Michael J Mathew mjm522

When your program runs, do the first instruction. If it is causing segmentation fault, put a while true or cin in the start of main loop, execute it. So it will be caught up in the while loop or may be waiting for the input. When it happens, run the first instruction to get the name of your executable.

First open a terminal and then type:

ps ax | grep <name of your executable>

This would list the complete path to the executable with any arguments present.

Note: To run a process dedicated on a CPU you can use the tasket command. For example

taskset -c 5,11 python -m timeit 'sum(range(10**7))'

However, this will not guarantee that the CPUs 5,11 will be used for that process alone. The same CPU can be interrupted by the scheduler and may not be properly isolated. So inorder to isolate the CPU the following steps are to be taken. Another feature that could be turned off for these CPUs are the interrupt feature 'IRQ'. However this is not covered here.

List all available cores and hyperthreads in your processor.

lscpu --all --extended

Steps

Install docker using instructions here. This installation instructions were tested on Ubuntu 18.04 with Docker version 19.03.3, build a872fc2f86.
To run docker as a non-root user, follow the post installation instruction here.
Install nvidia toolkit:Reference

$ sudo apt-get update && sudo apt-get install -y nvidia-container-toolkit

	function compress_pdf()
	{
	input_file=$1
	output_file=$1_compressed.pdf
	level=/$2
	#levels: screen > ebook > printer > prepress,
	gs -dNOPAUSE -dBATCH -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dPDFSETTINGS=$level -sOutputFile=$output_file $input_file
	echo "Saved in the same folder with name $output_file"
	}

	import numpy as np
	from multiprocessing import Queue, Process

	def worker(shared_queue):
	count = 0
	temp = np.zeros(6)
	while True:
	count += 1
	temp[-1] = count
	shared_queue.put(temp)

	import time
	import os, stat
	import numpy as np
	from multiprocessing import Process

	pipename = 'myfifo'
	if not stat.S_ISFIFO(os.stat(pipename).st_mode):
	os.mkfifo(pipename)

	def worker(pipe_name):

	import numpy as np
	from multiprocessing import Process

	def worker(array_file):
	count = 0
	temp = np.zeros(6)
	while True:
	count += 1
	temp[-1] = count
	array_file[:] = temp[:]

	import numpy as np
	from multiprocessing import shared_memory, Process

	def worker(shared_variable):
	existing_shm = shared_memory.SharedMemory(name=shared_variable.name)
	myarray = np.ndarray((6,), dtype=np.int64, buffer=existing_shm.buf)
	count = 0
	while True:
	count += 1
	myarray[-1] = count

	import sys
	import argparse
	import importlib

	def import_function_handle(config_file):
	file_path = [ROOT DIRECTORY WHERE YOUR PYTHON FILE EXIST]/{}.py'
	file_path = file_path.format(config_file)
	if not os.path.isfile(file_path):
	print("\033[31;40mSuch a config file does not exist, check name ...\033[m")
	print("\033[31;40mCurrent specified path: {}\033[m".format(file_path))

	Add this function to the `~/.bashrc` file or keep as a shell script file.

	```
	function combine_pdfs()
	{
	output_file=$1
	files=($2)
	gs -dBATCH -dNOPAUSE -q -sDEVICE=pdfwrite -dAutoRotatePages=/None -sOutputFile=$output_file "${files[@]}"
	}
	```