search

argiopetech / base

Bayesian Analysis for Stellar Evolution
http://webfac.db.erau.edu/~vonhippt/base9/
11 stars 4 forks source link

carbonicity seems to have no effect on mcmc results #13

Closed tedvh closed 11 years ago

tedvh commented 11 years ago

David S created a simulated cluster with globular cluster properties, using WDs with carbonicity=0. He then created two scripts for mpiMcmc, one with carbonicity=0 and one with carbonicity=1. I ran both of those scripts for him and they gave absolutely identical output files (all 4 of them). I'm thinking that mpiMcmc is not passing carbonicity properly at some point.

argiopetech commented 11 years ago

Can we verify that he was using WDFilterSet 1 (Montgomery)? The wood model will default it to 0.6.

tedvh commented 11 years ago

good call. I just check and both scripts have

set wdFilterSet = 1 # 0 = Wood, 1 = Montgomery set wdFilterSet = 1 # 0 = Wood, 1 = Montgomery

though perhaps that part is being incorrectly parsed?

-Ted

On 5/27/13 5:29 PM, Elliot Robinson wrote:

Can we verify that he was using WDFilterSet 1 (Montgomery)? The wood model will default it to 0.6.

— Reply to this email directly or view it on GitHub https://github.com/argiopetech/base/issues/13#issuecomment-18515367.

argiopetech commented 11 years ago

That's no good. I'll go ahead and mark this as a bug.

I'm currently running a Montgomery model test as part of finishing up #6, so I'll go ahead and change the carbonicity when I re-run the tests tomorrow. Hopefully I'll be able to confirm and fix this.

tedvh commented 11 years ago

It seems that carbonicity is having an effect on sampleWDMass, but not mpiMcmc. More details to follow in an email.

argiopetech commented 11 years ago

This is fixed by commit 4973bb3368480bf741b51bd909bad262224dfd3c.