py”, line 22, in
build\_python=False, rerun\_cmake=True, build\_dir=‘.’)
File “/home/nvidia/pytorch/tools/build\_pytorch\_libs.py”, line 282, in build\_caffe2
check\_call([‘make’, ‘-j’, str(max\_jobs), ‘install’], cwd=build\_dir, env=my\_env)
File “/usr/lib/python2.7/subprocess.py”, line 541, in check\_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command ‘[‘make’, ‘-j’, ‘6’, ‘install’]’ returned non-zero exit status 2
---
### Author:  [@brenozanchetta](https://forums.developer.nvidia.com/u/brenozanchetta)
#### Post date: [April 24, 2019, 6:26am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/13 "2019-04-24T06:26:00Z")
I solved the problem like this:
I went to the folder that contained nccl.h on my device.
Then, i wrote a single line that copies all .h extension files from that folder to a place I desire
Then I adapt this command to send all these .h to the problematic folder /home/nvidia/pytorch/torch/lib/c10d/…/c10d/
something like
$sudo -s
$cp /path/to/files{file1.h, nccl.h, file3.h, …} /home/nvidia/pytorch/torch/lib/c10d/…/c10d/
Then, I copy the command above and restart the installation.
Then, when it reaches close to 98% (maybe 96, maybe 97) I open a second terminal and paste the command and hit enter. Then, until the installation of the other terminal ends, I keep hitting Up-Arrow and enter every time a new line comes up on the installation. It gets stuck on 98%, then on 99% and even on 100%!
Apparently, the routine that is managing the installation does not recognize our .h files, and so, everytime we re-run the installation, it redefines the paths to the folder of these .h files. Besides that, the same installation refreshes the annoying folder that we are populating with these .h, go figure.
I did it exhaustively and finished it with success.
Tip: Do not attempt to import torch on python after installation if you forget to leave pytorch’s directory, it will not work. Do a cd … before.
Good luck!
---
### Author:  [@571777400](https://forums.developer.nvidia.com/u/571777400)
#### Post date: [April 25, 2019, 2:47am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/14 "2019-04-25T02:47:07Z")
hello!
i also have some problems in compile pytorch in tx2,my environment is ubuntu16.04LTS,cuda9+cudnn7,i compiled tensorflow successed,but when i install pytorch use:
git clone --recursive [GitHub - pytorch/pytorch: Tensors and Dynamic neural networks in Python with strong GPU acceleration](https://github.com/pytorch/pytorch)
cd pytorch
sudo pip install -U setuptools
sudo pip install -r requirements.txt
sudo python3 setup.py install
then,i got some problems like this:
nvlink error : entry function ‘\_Z30ncclReduceTreeLLKernel\_sum\_f328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceRingLLKernel\_sum\_f328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceTreeLLKernel\_sum\_f168ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceRingLLKernel\_sum\_f168ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceTreeLLKernel\_sum\_u648ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceRingLLKernel\_sum\_u648ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceTreeLLKernel\_sum\_i648ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceRingLLKernel\_sum\_i648ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceTreeLLKernel\_sum\_u328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceRingLLKernel\_sum\_u328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceTreeLLKernel\_sum\_i328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z30ncclReduceRingLLKernel\_sum\_i328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z29ncclReduceTreeLLKernel\_sum\_u88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z29ncclReduceRingLLKernel\_sum\_u88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z29ncclReduceTreeLLKernel\_sum\_i88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z29ncclReduceRingLLKernel\_sum\_i88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclBroadcastTreeLLKernel\_copy\_i88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclBroadcastRingLLKernel\_copy\_i88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllGatherTreeLLKernel\_copy\_i88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllGatherRingLLKernel\_copy\_i88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceTreeLLKernel\_sum\_f648ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceRingLLKernel\_sum\_f648ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceTreeLLKernel\_sum\_f328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceRingLLKernel\_sum\_f328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceTreeLLKernel\_sum\_f168ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceRingLLKernel\_sum\_f168ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceTreeLLKernel\_sum\_u648ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceRingLLKernel\_