search

cloudy-astrophysics / bug-tracker-migration-test

Trial run for importing the nublado.org Trac tickets as GitHub issues
0 stars 0 forks source link

single-zone sim causes unit integration error (trac #396) #398

Closed cloudy-bot closed 5 years ago

cloudy-bot commented 7 years ago

reported by: @CloudyLex

The attached was submitted to me privately. It does stop with this error on both the trunk and c17.00. It is very fast.

 PROBLEM PrtComment insanity - Continuum unit integration did not verify 
 exact vol= 0.909091, derived vol= 0.907029 relative difference is 0.0022676 
 ConInterOut= 9.07029e-11,  
 Something that cannot happen, has happened.
 This is TotalInsanity, I live in ../service.cpp.

Migrated from https://www.nublado.org/ticket/396

{
    "status": "closed",
    "changetime": "2019-02-04T16:18:34Z",
    "_ts": "1549297114434209",
    "description": "The attached was submitted to me privately.  It does stop with this error on both the trunk and c17.00.  It is very fast.\n\n{{{\n PROBLEM PrtComment insanity - Continuum unit integration did not verify \n exact vol= 0.909091, derived vol= 0.907029 relative difference is 0.0022676 \n ConInterOut= 9.07029e-11,  \n Something that cannot happen, has happened.\n This is TotalInsanity, I live in ../service.cpp.\n\n\n\n}}}",
    "reporter": "gary",
    "cc": "",
    "resolution": "fixed",
    "time": "2017-08-11T22:23:13Z",
    "component": "infrastructure",
    "summary": "single-zone sim causes unit integration error",
    "priority": "critical",
    "keywords": "",
    "version": "c17.00",
    "milestone": "c17.02",
    "owner": "peter",
    "type": "defect - code aborts"
}
cloudy-bot commented 7 years ago

Attachment: disaster.in

cloudy-bot commented 7 years ago

dropbox link to mod file https://www.dropbox.com/s/ycwuczxd0we29sw/inst_m1e6_Z004.mod?dl=0

The attached was submitted to me privately. It does stop with this error on both the trunk and c17.00. It is very fast.

 PROBLEM PrtComment insanity - Continuum unit integration did not verify 
 exact vol= 0.909091, derived vol= 0.907029 relative difference is 0.0022676 
 ConInterOut= 9.07029e-11,  
 Something that cannot happen, has happened.
 This is TotalInsanity, I live in ../service.cpp.

to:

The attached was submitted to me privately. It does stop with this error on both the trunk and c17.00. It is very fast.

 PROBLEM PrtComment insanity - Continuum unit integration did not verify 
 exact vol= 0.909091, derived vol= 0.907029 relative difference is 0.0022676 
 ConInterOut= 9.07029e-11,  
 Something that cannot happen, has happened.
 This is TotalInsanity, I live in ../service.cpp.
cloudy-bot commented 7 years ago

@CloudyLex commented:

dropbox link to mod file https://www.dropbox.com/s/ycwuczxd0we29sw/inst_m1e6_Z004.mod?dl=0

cloudy-bot commented 7 years ago

Original starburst file needed to run the sim Attachment: inst_m1e6_Z004.stb99.gz

cloudy-bot commented 7 years ago

Milestone renamed

cloudy-bot commented 7 years ago

Milestone renamed

cloudy-bot commented 7 years ago

@CloudyLex changed milestone from "C17.01" to "C17.02"

cloudy-bot commented 7 years ago

Milestone renamed

cloudy-bot commented 7 years ago

@peter-van-hoof-noaccount changed version from "trunk" to "c17.00"

cloudy-bot commented 7 years ago

second version of failing script Attachment: disaster2.in

cloudy-bot commented 7 years ago

The Starburst99 SED is an innocent bystander. Replacing it with a 50 kK BB yields the same problem. This sim is attached as disaster2.in.

cloudy-bot commented 7 years ago

This bug was caused by a somewhat inaccurate definition of Dilution in rt_continuum.cpp. The correct value obtained by integrating (r_in/r)^2^ over the thickness of the zone (drad) is: r_in^2^/(r(r-drad)). We used (r_in/(r-drad/2))^2^ = r_in^2^/(r(r-drad) + drad^2^/4). The term drad^2^/4 was big enough to make the test trip in this case.

Fixed on the trunk in r12576 and on c17_branch in r12577.