-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi Haydyn-- > I have a quick question to ask you about the random number > generation SEED in XPLOR-NIH. The scripts supplied in the eginputs use > either $seed=778, or $seed=65748309 for the initial velocities. Are > these differences related to the architecture of the computer used for > calculations? I'm running Mandrake 9.0 linux, and have had to adjust > things like the maximum for random number generation in other types of > programs (I typically use values for this of 2144563822.0), and use > seed values of say 82364 (which is actually what I use for CNS > calculations). The exact value of the seed shouldn't matter too much- but it shouldn't be too large- because the XPLOR interface stores the seed as a floating point number, strange things happen it the seed gets larger than about 10^8. > Will altering the SEED change things much? Or will it > just produce a slightly different set of results. The value of the seed controls the exact sequence of pseudo-random numbers used for things like setting initial random velocities. Thus, changing the seed will change the details of a particular structure. However, the set of final structures should be relatively insensitive to the seed value. If the set of results does depend sensitively on seed, it usually means that the optimization protocol needs some further work. best regards-- Charles -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/> iD8DBQE+ykatPK2zrJwS/lYRAnnvAJ4oF4+YeL7ukgY8eeW3Y4+ouR2jYQCfUV0g 5nYW0mk9PC/zl7rK99KJUnU= =jsfp -----END PGP SIGNATURE-----