sum\_u648ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceTreeLLKernel\_sum\_i648ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceRingLLKernel\_sum\_i648ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceTreeLLKernel\_sum\_u328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceRingLLKernel\_sum\_u328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceTreeLLKernel\_sum\_i328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z33ncclAllReduceRingLLKernel\_sum\_i328ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z32ncclAllReduceTreeLLKernel\_sum\_u88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z32ncclAllReduceRingLLKernel\_sum\_u88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z32ncclAllReduceTreeLLKernel\_sum\_i88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
nvlink error : entry function ‘\_Z32ncclAllReduceRingLLKernel\_sum\_i88ncclColl’ with max regcount of 80 calls function ‘\_Z29ncclReduceScatterRing\_max\_f64P14CollectiveArgs’ with regcount of 96
Makefile:68: recipe for target ‘/home/nvidia/pytorch/build/nccl/obj/collectives/device/devlink.o’ failed
make[5]: \*\*\* [/home/nvidia/pytorch/build/nccl/obj/collectives/device/devlink.o] Error 255
Makefile:44: recipe for target ‘/home/nvidia/pytorch/build/nccl/obj/collectives/device/colldevice.a’ failed
make[4]: \*\*\* [/home/nvidia/pytorch/build/nccl/obj/collectives/device/colldevice.a] Error 2
Makefile:25: recipe for target ‘src.build’ failed
make[3]: \*\*\* [src.build] Error 2
CMakeFiles/nccl\_external.dir/build.make:110: recipe for target ‘nccl\_external-prefix/src/nccl\_external-stamp/nccl\_external-build’ failed
make[2]: \*\*\* [nccl\_external-prefix/src/nccl\_external-stamp/nccl\_external-build] Error 2
CMakeFiles/Makefile2:67: recipe for target ‘CMakeFiles/nccl\_external.dir/all’ failed
make[1]: \*\*\* [CMakeFiles/nccl\_external.dir/all] Error 2
make[1]: \*\*\* Waiting for unfinished jobs…
[65%] Built target caffe2
Makefile:138: recipe for target ‘all’ failed
make: \*\*\* [all] Error 2
Traceback (most recent call last):
File “setup.py”, line 719, in
build\_deps()
File “setup.py”, line 285, in build\_deps
build\_dir=‘build’)
File “/home/nvidia/pytorch/tools/build\_pytorch\_libs.py”, line 281, in build\_caffe2
check\_call([‘make’, ‘-j’, str(max\_jobs), ‘install’], cwd=build\_dir, env=my\_env)
File “/usr/lib/python3.5/subprocess.py”, line 581, in check\_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command ‘[‘make’, ‘-j’, ‘6’, ‘install’]’ returned non-zero exit status 2
---
### Author:  [@brenozanchetta](https://forums.developer.nvidia.com/u/brenozanchetta)
#### Post date: [April 25, 2019, 4:33am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/15 "2019-04-25T04:33:36Z")
> [@](#):
>
> hello!
> i also have some problems in compile pytorch in tx2,my environment is ubuntu16.04LTS,cuda9+cudnn7,i compiled tensorflow successed,but when i install pytorch use:
> git clone --recursive [GitHub - pytorch/pytorch: Tensors and Dynamic neural networks in Python with strong GPU acceleration](https://github.com/pytorch/pytorch)
> cd pytorch
> sudo pip install -U setuptools
> sudo pip install -r requirements.txt
> sudo python3 setup.py install
>
> then,i got some problems like this:
> …
> Makefile:68: recipe for target ‘/home/nvidia/pytorch/build/nccl/obj/collectives/device/devlink.o’ failed
> make[5]: \*\*\* [/home/nvidia/pytorch/build/nccl/obj/collectives/device/devlink.o] Error 255
> Makefile:44: recipe for target ‘/home/nvidia/pytorch/build/nccl/obj/collectives/device/colldevice.a’ failed
> make[4]: \*\*\* [/home/nvidia/pytorch/build/nccl/obj/collectives/device/colldevice.a] Error 2
> Makefile:25: recipe for target ‘src.build’ failed
> make[3]: \*\*\* [src.build] Error 2
> CMakeFiles/nccl\_external.dir/build.make:110: recipe for target ‘nccl\_external-prefix/src/nccl\_external-stamp/nccl\_external-build’ failed
> make[2]: \*\*\* [nccl\_external-prefix/src/nccl\_external-stamp/nccl\_external-build] Error 2
> CMakeFiles/Makefile2:67: recipe for target ‘CMakeFiles/nccl\_external.dir/all’ failed
> make[1]: \*\*\* [CMakeFiles/nccl\_external.dir/all] Error 2
> make[1]: \*\*\* Waiting for unfinished jobs…
> [65%] Built target caffe2
> Makefile:138: recipe for target ‘all’ failed
> make: \*\*\* [all] Error 2
> Traceback (most recent call last):
> File “setup.py”, line 719, in
> build\_deps()
> File “setup.py”, line 285, in build\_deps
> build\_dir=‘build’)
> File “/home/nvidia/pytorch/tools/build\_pytorch\_libs.py”, line 281, in build\_caffe2
> check\_call([‘make’, ‘-j’, str(max\_jobs), ‘install’], cwd=build\_dir, env=my\_env)
> File “/usr/lib/python3.5/subprocess.py”, line 581, in check\_call
> raise CalledProcessError(retcode, cmd)
> subprocess.CalledProcessError: Command ‘[‘make’, ‘-j’, ‘6’, ‘install’]’ returned non-zero exit status 2
Hi there,
What post/tutorial/notes are you following?
