Single Instruction Multiple Data - (SIMD) Java Example

import jdk.incubator.vector.*;
import java.util.Arrays;

public class VectorAdditionSIMD {

    private static final VectorSpecies<Double> SPECIES = DoubleVector.SPECIES_PREFERRED;
    private static final int VECTOR_SIZE = SPECIES.length();

    public static double[] addVectorsSIMD(double[] vector1, double[] vector2) {
        if (vector1 == null || vector2 == null) {
            throw new IllegalArgumentException("Vectors cannot be null.");
        }
        if (vector1.length != vector2.length) {
            throw new IllegalArgumentException("Vectors must have the same dimensions.");
        }

        double[] result = new double[vector1.length];

        int i = 0;
        // Vectorized loop
        for (; i <= vector1.length - VECTOR_SIZE; i += VECTOR_SIZE) {
            DoubleVector v1 = DoubleVector.fromArray(SPECIES, vector1, i);
            DoubleVector v2 = DoubleVector.fromArray(SPECIES, vector2, i);
            DoubleVector sum = v1.add(v2);
            sum.intoArray(result, i);
        }

        // Scalar loop for remaining elements
        for (; i < vector1.length; i++) {
            result[i] = vector1[i] + vector2[i];
        }

        return result;
    }

    public static void main(String[] args) {
        // Example Usage
        double[] v1 = new double[100];
        double[] v2 = new double[100];
        double[] v3 = new double[7];

        for(int i = 0; i < 100; i++){
            v1[i] = i * 1.0;
            v2[i] = i * 2.0;
        }

        for(int i = 0; i < 7; i++){
            v3[i] = i * 3.0;
        }

        try {
            double[] sum = addVectorsSIMD(v1, v2);
            System.out.println("Vector 1: " + Arrays.toString(Arrays.copyOf(v1,10)));
            System.out.println("Vector 2: " + Arrays.toString(Arrays.copyOf(v2,10)));
            System.out.println("SIMD Sum: " + Arrays.toString(Arrays.copyOf(sum, 10)));

            double[] sum2 = addVectorsSIMD(v1, v3);

        } catch (IllegalArgumentException e) {
            System.err.println("Error: " + e.getMessage());
        }

    }
}

To verify, first install hsdis you can get a prebuilt binary (e.g. chriswhocodes.com/hsdis). Copy the shared library or dll to the appropriate location (e.g. on Linux copy hsdis-amd64.so to /usr/lib/jvm/java-21-openjdk-amd64/lib/server/) hsdis is the hotspot disassembler so we can see what instructions (and compiler we need c2/server not c1) is utilized.

Compile:

javac --add-modules jdk.incubator.vector VectorAdditionSIMD.java

Run/examine output:

java -XX:+UnlockDiagnosticVMOptions -XX:+PrintAssembly -XX:CompileThreshold=1 -XX:-TieredCompilation --add-modules jdk.incubator.vector VectorAdditionSIMD > c2output.txt

Now within c2output.txt you should find that the C2 compiler was utilized, and furthermore that you are using ymm (256bit) registers, this is a good sign that you are executing via SIMD.

If all is well you should see output along the lines of:

  0x00007f345ca993ca:   vmovdqu (%rdi,%rcx,1),%ymm0
  0x00007f345ca993cf:   vmovdqu (%rsi,%rcx,1),%ymm1
  0x00007f345ca993d4:   vpxor  %ymm1,%ymm0,%ymm0
  0x00007f345ca993d8:   vptest %ymm0,%ymm0
  0x00007f345ca993f1:   vmovdqu -0x20(%rdi,%rax,1),%ymm0
  0x00007f345ca993f7:   vmovdqu -0x20(%rsi,%rax,1),%ymm1
  0x00007f345ca993fd:   vpxor  %ymm1,%ymm0,%ymm0
  

Programmatically Determine if On Battery or Plugged In via Powershell

function onBattery { 
    if ((Get-WmiObject -Class Win32_Battery | Select-Object -ExpandProperty batterystatus) -eq 2) {
        write-output "Plugged In" 
    }
    else { 
        write-output "On Battery" 
    }
}

Toggle Controlled Folder Access in Programmatically in Powershell

function update-controlledfolderaccess { 
  param(
    [Parameter(Mandatory = $false, Position = 0)]
    [ValidateSet("toggle", "enable", "disable")]
    [String]$Action
  )

