Installing an older version of AFNI

Hi everyone,

I am trying to install an older version (18.2.15) of AFNI on my computer (MacOS 10.14.5) but I have been unsuccessful so far… From what I understand it seems like the only way to install an older version of AFNI is to recompile it from source with the command @get.afni.version 18.2.15.
After doing that I figured one needs to copy the right Makefile (in my case Makefile.macos_10.12_local) which does not work at all, specifically because it looks for some files that are simply not there. Specifically all the files in the /sw directory. I don’t know when or by what this directory is supposed to be created but it wasn’t in my case.
I then tried using Makefile.macosx_10.10_Intel_64 as there was no /sw involved, but here I get this output :
10 warnings generated.
/bin/rm -f to3d
/usr/bin/gcc -O2 -DDARWIN -m64 -Wall -Wcomment -Wformat -DUSE_TRACING -DHAVE_XDBE -o to3d to3d.o -L. -L/usr/X11R6/lib -L/usr/lib -L/usr/local/lib -L/usr/X11R6/lib -L/usr/local/lib -lpython2.7 -Wl,-x -Wl,-multiply_defined -Wl,warning -Wl,-bind_at_load -Wl,-no_compact_unwind -lmrix -lcoxplot -lmri -lf2c /usr/local/lib/libXm.a -lXt -lXft -lfontconfig /usr/local/lib/libpng.a -liconv -lXmu -lXp -lXpm -lXext -lX11 -lz /usr/local/opt/expat/lib/libexpat.a -lm -lc
Undefined symbols for architecture x86_64:
“_jpeg_CreateDecompress”, referenced from:
_load_jpeg in libXm.a(Jpeg.o)
“_jpeg_calc_output_dimensions”, referenced from:
_load_jpeg in libXm.a(Jpeg.o)
“_jpeg_destroy_decompress”, referenced from:
_load_jpeg in libXm.a(Jpeg.o)
“_jpeg_finish_decompress”, referenced from:
_load_jpeg in libXm.a(Jpeg.o)
“_jpeg_read_header”, referenced from:
_load_jpeg in libXm.a(Jpeg.o)
“_jpeg_read_scanlines”, referenced from:
_load_jpeg in libXm.a(Jpeg.o)
“_jpeg_start_decompress”, referenced from:
_load_jpeg in libXm.a(Jpeg.o)
“_jpeg_std_error”, referenced from:
_load_jpeg in libXm.a(Jpeg.o)
“_jpeg_stdio_src”, referenced from:
_load_jpeg in libXm.a(Jpeg.o)
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make: *** [to3d] Error 1

Any advice on how to install am older version as I was not able to find that answer on the board?
Best regards,
Augustin

Hi Augustin,

Compiling on a mac is non-trivial. But we do happen to have a copy of that set of binaries.

Try these: macos_10.12_local.AFNI_18.2.15.tgz

  • rick

I’ll also add that if you wanted to, you could build any version of AFNI using the github pull and an updated Makefile that I wrote specifically for 10.14 (probably 10.13 too) that uses homebrew for dependencies and Anaconda for Python. It’s part of the distribution and a solid starting point. Let me know if you have questions should you decide to build other iterations.

https://github.com/afni/afni/blob/master/src/Makefile.macosx_10.14

Hi Rick,
“Not trivial” to say the least indeed ^^" Thank you very much for the binaries! I’m running the script and it seems to be working like a charm.
Is there any conflict to expect if I want to switch back to the new version (the way I would do it is to rename the abin directory abin_18.2.15 and the abin_19.6 directory abin).
Thanks again,
Augustin

Thanks for the advice! For now I used Rick’s method but I’ll have a look at yours should the occasion present itself.
Augustin

Hi Augustin,

Indeed, it seems like a good idea to keep that directory separate.

Renaming them is an okay way to go, but using a link would allow you to preserve complete names for clarity, e.g.

mv abin abin_19.1.06
ln -s abin_18.2.15 abin

Then to switch back would be:

rm abin
ln -s abin_19.1.06 abin

The ‘rm’ is not so dangerous here, since it would fail if you accidentally used it on a directory (without -r).

That does seem a little easier than altering PATH.

  • rick

Dear colleagues,

I am working on data analyses of resting-state fMRI with children on a MacOs Mojave 10.14.6 In October 2018 I installed AFNI v18.3.01 and C-PAC v1.3.0 (https://fcp-indi.github.io/docs/user/index.html) as I was working at the Child Mind Institute with the C-PAC developers. The software worked well for me until I updated AFNI last July to v19.2.00, when C-PAC started crashing due to a bug that miscommunicated the pipeline with the upgraded AFNI.

I would like to try downgrading AFNI to that specific version and see if I could make the pipeline work again. Do you happen to have a compilation of that version for my Mac?

Thank you!

-victor

Hi Victor,

Was there some change on the AFNI side that is causing problems?

We do not have version AFNI_18.3.01, but we do have version AFNI_18.3.03, if that might be good enough.

It is at: macos_10.12_local.AFNI_18.3.03.tgz

  • rick

Thank you very much for the quick response and the compilation, Rick!
I am not sure of what is exactly the problem but had to do with the seed-based correlation analyses. However, C-PAC was updated as well and I had a merge of versions of AFNI and C-PAC that were not working. I am trying now to use the same software that worked for me in October, let’s see if it works.

Good evening!