I have not seen this exact error before, but I have a few thoughts:
1. This line
```auto
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/nccl_external.dir/all' failed
```
could be trying to explain that your building failed when processing you NCLL dependencies.
Go to your CMakesList.txt and other possible configuration files and check if USE\_NCCL is set to False (USE\_NCLL=False).
You can easily check if all NCCL options are set to False when you start a new build. Before the log outputs the percentages of the building process
```auto
...
USE_CUDA=True
USE_NCCL=True
...
[1%] Doing Whatever
...
[65%] Built target caffe2 )
```
It is challenging to turn the NCCL off, because sometimes you turn it off in the configuration script, but then the build process reveals NCCL on (Really confusing).
Everything that concerns NCCL on your PyTorch installation should be TURNED OFF on your TX2, because NCCL is meant only for state-of-the-art x86 devices (more or less, I could be wrong, but it definitely does not work on NVIDIA Jetson TX2).
If you feel lost, I suggest you take a look at this guide that I wrote with trial and error. It is a little outdated, but I repeated the process this month and it still is very helpful (90% or more :] ).
[https://github.com/Bfzanchetta/DLARM/blob/master/jetsontx2/pytorch\_install.sh](https://github.com/Bfzanchetta/DLARM/blob/master/jetsontx2/pytorch_install.sh)
Hope this helps you, if you need further assistance feel free to reply.
Have a good debugging!
---
### Author:  [@14beemjamal](https://forums.developer.nvidia.com/u/14beemjamal)
#### Post date: [May 4, 2019, 6:53am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/16 "2019-05-04T06:53:08Z")
> [@](#):
>
> > [@](#):
> >
> > hello!
> > i also have some problems in compile pytorch in tx2,my environment is ubuntu16.04LTS,cuda9+cudnn7,i compiled tensorflow successed,but when i install pytorch use:
> > git clone --recursive [GitHub - pytorch/pytorch: Tensors and Dynamic neural networks in Python with strong GPU acceleration](https://github.com/pytorch/pytorch)
> > cd pytorch
> > sudo pip install -U setuptools
> > sudo pip install -r requirements.txt
> > sudo python3 setup.py install
> >
> > then,i got some problems like this:
> > …
> > Makefile:68: recipe for target ‘/home/nvidia/pytorch/build/nccl/obj/collectives/device/devlink.o’ failed
> > make[5]: \*\*\* [/home/nvidia/pytorch/build/nccl/obj/collectives/device/devlink.o] Error 255
> > Makefile:44: recipe for target ‘/home/nvidia/pytorch/build/nccl/obj/collectives/device/colldevice.a’ failed
> > make[4]: \*\*\* [/home/nvidia/pytorch/build/nccl/obj/collectives/device/colldevice.a] Error 2
> > Makefile:25: recipe for target ‘src.build’ failed
> > make[3]: \*\*\* [src.build] Error 2
> > CMakeFiles/nccl\_external.dir/build.make:110: recipe for target ‘nccl\_external-prefix/src/nccl\_external-stamp/nccl\_external-build’ failed
> > make[2]: \*\*\* [nccl\_external-prefix/src/nccl\_external-stamp/nccl\_external-build] Error 2
> > CMakeFiles/Makefile2:67: recipe for target ‘CMakeFiles/nccl\_external.dir/all’ failed
> > make[1]: \*\*\* [CMakeFiles/nccl\_external.dir/all] Error 2
> > make[1]: \*\*\* Waiting for unfinished jobs…
> > [65%] Built target caffe2
> > Makefile:138: recipe for target ‘all’ failed
> > make: \*\*\* [all] Error 2
> > Traceback (most recent call last):
> > File “setup.py”, line 719, in
> > build\_deps()
> > File “setup.py”, line 285, in build\_deps
> > build\_dir=‘build’)
> > File “/home/nvidia/pytorch/tools/build\_pytorch\_libs.py”, line 281, in build\_caffe2
> > check\_call([‘make’, ‘-j’, str(max\_jobs), ‘install’], cwd=build\_dir, env=my\_env)
> > File “/usr/lib/python3.5/subprocess.py”, line 581, in check\_call
> > raise CalledProcessError(retcode, cmd)
> > subprocess.CalledProcessError: Command ‘[‘make’, ‘-j’, ‘6’, ‘install’]’ returned non-zero exit status 2
>
> Hi there,
>
> What post/tutorial/notes are you following?
