The performance advantage of OMX compared to more traditional ASCII-based formats will vary significantly between different models and PC hardware used. To illustrate the potential benefits of OMX, a typical forecast ‘large’ sized UK-based model dataset was converted to/from SATURN .UFM format and the total number of files created, the total read/write time and total file size recorded as shown below in Table 10.1. The test model consisted of 2,358 zones, four time periods, five user classes and two scenarios.
Table 10.1 – Indicative Performance using different Matrix Formats
|
Measure |
TUBA Format 2 |
CSV |
OMX |
UFM | |
|
(ASCII) |
(ASCII) |
(Binary) |
(Binary) | ||
|
No. of Files |
400 |
400 |
80 |
80 | |
|
Time Taken (mins) |
Export |
59 |
28 |
6 |
|
|
Import |
517 |
9 |
2 |
| |
|
Total Time |
576 |
36 |
8 |
| |
|
Time Ratio |
71.1 |
4.5 |
1.0 |
| |
|
Total Size (Gb) |
50.0 |
|
5.1 |
6.4 | |
|
Size Ratio |
9.9 |
3.2 |
1.0 |
| |
Source: HP Compaq 8200 Elite, Windows 7, 32Gb RAM, Mechanical Hard Drive