  • victor

I am having the same issue with 3dTcorr1d in CPAC after updating afni in July. Error is below. I am in Ubuntu 16.04.
I need to get an older version of afni (pre June 2019) for Ubuntu 16.04. Is this possible? Otherwise I am quite stuck.

error:
3dTcorr1D main
** Command line was:
3dTcorr1D -pearson -prefix scrubbed_preprocessed_correlation.nii.gz /home/jlee38/RS/wrk/resting_preproc_113051-100/sca_roi_0/_scan_funcf/_threshold_0.25/_compcor_ncomponents_5_selector_pc10.linear1.wm1.global1.motion1.quadratic0.gm0.compcor0.csf1/_bandpass_freqs_0.008.0.08/_mask_tt_mask_pad/3dTCorr1D/scrubbed_preprocessed.nii.gz /home/jlee38/RS/wrk/resting_preproc_113051-100/roi_timeseries_for_sca_0/_scan_funcf/_threshold_0.25/_compcor_ncomponents_5_selector_pc10.linear1.wm1.global1.motion1.quadratic0.gm0.compcor0.csf1/_bandpass_freqs_0.008.0.08/_mask_tt_mask_pad/3dROIstats/roi_stats.csv
** AFNI compile date = Jun 25 2019
** [[Precompiled binary linux_openmp_64: Jun 25 2019]]
** Program Crash **

------ CRASH LOG ------------------------------**
Fatal Signal 11 (SIGSEGV) received
… recent internal history …
-------THD_find_atr [7]: EXIT} (file=thd_atr.c line=452) to THD_datablock_apply_atr {500 ms}
+++++++THD_find_atr [7]: {ENTRY (file=thd_atr.c line=408) from THD_datablock_apply_atr {500 ms}
-------THD_find_atr [7]: EXIT} (file=thd_atr.c line=448) to THD_datablock_apply_atr {500 ms}
THD_datablock_apply_atr – brick statistics {500 ms}
+++++++THD_find_atr [7]: {ENTRY (file=thd_atr.c line=408) from THD_datablock_apply_atr {500 ms}
-------THD_find_atr [7]: EXIT} (file=thd_atr.c line=452) to THD_datablock_apply_atr {500 ms}
+++++++THD_find_atr [7]: {ENTRY (file=thd_atr.c line=408) from THD_datablock_apply_atr {500 ms}
-------THD_find_atr [7]: EXIT} (file=thd_atr.c line=448) to THD_datablock_apply_atr {500 ms}
+++++++THD_find_atr [7]: {ENTRY (file=thd_atr.c line=408) from THD_datablock_apply_atr {500 ms}
-------THD_find_atr [7]: EXIT} (file=thd_atr.c line=448) to THD_datablock_apply_atr {500 ms}
+++++++THD_find_atr [7]: {ENTRY (file=thd_atr.c line=408) from THD_datablock_apply_atr {500 ms}
-------THD_find_atr [7]: EXIT} (file=thd_atr.c line=448) to THD_datablock_apply_atr {500 ms}
+++++++THD_find_atr [7]: {ENTRY (file=thd_atr.c line=408) from THD_datablock_apply_atr {500 ms}
-------THD_find_atr [7]: EXIT} (file=thd_atr.c line=448) to THD_datablock_apply_atr {500 ms}
+++++++THD_find_atr [7]: {ENTRY (file=thd_atr.c line=408) from THD_datablock_apply_atr {500 ms}
-------THD_find_atr [7]: EXIT} (file=thd_atr.c line=448) to THD_datablock_apply_atr {500 ms}
------THD_datablock_apply_atr [6]: EXIT} (file=thd_initdblk.c line=1178) to THD_open_nifti {500 ms}
-----THD_open_nifti [5]: EXIT} (file=thd_niftiread.c line=800) to THD_open_one_dataset {500 ms}
+++++THD_patch_brickim [5]: {ENTRY (file=thd_loaddblk.c line=1274) from THD_open_one_dataset {500 ms}
-----THD_patch_brickim [5]: EXIT} (file=thd_loaddblk.c line=1313) to THD_open_one_dataset {500 ms}
+++++THD_report_obliquity [5]: {ENTRY (file=thd_coords.c line=704) from THD_open_one_dataset {500 ms}
++++++AFNI_process_environ [6]: {ENTRY (file=afni_environ.c line=112) from THD_report_obliquity {500 ms}
+++++++AFNI_suck_file [7]: {ENTRY (file=afni_environ.c line=25) from AFNI_process_environ {500 ms}
-------AFNI_suck_file [7]: EXIT} (file=afni_environ.c line=30) to AFNI_process_environ {500 ms}
------AFNI_process_environ [6]: EXIT} (file=afni_environ.c line=132) to THD_report_obliquity {500 ms}
-----THD_report_obliquity [5]: EXIT} (file=thd_coords.c line=711) to THD_open_one_dataset {500 ms}
----THD_open_one_dataset [4]: EXIT} (file=thd_opendset.c line=252) to THD_open_dataset {500 ms}
++++THD_patch_brickim [4]: {ENTRY (file=thd_loaddblk.c line=1274) from THD_open_dataset {500 ms}
----THD_patch_brickim [4]: EXIT} (file=thd_loaddblk.c line=1313) to THD_open_dataset {500 ms}
—THD_open_dataset [3]: EXIT} (file=thd_mastery.c line=171) to 3dTcorr1D main {500 ms}
+++mri_read_1D [3]: {ENTRY (file=mri_read.c line=2997) from 3dTcorr1D main {500 ms}

I am responding to the separated thread, thanks.

  • rick