> I have not seen this exact error before, but I have a few thoughts:
>
> 1. This line
>
> ```auto
> CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/nccl_external.dir/all' failed
>
> ```
>
> could be trying to explain that your building failed when processing you NCLL dependencies.
> Go to your CMakesList.txt and other possible configuration files and check if USE\_NCCL is set to False (USE\_NCLL=False).
>
> You can easily check if all NCCL options are set to False when you start a new build. Before the log outputs the percentages of the building process
>
> ```auto
> ...
> USE_CUDA=True
> USE_NCCL=True
> ...
> [1%] Doing Whatever
> ...
> [65%] Built target caffe2 )
>
>
> ```
>
> It is challenging to turn the NCCL off, because sometimes you turn it off in the configuration script, but then the build process reveals NCCL on (Really confusing).
>
> Everything that concerns NCCL on your PyTorch installation should be TURNED OFF on your TX2, because NCCL is meant only for state-of-the-art x86 devices (more or less, I could be wrong, but it definitely does not work on NVIDIA Jetson TX2).
>
> If you feel lost, I suggest you take a look at this guide that I wrote with trial and error. It is a little outdated, but I repeated the process this month and it still is very helpful (90% or more :] ).
>
> [https://github.com/Bfzanchetta/DLARM/blob/master/jetsontx2/pytorch\_install.sh](https://github.com/Bfzanchetta/DLARM/blob/master/jetsontx2/pytorch_install.sh)
>
> Hope this helps you, if you need further assistance feel free to reply.
>
> Have a good debugging!
Hi brenozanchetta have you tried building the newer version of pytorch 1.0.1.post2 on Jetpack 3.3 or above. I have previously succesfully compilled pytorch 0.4 on CUDA 8, but building this version of pytorch messed my Jetson TX2 causing me to have to flash it again with Jetpack 3.3.
---
### Author:  [@brenozanchetta](https://forums.developer.nvidia.com/u/brenozanchetta)
#### Post date: [May 5, 2019, 12:05am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/17 "2019-05-05T00:05:25Z")
14beemjamal, I have built pytorch 1.1.0a0+95ca667 on Jetpack 3.3, but higly recommend 3.2.1.
The key point is to turn NCCL off.
NCCL does not work on Jetson TX2, therefore causes installation errors.
Hope this could help.
Did you try the steps I mentioned before?
---
### Author:  [@14beemjamal](https://forums.developer.nvidia.com/u/14beemjamal)
#### Post date: [May 5, 2019, 9:12am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/18 "2019-05-05T09:12:38Z")
I havent yet. Will try them tommorow. Did you need any swap space for the build process. Becuase for me when i run the command.
```auto
python3 setup.py install
```
It slows my Jetson TX2 down completely to the point i cant even move the cursor and starts killing processes. Is it supposed to do that.
---
### Author:  [@14beemjamal](https://forums.developer.nvidia.com/u/14beemjamal)
#### Post date: [May 6, 2019, 6:14pm UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/19 "2019-05-06T18:14:16Z")
brenozanchetta, So i followed the steps mentioned in your script[url][https://github.com/Bfzanchetta/DLARM/blob/master/jetsontx2/pytorch\_install.sh[/url]](https://github.com/Bfzanchetta/DLARM/blob/master/jetsontx2/pytorch_install.sh%5B/url%5D) manually. I was able to succesfully build Pytorch 1.0.1 on Jetpack 3.3, CUDA 9.0 and CUDNN 7.5. Thanks for your work
---
### Author:  [@brenozanchetta](https://forums.developer.nvidia.com/u/brenozanchetta)
#### Post date: [May 7, 2019, 11:12am UTC](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146/20 "2019-05-07T11:12:11Z")
You are welcome! Happy to help.
[Next page](https://forums.developer.nvidia.com/t/pytorch-install-with-python3-broken/66146.md?page=2)