ChildOSX 2008 vs. Previous 2006 Test

3d version of images shown below. Click for larger image.

Should the following image be alarming?:

This is a comparison of the ChildOSX (2/10/08) and previous (8/17/06) test_P4_mb1_nb1_1 outputs for 3 time segments: {3000000, 3005000, 3010000}, from the files .nodes, .edges, and .q. Only (x, y) data is plotted. Red is edges, yellow are nodes having a discharge >= 1.6e9 (CRITICAL_FLOW in the input file).

The reason for this appears to be instability due to input precision loss (otherwise known as 'chaos'). I reran the 2/10/08 inputs and got the exact same results in time segment 2 (to all decimal places in .nodes, .z, and .q):

111: ~/Projects/Lancaster/Child/Child_CU_STL_2006/OSX_test > diff Outputs021008/test.nodes Outputs021308/test.nodes
112: ~/Projects/Lancaster/Child/Child_CU_STL_2006/OSX_test > diff Outputs021008/test.z Outputs021308/test.z
113: ~/Projects/Lancaster/Child/Child_CU_STL_2006/OSX_test > diff Outputs021008/test.q Outputs021308/test.q
114: ~/Projects/Lancaster/Child/Child_CU_STL_2006/OSX_test > diff Outputs021008/test.inputs Outputs021308/test.inputs
1c1
< # Created by CHILD on Sun Feb 10 13:36:18 2008.
---
> # Created by CHILD on Wed Feb 13 16:22:54 2008.

I then truncated the .z input to 3 decimal places (it was 9-12) and reran:

115: ~/Projects/Lancaster/Child/Child_CU_STL_2006/OSX_test > head Outputs021008/test.z
 0
2602
165.75439229
134.242538713
109.532273688
131.702889786
172.057538389
95.6048508439
143.272007469
109.344034121
116: ~/Projects/Lancaster/Child/Child_CU_STL_2006/OSX_test > head Outputs/test.z
 0
2602
165.754
134.243
109.532
131.703
172.058
95.605
143.272
109.344

The results are somewhat different. Here is a comparison of the two previous 'medium precision' runs and the current 'low precision' run:

It would appear that the full precision of the 8/17/06 run was truncated in the text files, and is the reason why the current runs using those text files as input differ. If possible, I would like you to try the 'medium precision' inputs (available from the 02/10/08 page) and see if you get the same (and different from 8/17/06) results I do.