Topics

Non-positive photo cs with high resolution


Stefano Bianchi
 

Hi all,

this simple input file:

set continuum resolution 0.03
agn kirk
hden 6.6
xi 3.
stop column density 21.

ends in a stop in SanityCheck, due to 'non-positive photo cs' for various elements. This happens both for c17.00 and c17.01.
The error disappears if I use a lower continuum resolution (i.e. a larger number instead of 0.03).
I had an 'old' branch pre-release installed (r12745M): with that version, it also works with the high resolution. I haven't tried the current branch pre-release.

Any clues?

Thanks,

Stefano


Gary Ferland
 

this ran fine for me.  I used the current LLVM on a Mac.  You might check other compilers/machines or try updating the compiler you have.  What compiler are you using?

We use Opacity Project photo cs where possible and they have negative values below the photoionization threshold as a sentinel for its location.  The code knows about this and protects against it - hence the test you tripped over.  There should be code in place to ensure that we do not evaluate the OP data below the threshold.

On Mon, May 11, 2020 at 4:06 AM Stefano Bianchi <bianchi@...> wrote:
Hi all,

this simple input file:

set continuum resolution 0.03
agn kirk
hden 6.6
xi 3.
stop column density 21.

ends in a stop in SanityCheck, due to 'non-positive photo cs' for various elements. This happens both for c17.00 and c17.01.
The error disappears if I use a lower continuum resolution (i.e. a larger number instead of 0.03).
I had an 'old' branch pre-release installed (r12745M): with that version, it also works with the high resolution. I haven't tried the current branch pre-release.

Any clues?

Thanks,

Stefano



--
Gary J. Ferland
Physics, Univ of Kentucky
Lexington KY 40506 USA
Tel: 859 257-8795
https://pa.as.uky.edu/users/gary


Stefano Bianchi
 

Hi Gary,

I got the crash in two different machines, with different compilers:

GCC: (Ubuntu 7.4.0-1ubuntu1~18.04) 7.4.0
GCC: (GNU) 4.8.5 20150623 (Red Hat 4.8.5-36)

Note that the branch version r12745M does NOT fail with the same machines/compilers.

Thanks,

Stefano


Gary Ferland
 

hang on - i got it to fail

On Tue, May 12, 2020 at 9:46 AM Stefano Bianchi <bianchi@...> wrote:
Hi Gary,

I got the crash in two different machines, with different compilers:

GCC: (Ubuntu 7.4.0-1ubuntu1~18.04) 7.4.0
GCC: (GNU) 4.8.5 20150623 (Red Hat 4.8.5-36)

Note that the branch version r12745M does NOT fail with the same machines/compilers.

Thanks,

Stefano



--
Gary J. Ferland
Physics, Univ of Kentucky
Lexington KY 40506 USA
Tel: 859 257-8795
https://pa.as.uky.edu/users/gary


Gary Ferland
 

It turns out that your sim runs fine with c17.02, which I am releasing today.

we had not come across your problem during our development since we had not tried doing this exact use case.  Our versions of c17.01 do fail with all of our machines and compilers, and c17.02 passes, as does the trunk.  Please give c17.02 a try. You can get it from the download page.

thanks,
Gary

On Tue, May 12, 2020 at 2:50 PM Gary Ferland via groups.io <gary=g.uky.edu@groups.io> wrote:
hang on - i got it to fail

On Tue, May 12, 2020 at 9:46 AM Stefano Bianchi <bianchi@...> wrote:
Hi Gary,

I got the crash in two different machines, with different compilers:

GCC: (Ubuntu 7.4.0-1ubuntu1~18.04) 7.4.0
GCC: (GNU) 4.8.5 20150623 (Red Hat 4.8.5-36)

Note that the branch version r12745M does NOT fail with the same machines/compilers.

Thanks,

Stefano



--
Gary J. Ferland
Physics, Univ of Kentucky
Lexington KY 40506 USA
Tel: 859 257-8795
https://pa.as.uky.edu/users/gary



--
Gary J. Ferland
Physics, Univ of Kentucky
Lexington KY 40506 USA
Tel: 859 257-8795
https://pa.as.uky.edu/users/gary


Stefano Bianchi
 

Great!

Thanks,
Stefano