Hello,
I ran freesurfer ReconAll process on 27 subjects using linux opensuse 11.3. freesurfer version is 5.0.0 (freesurfer-Linux-centos4-stable-pub-v5.0.0.tar.gz). Then I ran the same process on some of those subjects using same version of freesurfer (freesurfer-Darwin-leopard-ppc-stable-pub-v5.0.0.dmg) on mac machine (snow leopard 10.6.5). Surprisingly, result that saved in stats folder are not the same even surface and volume images seem to be the same. I attached resulted files from both systems for the same subject, which issue do you think caused that difference and will it affect my overall results as I am going to use gdec with same data. In addition, what do you think is more correct, mac or linux version.
Best wishes, Jamaan
Hello,
I am sorry I forgot to attach file.
On 9 December 2010 19:39, Jamaan Alghamdi jamaan.alghamdi@gmail.com wrote:
Hello,
I ran freesurfer ReconAll process on 27 subjects using linux opensuse 11.3. freesurfer version is 5.0.0 (freesurfer-Linux-centos4-stable-pub-v5.0.0.tar.gz). Then I ran the same process on some of those subjects using same version of freesurfer (freesurfer-Darwin-leopard-ppc-stable-pub-v5.0.0.dmg) on mac machine (snow leopard 10.6.5). Surprisingly, result that saved in stats folder are not the same even surface and volume images seem to be the same. I attached resulted files from both systems for the same subject, which issue do you think caused that difference and will it affect my overall results as I am going to use gdec with same data. In addition, what do you think is more correct, mac or linux version.
Best wishes, Jamaan
-- Jamaan Alghamdi 07779801950 0151 795 4627 Magnetic Resonance & Image Analysis Research Centre (MARIARC) University of Liverpool Pembroke place Liverpool L69 3GE
http://web.mac.com/jamaan/Jamaan/Welcome.html http://jalghamdi.kau.edu.sa/Default.aspx?Site_ID=0005294&lng=EN
the results differ because the math operations are conducted identically on the two platforms. each OS uses different math libs, and probably more importantly, your processors are different (x86 vs powerpc).
the qdec analysis of each platform should give the same results though, as the stats differences you are seeing due to platform difference should be small.
i suppose i would tend to trust the linux results more, if you had to pick one.
n.
On Thu, 2010-12-09 at 20:11 +0000, Jamaan Alghamdi wrote:
Hello,
I am sorry I forgot to attach file.
On 9 December 2010 19:39, Jamaan Alghamdi jamaan.alghamdi@gmail.com wrote: Hello,
I ran freesurfer ReconAll process on 27 subjects using linux opensuse 11.3. freesurfer version is 5.0.0 (freesurfer-Linux-centos4-stable-pub-v5.0.0.tar.gz). Then I ran the same process on some of those subjects using same version of freesurfer (freesurfer-Darwin-leopard-ppc-stable-pub-v5.0.0.dmg) on mac machine (snow leopard 10.6.5). Surprisingly, result that saved in stats folder are not the same even surface and volume images seem to be the same. I attached resulted files from both systems for the same subject, which issue do you think caused that difference and will it affect my overall results as I am going to use gdec with same data. In addition, what do you think is more correct, mac or linux version. Best wishes, Jamaan -- Jamaan Alghamdi 07779801950 0151 795 4627 Magnetic Resonance & Image Analysis Research Centre (MARIARC) University of Liverpool Pembroke place Liverpool L69 3GE http://web.mac.com/jamaan/Jamaan/Welcome.html http://jalghamdi.kau.edu.sa/Default.aspx?Site_ID=0005294&lng=EN-- Jamaan Alghamdi 07779801950 0151 795 4627 Magnetic Resonance & Image Analysis Research Centre (MARIARC) University of Liverpool Pembroke place Liverpool L69 3GE
http://web.mac.com/jamaan/Jamaan/Welcome.html
http://jalghamdi.kau.edu.sa/Default.aspx?Site_ID=0005294&lng=EN
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Jamaan Alghamdi -
Nick Schmansky