| ??? 03/15/05 07:55 Read: times |
#89699 - application dependent Responding to: ???'s previous message |
The "poorness" or "goodness" of a pRNG is always dependent on how it is going to be used.
One can be quite satisfied with the performance of simplistic, "poor" pRNG in one application. Other can completely spoil even the most sophisticated pRNG by improper use. The state-of-the-art pRNG is the Mersenne Twister and alike; however, even that has its strengths and weaknesses. Nevertheless, they require slightly more resources than available in many typical '51 applications, while simpler pRNGs (either LFSR or LC) may be fully adequate, especially if "salted" by a real-world input (e.g. timing of human response, or network traffic). Jan Waclawek |
| Topic | Author | Date |
| Randomization? | 01/01/70 00:00 | |
| No RND op code in 8051/52 | 01/01/70 00:00 | |
| Random Number | 01/01/70 00:00 | |
| random stuff | 01/01/70 00:00 | |
| We can also count.... | 01/01/70 00:00 | |
| But Good Enough!!! | 01/01/70 00:00 | |
| LC RNG | 01/01/70 00:00 | |
| Random what? | 01/01/70 00:00 | |
| maxim's lfsr | 01/01/70 00:00 | |
| the dangers of thinking random is random | 01/01/70 00:00 | |
| wibbly maths | 01/01/70 00:00 | |
| Good old Galois Fields | 01/01/70 00:00 | |
| Bet you didnt know | 01/01/70 00:00 | |
| Try this link | 01/01/70 00:00 | |
| While browsing | 01/01/70 00:00 | |
| rand() | 01/01/70 00:00 | |
| still pseudo | 01/01/70 00:00 | |
| better pseudo! | 01/01/70 00:00 | |
| rand() is part of stdlib.h | 01/01/70 00:00 | |
| true, but not the point | 01/01/70 00:00 | |
application dependent | 01/01/70 00:00 |