  $setCFAcmd = "Set-MpPreference -EnableControlledFolderAccess " 
  $setCFAaction = $null 
  $ctlfldracc_enabled = $false 

  if ($Action -eq "toggle") {  
    $Preferences = Get-MpPreference
    $ctlfldracc_enabled = $Preferences.EnableControlledFolderAccess
    if ($ctlfldracc_enabled) { $setCFAaction = "disable" } else { $setCFAaction = "enable" } 
  }
  else {
    $setCFAaction = $Action
  }

  $add2profile_process = Start-Process powershell -Verb RunAs "-Command", ($setCFAcmd += $setCFAaction) -wait -PassThru 
  
  if ($add2profile_process.ExitCode -eq 0) {
    Write-Host (">> Successfully {0}d EnableControlledFolderAccess <<" -f $setCFAaction) -ForegroundColor Green
  }
  else {
    throw ("Couldn't {0} EnableControlledFolderAccess .." -f $setCFAaction)
  }
}

AutoHotKey Script - Change Windows Paths to Linux/WSL On Paste

If you use a hybrid Windows/WSL environment (like I do) you probably have run into the situation where you copy a path from a Windows program and want to paste it into a Linux program (like a WSL commandline).

The following script does the job and quotes it in case there's spaces.

To use it:

• Install AHK
• Copy the following script to an AHK extension and double click it
• Copy a Windows path (e.g. from Windows Explorer) and paste into a Linux client (e.g. WSL) with Ctl+Alt+v

; Created by Adam Danischewski Copyrighted 2024
; Any use permissible, no warranty expressed or implied.  
; AHK Script - Convert Windows path to WSL path on Ctrl+Alt+V
^!v::
    ; Get clipboard content
    path := A_Clipboard

    ; Ensure path starts with a drive letter
    if RegExMatch(path, "^[A-Z]:", match) {
        ; Replace drive letter with "/mnt/" prefix
        path := StrReplace(path, match, "/mnt/" Format("{:L}",match), 1)  ; Limit replacements to 1
        path := StrReplace(path, ":", "")
    }

    ; Replace backslashes with forward slashes
    path := StrReplace(path, "\", "/")

    ; Wrap the path in quotes and send it
    SendInput % """" path """"
    return

AWK - Fast Urlencoder - Handles Multibytes

#!/bin/bash
function urlencode() { 
LC_ALL=C awk 'BEGIN {
  for (i = 0; i <= 255; i++)
    hex[sprintf("%c", i)] = sprintf("%%%02X", i)
  safe = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-_.:/!~*@$&[]+=,;?"
  split(safe, safechars, "")
  for (i in safechars)
    hex[safechars[i]] = safechars[i]
}
{
  encoded = ""
  for (i = 1; i <= length($0); i++)
    encoded = encoded hex[substr($0, i, 1)]
  print encoded
}
'
}

 This little bad boy rivals the speed of my C implementation below! 
 Slower by less than 10%.
 
 [Btw: this example is wrapped in a Bash function for ease of use]

Small Fast URL Encoder - vx_urlenc

// SPDX-License-Identifier: Apache-2.0
/***********************************************************
 **********************************************************
 
 * Fast URL Encoder - Closely conforms to RFC 3986
 *
 * Takes piped in uri's similar to jq but is quicker and conforms 
 * more to RFC 3986 [diverges w/most browsers on #'s]. 
 *
 * Handles multibyte unicode characters 
 *  
 * Copyright (C) 2024 Victrixsoft
 *
 * Issues: If you find any issues please open a ticket on github!
 *         https://github.com/victrixsoft/vx_urlenc/issues 
 * 
 * Author: Adam Danischewski <my first nm(dot)my last nm@gmail.com>
 * 
 **********************************************************
 ***********************************************************/

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#ifdef _WIN32
#include <io.h>
#define access _access
#else
#include <unistd.h>
#endif

