ctffind4 and amplitude contrast [0.05]

Submitted by mstrauss on Fri, 2014-11-28 15:28

Forums

ctffind & ctftilt

Hi,

I've been trying to get ctffind4 to work for my data. My initial tests seemed fine, especially using the default values, so it runs. Indeed it almost always runs, except when I use an amplitude contrast value of 0.05. Then I get the following error:

forrtl: error (65): floating invalid
Aborted

It works if I use 0.052 (so I have a workaround), but for some reason it doesn't like 0.05.

mike

Alexis Tue, 2014-12-02 15:34

Hey Mike,

Thanks for your bug report. I was able to reproduce it and found where it was happening. It's now fixed here.

Since you have a workaround and the bug only affects the generation of the diagnostic image (and even then it really shouldn't have a significant effect), I won't rush out a release to fix this. I have a couple of other things I want to fix before I make a new version.

If you do want a fix now though, I can upload a static binary so just let me know.

Cheers,
Alexis

Alexis Thu, 2014-12-04 13:26

This is fixed in version 4.0.8, now available for download.

Biocheming Wed, 2015-02-25 04:38

I'm afraid that the bug is still available. At first I thought it is the problem of using precompiled program. However, when I tried it with ctffind4.0.11 compiled myself, this bug came to the screen again...
Another thing is that I found intel compiler is not free now (only for 30 days free trial). Please consider the source code updated for gfortran, although the efficiency may decrease.

Alexis Wed, 2015-02-25 09:43

Thanks for reporting this. Indeed, 4.0.11 had a regression and I re-introduced the bug. I will release 4.0.12 later today to fix this.

Regarding gfortran: you should be able to build ctffind4 (versions 4.0.9 and above) with gfortran, as long as you have a version of gfortran >= 4.9.1.