Performance question

[dy]

Hi @mikolalysenko @hughsk!
There is another solution from here:

float shift_right (float v, float amt) { 
    v = floor(v) + 0.5; 
    return floor(v / exp2(amt)); 
}
float shift_left (float v, float amt) { 
    return floor(v * exp2(amt) + 0.5); 
}
float mask_last (float v, float bits) { 
    return mod(v, shift_left(1.0, bits)); 
}
float extract_bits (float num, float from, float to) { 
    from = floor(from + 0.5); to = floor(to + 0.5); 
    return mask_last(shift_right(num, from), to - from); 
}
vec4 encode_float (float val) { 
    if (val == 0.0) return vec4(0, 0, 0, 0); 
    float sign = val > 0.0 ? 0.0 : 1.0; 
    val = abs(val); 
    float exponent = floor(log2(val)); 
    float biased_exponent = exponent + 127.0; 
    float fraction = ((val / exp2(exponent)) - 1.0) * 8388608.0; 
    float t = biased_exponent / 2.0; 
    float last_bit_of_biased_exponent = fract(t) * 2.0; 
    float remaining_bits_of_biased_exponent = floor(t); 
    float byte4 = extract_bits(fraction, 0.0, 8.0) / 255.0; 
    float byte3 = extract_bits(fraction, 8.0, 16.0) / 255.0; 
    float byte2 = (last_bit_of_biased_exponent * 128.0 + extract_bits(fraction, 16.0, 23.0)) / 255.0; 
    float byte1 = (sign * 128.0 + remaining_bits_of_biased_exponent) / 255.0; 
    return vec4(byte4, byte3, byte2, byte1); 
}

The benefit is instant conversion of result to Float32Array:

var result = new Uint8Array(w * 4);
gl.readPixels(0, 0, w, 1, gl.RGBA, gl.UNSIGNED_BYTE, result);
result = new Float32Array(result.buffer); //← this guy is in theory very fast, instead of converting per-element

Just wanted to know your thoughts on that, because there are gl-texture2d-read-float and others depending on that.

[dy]

Tiny benchmark has shown that there is no difference.
I guess it would be easier for a user to get float32array in return, if uint8array is passed as an argument:

var floatArray = unpackFloat(buffer);

@mikolalysenko what do you think?

[mikolalysenko]

This module also allows for instant conversion to a float32array. Look at how the decoding works: https://github.com/mikolalysenko/glsl-read-float/blob/master/index.js#L1-L12