a mostly pseudo script of the gpu version of sigmoid runInput

gpu-implementation-comparison.js

//existing neural net
class NeuralNetwork {
  runInput(input) {
    this.outputs[0] = input;  // set output state of input layer

    let output = null;
    for (let layer = 1; layer <= this.outputLayer; layer++) {
      for (let node = 0; node < this.sizes[layer]; node++) {
        let weights = this.weights[layer][node];

        let sum = this.biases[layer][node];
        for (let k = 0; k < weights.length; k++) {
          sum += weights[k] * input[k];
        }
        this.outputs[layer][node] = 1 / (1 + Math.exp(-sum));
      }
      output = input = this.outputs[layer];
    }
    return output;
  }
}

//with gpu.js
class NeuralNetwork {
  makeLayers() {
    const layers = this.layers;
    const kernels = this.kernels;
    for (let i = 0; i < layers.length; i++) {
      const layer = layers[i];
      const kernel = gpu.createKernel(`function(input, weights, biases) {
        var sum = biases[this.thread.x];
        for (var i = 0; i < ${ weights.length }; i++) {
          sum += weights[i] * input[i];
        }
        return 1 / (1 + Math.exp(-sum));
      }`, {
        dimensions: [layer.length - 1]
      });
      kernel.build(new Array(this.size[i]), layer.weights, layer.biases);
      kernels.push(kernel);
    }
  }
  runInput(input) {
    this.outputs[0] = input;  // set output state of input layer

    let output = null;
    for (let layer = 1; layer <= this.outputLayer; layer++) {
      this.outputs[layer] = this.kernels[layer](input, this.weights[layer], this.biases[layer]);
      output = input = this.outputs[layer];
    }
    return output;
  }
}