#define ASCII_SIZE 128

char *url_encode(const char *);
void init_lookup_table(void);
int fileno(FILE *);
void __attribute__((constructor)) init(void);

char lookup_table[ASCII_SIZE] = {0};

void init_lookup_table(void) {
  for (int i = 'a'; i <= 'z'; i++)
    lookup_table[i] = 1;
  for (int i = 'A'; i <= 'Z'; i++)
    lookup_table[i] = 1;
  for (int i = '0'; i <= '9'; i++)
    lookup_table[i] = 1;
  const char *safe = "-_.:/!~*@$&[]+=,;?";
  while (*safe)
    lookup_table[(unsigned char)*safe++] = 1;
}

void __attribute__((constructor)) init() { init_lookup_table(); }

char *url_encode(const char *str) {
  if (str == NULL)
    return NULL;

  char *encoded =
      malloc(strlen(str) * 4 * 3 + 1); // Worst case: every char 4 byte mb
  if (encoded == NULL)
    return NULL;

  char *penc = encoded;
  size_t output_idx = 0;
  while (*str && *str != '\n') { // Stop at newline
    if (lookup_table[(unsigned char)*str]) {
      *penc++ = *str;
      output_idx++;
    } else {
      sprintf(penc, "%%%02X", (unsigned char)*str);
      penc += 3;
      output_idx += 3;
    }
    str++;
  }
  *penc = '\0';
  char *resized_enc = realloc(encoded, output_idx + 1);
  if (resized_enc == NULL) {
    return encoded;
  } else {
    return resized_enc;
  }
}

int main(int argc, char *argv[]) {
  (void)argc;
  (void)argv;

  if (isatty(fileno(stdin))) {
    perror("It's not a pipe\n");
    return 1;
  }

  char strbuf[65536];
  while (fgets(strbuf, sizeof strbuf, stdin) != NULL) {
    char *encoded = url_encode(strbuf);
    if (encoded) {
      printf("%s\n", encoded);
      free(encoded);
    } else {
      fprintf(stderr, "Error encoding line\n");
    }
  }

  return 0;
}

 Compile: gcc -o vx_urlenc vx_urlenc.c
   Usage: echo 'http://www.gaggle.com/🐀🐁🐂🐃🐄🐅🐆🐇🐈🐉🐊🐋🐌🐍!!![]*?#@.mp4' | ./vx_urlenc
 

https://github.com/victrixsoft/vx_urlenc

Unicode Codepoint to Hex - The Bash Way

function cp2hex() { 
 local codepoint="${1#U}"
 local -i bytes=0
 codepoint=${codepoint^^}
 local binary=$(echo "ibase=16;obase=2;${codepoint}"|bc)
 local -a binary_arr
  if ((0x${codepoint} <= 0x007F)); then
  bytes=1
  binary=$(printf "%08d" $binary)
  read -d=' ' -a binary_arr < <(sed 's/./ &/g' <<< "$binary")
  elif ((0x${codepoint} <= 0x07FF)); then
   bytes=2
   binary=$(printf "%011d" $binary)
   read -d=' ' -a binary_arr < <(sed 's/./ &/g' <<< "$binary")
   binary_arr[0]="110${binary_arr[0]}"
   binary_arr[5]="10${binary_arr[5]}"
  elif ((0x${codepoint} <= 0xFFFF)); then
   bytes=3
   binary=$(printf "%016d" $binary)
   read -d=' ' -a binary_arr < <(sed 's/./ &/g' <<< "$binary")
   binary_arr[0]="1110${binary_arr[0]}"
   binary_arr[4]="10${binary_arr[4]}"
   binary_arr[10]="10${binary_arr[10]}"
  elif ((0x${codepoint} <= 0x10FFFF)); then
   bytes=4
   binary=$(printf "%021d" $binary)
   read -d=' ' -a binary_arr < <(sed 's/./ &/g' <<< "$binary")
   binary_arr[0]="11110${binary_arr[0]}"
   binary_arr[3]="10${binary_arr[3]}"
   binary_arr[9]="10${binary_arr[9]}"
   binary_arr[15]="10${binary_arr[15]}"
  fi 
  printf "${bytes} Byte unicode: \U${codepoint}\n"
  echo "Codepoint: U${codepoint}"
  echo "Codepoint binary: $binary"
  binary=$(sed 's/ //g' <<< "${binary_arr[@]}")
  echo "UTF-8 binary: ${binary}"
  echo "UTF-8 hex: $(printf "%X" $((2#${binary})))"
}

$>cp2hex U0024
1 Byte unicode: $
Codepoint: U0024
Codepoint binary: 00100100
UTF-8 binary: 00100100
UTF-8 hex: 24

