Take a look back through 2015

[-hh]

> Mark Wie. wrote:
> But "3.5x Faster" than what?
> It's like those ads that feature "20% less fat"

:-)

But: If oldWeight is 3.5*newWeight then newWeight is 1/3.5 of oldWeight what is
71.43% less fat (or whatever).

> Richard G. wrote:
> I dunno, the performance boost seems reasonably well reflected in my 
> most recent benchmarks.

Honestly, how do you conclude the factor 3.5 from these results? Must be an exact calculation because of the decimal. Kind of special weighted geometric mean?

Here are your results in one list, if anybody would like to do his own computations.

-- item1 = 6.7.1-rc2, 2 = 7.0.1-rc2, 3=8.0.0-dp11
Test_ArrayAccess: 19,48,27
Test_ArraySplit: 35,75,57
Test_BuildFilePath: 0,22,14 <-- Take here 1 instead of 0
Test_BuildList: 72,197,98
Test_DecodeArray: 46,79,63
Test_EncodeArray: 9,58,42
Test_FileBinRead: 1,14,14
Test_FileBinWrite: 2,80,78
Test_FileTextRead: 1,156,77
Test_FileTextWrite: 2,96,94
Test_LineAccessByNumber: 518,4684,319
Test_LineAccessForEach: 8,41,17
Test_Lineoffset: 2544,7568,2521
Test_Merge: 220,620,349

[With equal weights one would have:
The average factor of 7.0.1/8.0.0 is around 1.8
(that is "45% less fat" would be comparable)
the average factor of 6.7.1/8.0.0 is around 0.2
(that is "500% more fat" would be comparable)]