This page is part of the GENIUS project.
Operating System and CPU Calibration Tests
This test case involved comparing various computers we used in the lab for sniffing packet traffic. A simultaneous test was performed with fixed-spaced ethernet packets of 125us for 5 minutes with various packet sniffing computer configurations.
Packet Sniffing Computer Configurations:
- 1.6Ghz P4 Compaq EVO 500 running FreeBSD 4.6
- 1.6Ghz P4 Compaq EVO 500 running Microsoft Windows 2000 (with windump)
- 600MHz Celeron CA810e Intel motherboard running FreeBSD 4.6
- 166MHz Pentium IBM 300GL running FreeBSD 4.6
Figure 2.
Setup of Hardware for Operating System and CPU Calibration Tests
Tests cases conducted were:
| Interval | Duration |
|---|---|
| 125usec (124.8usec) | 5min |
Results
The tables below detail the Mean, Variance and Standard Deviation values for the test cases generated. All values are in microseconds.
| September 3rd trial: 1.6Ghz P4 Compaq EVO 500 running FreeBSD 4.6 | |||
|---|---|---|---|
| Interval | Mean | Variance | Std Dev |
| 125usec | 208.0 | 4.9 | 2.2 |
| September 3rd trial: 1.6Ghz P4 Compaq EVO 500 running Microsoft Windows 2000 | |||
|---|---|---|---|
| Interval | Mean | Variance | Std Dev |
| 125usec | 208.0 | 9.0 | 3.0 |
| September 3rd trial: 600MHz Celeron CA810e Intel motherboard running FreeBSD 4.6 | |||
|---|---|---|---|
| Interval | Mean | Variance | Std Dev |
| 125usec | 208.0 | 7.0 | 2.6 |
| September 3rd trial: 166MHz Pentium IBM 300GL running FreeBSD 4.6 | |||
|---|---|---|---|
| Interval | Mean | Variance | Std Dev |
| 125usec | 208.0 | 27.4 | 5.2 |
 |
 |
| 1.6Ghz P4 Compaq EVO 500 FreeBSD 4.6 Figure 3. |
1.6Ghz P4 Compaq EVO 500 Windows 2000 Figure 4. |
 |
 |
| 600MHz Celeron CA810e Intel FreeBSD 4.6 Figure 5. |
166MHz Pentium IBM 300GL FreeBSD 4.6 Figure 6. |
Note: Although the above tests were run for 5 minutes duration, only a snap shot of this data is shown above due to limitations with Microsoft Excel. The entire data that was collected has been checked and is consistent with the above representative samples.
It was found that all machines appeared to time stamp packets to a reasonable amount of accuracy. Also interesting to note is the variation in time stamped packets that occurs as a result of the speed of the CPU.
| ||||||||||||||||
|