$>cp2hex U00A2
2 Byte unicode: ¢
Codepoint: U00A2
Codepoint binary: 00010100010
UTF-8 binary: 1100001010100010
UTF-8 hex: C2A2

$>cp2hex 20AC
3 Byte unicode: €
Codepoint: U20AC
Codepoint binary: 0010000010101100
UTF-8 binary: 111000101000001010101100
UTF-8 hex: E282AC

$>cp2hex U1F604
4 Byte unicode: 😄
Codepoint: U1F604
Codepoint binary: 000011111011000000100
UTF-8 binary: 11110000100111111001100010000100
UTF-8 hex: F09F9884

Powershell Function to List Ports in Use by a Java Prog

Replace"YourCurrentProject" default with your most frequent query.

Add the following to your $profile (PS> code $profile).

function lsAppPorts($prog = "YourCurrentProject") {
    foreach($a in jps -l | C:\Progra~2\GnuWin32\bin\awk "/$prog/{ print `$1; }") {
        netstat -ano | findstr $a
    }
}

Note: Requires Gnu Awk for Windows

Generate Tough Passwords From the Commandline

Note: Requires openssl
Usage: $> genpasswd [int size, default: 8]

alias genpasswd='_(){ if (($#<1)); then PASSWDSIZE=8; else PASSWDSIZE=${1};fi; openssl rand -base64 32|sed -r "s/^(.{${PASSWDSIZE}})(.*)/\1/g"; }; _'

Custom SpringBoot Serializer for REST Controller - Fairly Elegant Way

SpringBoot is great for speed of development when you are familiar with it. Modern development often is centered on microservices which is heavily reliant upon a REST controller that returns JSON formatted data.

At some point there is a high chance that you'll need to customize the Serialization of data returned from the REST controller.

At first glance it seems straightforward enough, just define your custom serializer and reference it within in the @JsonSerializer annotation: @JsonSerialize(using = MyCustomSerializer.class)

If you do not register your custom Serializer you will likely get a 500 error stating that a serializer could not be found.

You must also make sure that you define a handledType() in your custom Serializer. If not you will see a message along the lines of:
... does not define valid handledType()

Furthermore you cannot use @JsonSerializer(using = ...) in your REST controller where your Serializer is defined, for that you need a custom annotation.

Most tutorials online I've found have circumvented these steps and provided kludgey solutions reliant on creation of another class - unregistered serializers are found when a wrapper class for instance is utilized defined within the REST controller.

Of course, you would probably put your custom Serializer somewhere outside of your REST controller, but for simplicity sake this example has it all in one.

import ... 
...
@RestController
@Component // Make sure our @Bean is found 
... 

////
// Define our Serializer with handler 
////
public class MyLongSerializer extends JsonSerializer<Long> {
    private static final long serialVersionUID = -8885817712043352432L;

    @Override
    public void serialize(Long value, JsonGenerator jsonGenerator, SerializerProvider serializerProvider)
            throws IOException {
    	jsonGenerator.writeStartObject();
        jsonGenerator.writeNumberField("my val", value);
        jsonGenerator.writeEndObject();
    }
    
    @Override
    public Class<Long> handledType() {
	return Long.class;
    }

}

////
// Define our custom Serializer annotation - @MyLong
////
@Retention(RetentionPolicy.RUNTIME)
@JacksonAnnotationsInside
@JsonInclude(value = Include.NON_NULL)
@JsonSerialize(using = MyLongSerializer.class)
public @interface MyLong {
}

////
// Define our REST controller test 
////
@RequestMapping(value = { "/testLong" }, method = RequestMethod.GET, produces = {
		MediaType.APPLICATION_JSON_VALUE }, consumes = MediaType.ALL_VALUE)
public Long testLong() { 
    @MyLong
    Long myLong = 293874928749284L;
	return myLong;
}

////
// Add our dear custom serializers
////
@Bean
public Jackson2ObjectMapperBuilder jackson2ObjectMapperBuilder() {
    SimpleModule m = new SimpleModule();
    m.addSerializer(new MyLongSerializer());
    // Add here as many as you have: 
     // m.addSerializer(...); ... 
    return new Jackson2ObjectMapperBuilder().modules(m);
}	

Hit your url: http://.../testLong and you should see:
{
"my val": 293874928749284
}