# GPU parallel processing with APARAPI
**URL:** https://discourse.processing.org/t/gpu-parallel-processing-with-aparapi/46223
**Category:** Libraries
**Created:** [April 15, 2025, 3:48pm UTC](https://discourse.processing.org/t/gpu-parallel-processing-with-aparapi/46223 "2025-04-15T15:48:18Z")
**Posts on this page:** 1
**Page:** 1
### Author:  [@williamfraser](https://discourse.processing.org/u/williamfraser)
#### Post date: [April 15, 2025, 3:48pm UTC](https://discourse.processing.org/t/gpu-parallel-processing-with-aparapi/46223/1 "2025-04-15T15:48:18Z")
Good day all.
I have written a particle simulator and would like to try improving speed using parallel processing on a GPU.
APARAPI is something used to run Java on GPU, but will it work with Processing?
I have no experience with Java or any of the technical stuff that happens under the hood with Processing. Below is a link to the APARAPI user guide. I have implemented threads in my code for multithreading on the CPU so I understand there is some similarity between launching new threads and launching new kernels, but I have no idea what to make of the rest.
>
For what it is worth, I tried out a simple sketch with elements from the “Squares” example that comes with the download. I somehow managed to get it running but only after doing the following:
In the aparapi zip file “dist\_windows\_x86\_64” I found the files “aparapi.jar” and “aparapi\_x86\_64.dll”. I placed a copy of the dll file in Processing-4.3 \> core \> library \> windows-amd64.
I made a folder called “code” in the same folder as the sketch and copied the jar file into that folder.
```auto
import com.amd.aparapi.Kernel;
import com.amd.aparapi.Range;
int numBalls = 100;
float dx = 0.2;
float balldiameter = 20;
float[] values = new float[numBalls];
Ball[] balls = new Ball[numBalls];
void setup() {
size(1500, 300);
for (int i = 0; i < numBalls; i++) {
values[i] = i;
balls[i] = new Ball(20*i, 20);
}
}
void draw() {
background(205);
kernel.execute(Range.create(numBalls));
for (int i=0; i