note
roboticus
<p>Oh, something I forgot to mention: I tried using a constant number of bits but varying the vector size/quantity to see how things scaled. In other words, I compared:</p>
<table>
<tr><th>vec size</th><th># vectors</th></tr>
<tr align="right"><td>10,000 </td><td> 1</td></tr>
<tr align="right"><td> 5,000</td><td> 2</td></tr>
<tr align="right"><td> 3,333</td><td> 3</td></tr>
<tr align="right"><td> 2,500</td><td> 4</td></tr>
<tr align="right"><td> 2,000</td><td> 5</td></tr>
<tr align="right"><td> 1,000</td><td>10 </td></tr>
</table>
<p>I found more smaller vectors works better until the number of samples matches the number if bits in the smaller vector. Plotting the functions:</p>
<c>
(1-exp(-x/1000))^10
(1-exp(-x/2000))^5
(1-exp(-x/2500))^4
(1-exp(-x/3333))^3
1-exp(-x/10000)
</c>
<p>using a [http://www.coolmath.com/graphit/|graphing calculator] shows that's where the curves cross:</p>
<p>...[roboticus]</p>
<p><i>When your only tool is a hammer, all problems look like your thumb.</i></p>
962802
963012