scivision

There are a few node types on BU SCC that are designed for MPI jobs. The job script syntax can be used for interactive session start.

qrsh -pe mpi_28_tasks_per_node 56

This becomes an issues with MPI workers with synchronized parallelism. There are "14 core" laptops with two P-cores and 12 E-cores. I have seen project teams that configure their programs (say via enviornment variables) to only use the P-cores, but this can lead to a performance drop vs. their much older all P-core computer.

fastbuild/fastbuild#995

Configure GNU Octave JVM

GNU Octave JVM can be configured with the JAVA_HOME environment variable. Some install packages don't include Java.

For example, with Homebrew:

GitHub Action to SSH into CI run

Just signal boosting this Action action-tmate that allows SSHing into a remote GitHub Actions server to help debug.

	cmake_minimum_required(VERSION 3.20)

	project(mini LANGUAGES Fortran)

	message(STATUS "CMAKE Generator: ${CMAKE_GENERATOR}")

	add_executable(mini mini.F)

	Notes on WSJT-X Installation for Mac OS X
	-----------------------------------------
	First just drop the wsjtx.app onto the Applications folder in the dmg window to
	install to Applications

	If you have already downloaded a previous version of WSJT-X then I suggest
	you change the name in the Applications folder from WSJT-X to WSJT-X_previous
	before proceeding.

	I recommend that you follow the installation instructions especially if you

	//Copyright Kevin Lawler, released under ISC License
	// https://getkerf.wordpress.com/2016/03/30/the-best-algorithm-no-one-knows-about/

	double random_double() //your random double function
	{
	//[0,1) uniformly random double, mt19937-64.c source is fine
	//keep in mind most double-based implementations drop randomness to 52-bits
	return genrand64_real2();
	}

	#!/usr/bin/env -S uv run --script
	# /// script
	# requires-python = ">=3.9"
	# dependencies = ["requests"]
	# ///

	from __future__ import annotations
	import requests

	1: \| Input Size (bytes) \| std::find (GB/s) \| simdutf::find (GB/s) \| naive_find (GB/s) \|
	1: \|--------------------\|------------------\|-----------------------\|-------------------\|
	1: \| 1024 \| 4.56 \| 26.21 \| 2.74 \|
	1: \| 8192 \| 4.61 \| 59.72 \| 2.70 \|
	1: \| 65536 \| 4.75 \| 54.45 \| 2.68 \|
	1: \| 524288 \| 4.77 \| 55.61 \| 2.69 \|
	1: \| 2097152 \| 4.78 \| 26.14 \| 2.69 \|
	1/2 Test #1: BenchmarkNative .................. Passed 37.91 sec

	2: \| Input Size (bytes) \| std::find (GB/s) \| simdutf::find (GB/s) \| naive_find (GB/s) \|

	[flake8]
	max-line-length = 100