 |
 |
Hi - there are no major conceptual changes - just a series of minor tweaks in different programs used. On linux the two versions take pretty much the same time. On a G5 Mac here I'm getting 1765s for FSL3.2 and 2543s for FSL3.3 actual compute time on an example dataset. I think this is mostly explained by extra robustness checking in FLIRT and in the conversion of siena_diff from C to C++ for compatibility with the rest of FSL (we've nearly finished killing off all the old C code). Both versions give almost the same answer (-3.708 and -3.658). Hence I think all is well, and would just recommend using purely FSL3.3 for these analyses. Cheers. On 30 May 2007, at 23:38, Zografos Caramanos wrote: > Hello again! > > Can anyone please describe/summarize the differences between > running SIENA > v-2.3 (FSL v-3.2beta, FLIRT version 5.2, FAST Version 3.5) and the > current > SIENA v-2.4 (FSL v-3.3.11, FLIRT version 5.4.2, FAST Version 3.53) > -- or > point me to such a description/summary. > > My Scan-Rescan results with v-2.4 seem to suggest that it is more > robust > than v-2.3, but each run of SIENA v-2.4 takes about 60 minutes as > opposed to > about 20 minutes for SIENA v-2.3 (running on a Dual 2GHz PowerPC G5 > with > 2-GB DDR SDRAM and Mac OS X v-10.4.8) and I am curious as to what is > responsible for this. > > Cheers! > Aki > > Zografos Caramanos, M.A. > Magnetic Resonance Studies Unit > Montreal Neurological Institute, McGill University > (e-mail) [log in to unmask] ------------------------------------------------------------------------ --- Stephen M. Smith, Professor of Biomedical Engineering Associate Director, Oxford University FMRIB Centre FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK +44 (0) 1865 222726 (fax 222717) [log in to unmask] http://www.fmrib.ox.ac.uk/~steve ------------------------------------------------------------------------ ---