I'm completely new to OpenCL and decided to experiment with it using the code provided on a website. I made some changes to the code and am attempting to send a 4x4 matrix filled with all 1s to the kernel and retrieve it back. It may seem like a simple task, but it's important for me to grasp the workings of OpenCL. Here is the input matrix:
[[ 1. 1. 1. 1.]
[ 1. 1. 1. 1.]
[ 1. 1. 1. 1.]
[ 1. 1. 1. 1.]]
However, the output from the kernel is different from the input as shown below:
[[ 1. 1. 1. 1.]
[ 0. 0. 0. 0.]
[ 0. 0. 0. 0.]
[ 0. 0. 0. 0.]]
This is the complete code that I've been working on:
import pyopencl as cl
from pyopencl import array
import numpy as np
## Step #1. Obtain an OpenCL platform.
platform = cl.get_platforms()[0]
##... (code continues)
print(matrix_dot_vector)
The issue I am facing is that the value of int y = get_global_id(1); always seems to be 0, which leads to this discrepancy in results. I can't figure out why this happens, especially since I pass the correct shape to the kernel function. Any ideas on what could be wrong?
Thank you!