Julia threads hanging when > 10^7 loops - multithreading

Julia script hangs if I set nloops to 10^8, runs fine at 10^7.
I execute using julia -t auto , using a 2021 Macbook Pro.
using Base.Threads
nloops=10^7
nt=nthreads()
println("nthreads: ",nt)
x=0
#threads for i in 1:nloops
global x+=1
end
println("x= ",x)

The answer was simply updating to 1.8.0 and making sure I downloaded the ARM version. Credit to Jerome Richard.

Related

Subprocess.wait() in Python 2.7 vz Pthon 3.x

I'm just migrating my script from Python27 to Python 394 to run my build process like this:
p = subprocess.Popen(_make_clean, creationflags = subprocess.CREATE_NEW_CONSOLE)
p.wait()
p = subprocess.Popen(_make_rebuild, creationflags = subprocess.CREATE_NEW_CONSOLE)
p.wait()
Was working pretty nice in Python 27 but now I can see in Python 394 the Python script is not waiting for subprocesses. Of course, here it is necessary to wait until the first one (clean) is finished to start the second one (rebuild). I was able to make it work without the flag 'CREATE_NEW_CONSOLE'. Unfortunately, I need to separate the outputs for users to see what is the progress. It is just a mess if all scripts are running in the same console.
Why it is not working any more?
Ok, this is an interesting finding. I'm using ConEmu as my main terminal and it is fully integrated into Windows. In the past, I was NOT running my scripts over this emulated terminal and that seems to be the reason why I have this issue today. ConEmu is opening a new terminal window for each subprocess but in a different way than the native Windows Command-Line. When the script is called over Windows CMD everything is fine.

opencv Code keep crashing and didn't execute afterward

I was working on a project that was running fine but now when I try to setup the whole code again for the client. my opencv code execute and crashes the first time I try to run it.afterwards the code doesn't even execute until I restart my system again.then code execute and crash again then it doesn't even execute. the terminal shows the msg(attached image) and nothing happen afterwards
os: ubuntu 18.04
python version: python3
opencv version 4
The problem seems to come from pytroch library as te error comes from torch.nn.modules.
I recommend you to check if you have the last version of Pytorch or if there are some issue within the test_2.py code in line 139 when calling:
utils.predict(model_ft, labels, transforms, test_dir, imgFile)

Python version doesn't seem to be updating

Here is what I get when I run the following commands in my terminal:
python --version
python 2.7.1
python3 --version
python 3.7.3
I ran the 'Update Shell Profile.command' file that comes with the Python3 download, but it still doesn't change the value of
python version
when I run it. My question is, is this normal? Does it change anything important other than needing to type 'python3 file.py' to run 'file.py'? If so, how do I change this? (Also, the computer I am using uses Homebrew which I am not familiar with, I am not sure if this changes anything or how.)
Python 2.7 and Python 3.7 are the latest releases of python 2 and 3 repsectively. Those two are two different version which can co-exist together. You can have both versions installed in your computer at the same time. If u dont like that just uninstall one(I'd suggest python2). The main difference is that python 3 offers many advantages compared to python 2. Some of them would be that some libraries work only for that version other legacy ones work for the other version etc. So there is no problem with having both but like i said there is really no point. Just choose whch version u like and uninstall the other one.

Python - execute sequential commands in same windows command prompt

I am very new to Python (been learning for a few days) . Can anyone help me with the following issue please?
I am trying to write a program to automate the testing of built in commands of a product which is installed on windows and linux.
Before any of the commands can run, the profile needs to be set using profile.cmd.
The commands ( command1.cmd, command2.cmd etc) then need to run in the same windows command prompt as profile.cmd as they rely on the environment which this sets.
The problem I find is that after profile.cmd has been run, python closes the windows command prompt and opens a new one for command1.cmd.
command1.cmd therefore fails to run properly as it is in a new command prompt which has not had the environment set by profile.cmd.
My code is shown below. I am currently testing on windows but will also need to test on linux.
Any suggestions would be much appreciated. Thanks - Kevin
import subprocess
rc = subprocess.run (["profile.cmd"],stdout=subprocess.PIPE)
print(rc.stdout.decode("ascii"))
rc = subprocess.run (["command1.cmd"],stdout=subprocess.PIPE)
print(rc.stdout.decode("ascii"))
rc = subprocess.run (["command2.cmd"],stdout=subprocess.PIPE)
print(rc.stdout.decode("ascii"))
rc = subprocess.run (["command3.cmd"],stdout=subprocess.PIPE)
print(rc.stdout.decode("ascii"))

