3dPval approximate computation time?


I want to convert a statistical array (F statistic) to p-values. I’ve chosen 3dPval to do this. The input file contains a single sub-brik of resolution:64 x 64 x 36 where each voxel is 4mm isotropic.This process is running on an Intel Xeon CPU E5-2640 2.4 Ghz.

So far the computation has been running for over 24 hours. I expected it to be much faster as the GUI can seemingly do this almost instantaneously.

Does anyone have any experience with this function that could tell me if this length of computation time is typical? Or suggest a faster alternative?

Kind regards,

Hmmm, it should be fast. On my Macbook Air, that takes about 0.1 seconds total, for a slightly larger dataset. What is the output of 3dinfo on this dataset? Does it show what type of data is there? Post that output and the original command you used.

Thanks for taking a look Daniel.

++ 3dinfo: AFNI version=AFNI_19.1.08 (May 10 2019) [64-bit]

Dataset File: sub-MSC02_glasslexical_SPMG1_GLM_brik21.nii
Identifier Code: XYZ_sAQSPDu6qhaFuGm2WCEhQw Creation Date: Mon Feb 7 13:26:45 2022
Template Space: ORIG
Dataset Type: Func-Bucket (-fbuc)
Byte Order: LSB_FIRST {assumed} [this CPU native = LSB_FIRST]
Storage Mode: NIFTI
Storage Space: 589,824 (590 thousand) bytes
Geometry String: “MATRIX(3.999912,-0.021446,-0.015559,-128.6708,-0.022576,-3.988152,-0.306819,127.6224,0.013868,-0.3069,3.988186,-31.27849):64,64,36”
Data Axes Tilt: Oblique (4.411 deg. from plumb)
Data Axes Approximate Orientation:
first (x) = Right-to-Left
second (y) = Posterior-to-Anterior
third (z) = Inferior-to-Superior [-orient RPI]
R-to-L extent: -128.671 [R] -to- 123.329 [L] -step- 4.000 mm [ 64 voxels]
A-to-P extent: -124.378 [A] -to- 127.622 [P] -step- 4.000 mm [ 64 voxels]
I-to-S extent: -31.278 [I] -to- 108.722 [S] -step- 4.000 mm [ 36 voxels]
Number of values stored at each pixel = 1
– At sub-brick #0 ‘Glass_coherentGlass_random-Noun-Verb_GLT_Fstat’ datum type is float: 0 to 65.1411
statcode = fift; statpar = 1 372

3dPval -input sub-MSC02_glasslexical_SPMG1_GLM_brik21.nii -output sub-MSC02_glasslexical_SPMG1_GLM_brik21_PVAL

OK, so that version is more than 2 years old. Can you try with a newer version? If you convert the dataset to AFNI format with 3dcopy, does it work any faster?

Thanks Daniel, I solved it.

3dPval doesn’t take -input -output arguements. It only accepts -prefix for the output. It must have just been running and hanging with no error messages for silly users!

Yes, the original version did not trap bad options, later it would infinitely warn.
As of the next build, they will be terminal.

It is also a nice habit to post any troublesome commands, otherwise it is hard to guess what the problem may be.


  • rick