Quote:
Originally Posted by Koko
I timed 31 searches on the kindle touch (KT) and 31 searches on kindle dx (kdx) using a manual stopwatch. The words used where 'tante', 'apert', 'acrocephaly', 'myopia', 'amblyopia' and 'retina'. For each search a new word from the list was used.
Results: Using an paired sample t-test the results could not be said to be stastistically different (sig. 2 tailed=0,102) within a 95% confidence interval. The mean time for KT was 2,04 seconds and for KDX 2,32 seconds.
Here are the data:
|
You could preserve a tabular format if you use a CODE block rather than a QUOTE block:
Code:
-------------------------------------------------
| Column 1 | Column 2 | Column 3 | Column 4 |
| measurement | word | device | time |
| number | | | |
|-------------+-------------+----------+----------|
| 1 | tante | 0 | 2.88 |
| 2 | tante | 1 | 1.98 |
| 3 | apert | 0 | 2.14 |
| 4 | apert | 1 | 2.29 |
| 5 | acrocephaly | 0 | 1.58 |
| 6 | acrocephaly | 1 | 1.85 |
| 7 | myopia | 0 | 1.51 |
| 8 | myopia | 1 | 1.87 |
| 9 | amblyopia | 0 | 1.71 |
| 10 | amblyopia | 1 | 1.86 |
| 11 | tante | 0 | 1.51 |
| 12 | tante | 1 | 2.13 |
| 13 | apert | 0 | 2.14 |
| 14 | apert | 1 | 2.05 |
| 15 | acrocephaly | 0 | 2.02 |
| 16 | acrocephaly | 1 | 1.93 |
| 17 | myopia | 0 | 1.58 |
| 18 | myopia | 1 | 1.95 |
| 19 | amblyopia | 0 | 1.53 |
| 20 | amblyopia | 1 | 1.93 |
| 21 | tante | 0 | 1.57 |
| 22 | tante | 1 | 1.89 |
| 23 | apert | 0 | 1.32 |
| 24 | apert | 1 | 2.00 |
| 25 | acrocephaly | 0 | 1.73 |
| 26 | acrocephaly | 1 | 1.90 |
| 27 | myopia | 0 | 1.45 |
| 28 | myopia | 1 | 2.01 |
| 29 | amblyopia | 0 | 1.42 |
| 30 | amblyopia | 1 | 1.88 |
| 31 | tante | 0 | 1.43 |
| 32 | tante | 1 | 1.90 |
| 33 | apert | 0 | 1.30 |
| 34 | apert | 1 | 1.91 |
| 35 | acrocephaly | 0 | 1.39 |
| 36 | acrocephaly | 1 | 1.91 |
| 37 | myopia | 0 | 2.80 |
| 38 | myopia | 1 | 1.92 |
| 39 | amblyopia | 0 | 2.20 |
| 40 | amblyopia | 1 | 2.06 |
| 41 | tante | 0 | 2.50 |
| 42 | tante | 1 | 1.94 |
| 43 | apert | 0 | 2.20 |
| 44 | apert | 1 | 1.93 |
| 45 | acrocephaly | 0 | 2.09 |
| 46 | acrocephaly | 1 | 1.80 |
| 47 | myopia | 0 | 2.15 |
| 48 | myopia | 1 | 1.91 |
| 49 | amblyopia | 0 | 4.10 |
| 50 | amblyopia | 1 | 3.18 |
| 51 | tante | 0 | 2.32 |
| 52 | tante | 1 | 1.93 |
| 53 | apert | 0 | 2.39 |
| 54 | apert | 1 | 1.93 |
| 55 | acrocephaly | 0 | 2.48 |
| 56 | acrocephaly | 1 | 1.86 |
| 57 | myopia | 0 | 2.44 |
| 58 | myopia | 1 | 1.38 |
| 59 | amblyopia | 0 | 2.84 |
| 60 | amblyopia | 1 | 1.99 |
| 61 | retina | 0 | 2.73 |
| 62 | retina | 1 | 4.02 |
-------------------------------------------------
Also, because this is an English language forum, I "normalized" the decimal points to English (US/UK) format, and I added leading spaces and trailing zeros as needed for uniform consistent numeric formatting.
EDIT: By the way, search speed is not only dependent on CPU speed, memory speed, eMMC flash speed, and firmware version differences, but can be affected by number of background processes running, by services installed, by whether wireless is on or off, and by file fragmentation. Do both computers have the same hacks installed and active? And because of the firmware-controlled CPU clock speed governor, search speed could also be affected by charging status of the battery, and whether it is connected to a host PC or a wall charger. You need to control all of those factors to yield scientifically and statistically valid test results.
EDIT2: Your interpretation of the test results shows that despite hardware and firmware differences, the bottleneck is probably the choice of search algorithm and indexing method, and the speed of the eMMC (USB storage partition).