Anaconda install on Windows is stuck on executing install.py

Update: Two months after posting this question, I switched to a fresh install of Windows 10 and was able to install the correct version of Anaconda Python. Unfortunately the answers came too late for me to test them.
I want to switch from the 32-bit version of Anaconda Python 3.5 to the 64-bit version. First I uninstalled Anaconda through the 'Programs and Features' window, then I downloaded the Anaconda3-4.2.0-Windows-x86_64.exe installer and attempted to run it. For reference, I am using the 64-bit version of Windows 7. After some initial processing the installer is stuck on Execute: "C:\Users\user.name\AppData\Local\Continuum\Anaconda3\pythonw.exe" -E -s "C:\Users\user.name\AppData\Local\Continuum\Anaconda3\pkgs\.install.py" --root-prefix "C:\Users\user.name\AppData\Local\Continuum\Anaconda3" --post root
It does not throw any errors, it just hangs. I let it run throughout the night to see if it would progress any further, but unfortunately it was still stuck in the same place in the morning. Afterwards I tried looking for hidden Continuum files from the previous installation and removed those - also removed Anaconda from my PATH variable. I even tried installing miniconda instead of anaconda, but both installers get stuck in the exact same place. Am I missing a hidden file somewhere which is causing the Anaconda installer to hang?
Below is part of the installation log up until the execution command where it gets stuck.
Installing: dill-0.2.5-py35_0 (into root)
untgz::extract -d 'C:\Users\user.name\AppData\Local\Continuum\Anaconda3' -zbz2 'C:\Users\use.name\AppData\Local\Continuum\Anaconda3\pkgs\dill-0.2.5-py35_0.tar.bz2'
Writing Lib/site-packages/dill-0.2.5-py3.5.egg-info
Writing Lib/site-packages/dill/__diff.py
Writing Lib/site-packages/dill/__init__.py
Writing Lib/site-packages/dill/__pycache__/__diff.cpython-35.pyc
Writing Lib/site-packages/dill/__pycache__/__init__.cpython-35.pyc
Writing Lib/site-packages/dill/__pycache__/_objects.cpython-35.pyc
Writing Lib/site-packages/dill/__pycache__/detect.cpython-35.pyc
Writing Lib/site-packages/dill/__pycache__/dill.cpython-35.pyc
Writing Lib/site-packages/dill/__pycache__/info.cpython-35.pyc
Writing Lib/site-packages/dill/__pycache__/objtypes.cpython-35.pyc
Writing Lib/site-packages/dill/__pycache__/pointers.cpython-35.pyc
Writing Lib/site-packages/dill/__pycache__/settings.cpython-35.pyc
Writing Lib/site-packages/dill/__pycache__/source.cpython-35.pyc
Writing Lib/site-packages/dill/__pycache__/temp.cpython-35.pyc
Writing Lib/site-packages/dill/_objects.py
Writing Lib/site-packages/dill/detect.py
Writing Lib/site-packages/dill/dill.py
Writing Lib/site-packages/dill/info.py
Writing Lib/site-packages/dill/objtypes.py
Writing Lib/site-packages/dill/pointers.py
Writing Lib/site-packages/dill/settings.py
Writing Lib/site-packages/dill/source.py
Writing Lib/site-packages/dill/temp.py
Writing Scripts/get_objgraph.py
Writing Scripts/unpickle.py
Writing info/LICENSE.txt
Writing info/files
Writing info/has_prefix
Writing info/index.json
Writing info/platform
Writing info/recipe/bld.bat
Writing info/recipe/build.sh
Writing info/recipe/meta.yaml
extraction complete.
Execute: "C:\Users\user.name\AppData\Local\Continuum\Anaconda3\pythonw.exe" -E -s "C:\Users\user.name\AppData\Local\Continuum\Anaconda3\pkgs\.install.py" --root-prefix "C:\Users\user.name\AppData\Local\Continuum\Anaconda3" --post root
I simply ended all python processes (any stale anaconda install/uninstall process running) and deleted all directories created during previous installation attempt. Then I could install without trouble.
If above effort does not work, check the environment path variables. It should not include any python path.
I have ended pythow process from taskmanager and removed unnecessary/irrelevant from environment path. It resolves my issues.
Try this, it worked for me:
conda install python=3.5
conda install tensorflow
Temporary disable antivirus and then try.
For me the installation was extremely slow, especially during the "unpacking" phase. After I disabled the "Real-time protection" under "Virus & threat protection" - the installation continued much faster.
I solved this issue by doing the following:
The installation was stuck on "extracting: anaconda-2020.02-py37_0.tar.bz2".
I went to https://anaconda.org/anaconda/anaconda/files and installed that file manually.
I extracted that file in the anaconda folder and it was all good!
As mentioned by ClaraJacintho in https://github.com/ContinuumIO/anaconda-issues/issues/6258#issuecomment-565063685:
I found that the "security module" called Warsaw (or GAS antifraud), used by many banks in Brazil for internet banking security, was interfering with the installation. I uninstalled it and was able to install Anaconda3.2019.10 64 bits on my Windows 10 machine.
It worked for me.

Resources