Hi Rich,
I'm getting the following at the start of the Adiabatic Mode Finding for the examples using models b3/degroote and mesa/bcep:
Adiabatic Mode Finding
======================
Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.
Backtrace for this error:
#0 0x100f5b092
#1 0x100f5b85e
#2 0x7fff93063949
#3 0x100607603
#4 0x1006081d6
#5 0x10066ed8a
#6 0x1005d652c
#7 0x1005d6a9d
I'll also point out the huge jump in the mode identification number in the red giant model:
Mode : -353 7 360 0.1121962284733227E+02 0.0000000000000000E+00 0.7264204676687335E-11 6
Mode : -352 7 359 0.1125019601084189E+02 0.0000000000000000E+00 0.4763779933772715E-11 6
Mode : -351 7 358 0.1128019668084656E+02 0.0000000000000000E+00 0.1243273109485961E-11 10
*Mode : -274 10 284 0.1130705063783703E+02 0.0000000000000000E+00 0.1203083274905371E-09 14
*Mode : -193 8 201 0.1132699805051222E+02 0.0000000000000000E+00 0.1208018410339823E-08 8
Mode : -348 7 355 0.1135151363881633E+02 0.0000000000000000E+00 0.2956567722155446E-10 7
Mode : -347 7 354 0.1138139220430958E+02 0.0000000000000000E+00 0.1426072489435195E-09 10
Mode : -346 7 353 0.1141271544665557E+02 0.0000000000000000E+00 0.1148942633976547E-10 6
Mode : -345 7 352 0.1144457673327924E+02 0.0000000000000000E+00 0.1393297851301246E-10 5
I know the difficulty with labelling rules in evolved models and I admittedly didn't check out the eigenfunctions for the modes before, during, and after the jump in the labelling, but, I thought I'd point it out just in case.
Chris
floating point exception in example models
-
rhtownsend
- Site Admin
- Posts: 397
- Joined: Sun Mar 31, 2013 4:22 pm
Re: floating point exception in example models
Thanks for the report; the fix is easy. I'll do a version bump today.
cheers,
Rich
cheers,
Rich
Re: floating point exception in example models
Hi Rich,
I get a similar error (see below) when trying to compute the frequencies in the attached model / input file "gyre.in.tar.gz". Honestly I would not exactly mind having the error, but it also freezes up my terminal, so I have to manually interrupt the process before proceeding.
I'm trying to compute frequencies for a whole range of models automatically, checking consistency only after computation. I could handle an exit code different from zero (I could then afterwards check what went wrong with these models), but if it doesn't exit at all, I can't try the next model.
So you might find that the model/parameters are rubbish, but still it would be nice if the terminal exits. As a matter of fact, it's a pre-main sequence model, so perhaps the conditions to run GYRE are not satisfied. I did not check e.g. conditions for hydrostatic equilibrium.
If I try the model right before it (profile1000 in attachment), GYRE also fails, but at least it exits nicely, so I'm not worried about this one. The error I get from that one is also appended below.
Setup: Kubuntu 13.04.
Output:
gyre_ad [1.0.1]
===============
Compler : GCC version 4.9.0 20130404 (experimental)
Compler options : -I /usr/include/x86_64-linux-gnu -I /home/pieterd/software/gyre/madsdk/include -mtune=generic -march=x86-64 -g -Wall -Wno-unused-dummy-argument -std=f2008 -finit-real=snan -ffpe-trap=invalid,overflow,zero -fbacktrace -fcheck=all -finline-limit=0 -fopenmp
OpenMP Threads : 4
MPI Processors : 1
Initialization
==============
Adiabatic Discriminant Scan
===========================
Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.
Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.
Backtrace for this error:
Backtrace for this error:
#0 0x7F2FA70EC4C7
#1 0x7F2FA70ECACE
#2 0x7F2FA61BF0AF
#0 0x7F2FA70EC4C7
#1 0x7F2FA70ECACE
#2 0x7F2FA61BF0AF
#3 0x63BC10 in __gyre_ext_arith_MOD_exp_r at gyre_ext_arith.f90:530
#4 0x63BADD in __gyre_ext_arith_MOD_exp_c at gyre_ext_arith.f90:553
#5 0x4D1112 in solve_magnus at gyre_ivp_magnus.f90:271
#6 0x4D7B6D in __gyre_ivp_magnus_MOD_solve_magnus_gl2 at gyre_ivp_magnus.f90:126
#7 0x4648E7 in __gyre_ivp_MOD_solve at gyre_ivp.f90:95
#8 0x461F59 in __gyre_ad_shooter_MOD_shoot._omp_fn.1 at gyre_ad_shooter.f90:138
#9 0x7F2FA6BC7479
#10 0x7F2FA6557F8D
#11 0x7F2FA6281E1C
#12 0xFFFFFFFFFFFFFFFF
#3 0x63BC10 in __gyre_ext_arith_MOD_exp_r at gyre_ext_arith.f90:530
#4 0x63BADD in __gyre_ext_arith_MOD_exp_c at gyre_ext_arith.f90:553
#5 0x4D1112 in solve_magnus at gyre_ivp_magnus.f90:271
#6 0x4D7B6D in __gyre_ivp_magnus_MOD_solve_magnus_gl2 at gyre_ivp_magnus.f90:126
#7 0x4648E7 in __gyre_ivp_MOD_solve at gyre_ivp.f90:95
#8 0x461F59 in __gyre_ad_shooter_MOD_shoot._omp_fn.1 at gyre_ad_shooter.f90:138
#9 0x7F2FA6BC7479
#10 0x7F2FA6557F8D
#11 0x7F2FA6281E1C
#12 0xFFFFFFFFFFFFFFFF
^C
Output for profile1000
gyre_ad [1.0.1]
===============
Compler : GCC version 4.9.0 20130404 (experimental)
Compler options : -I /usr/include/x86_64-linux-gnu -I /home/pieterd/software/gyre/madsdk/include -mtune=generic -march=x86-64 -g -Wall -Wno-unused-dummy-argument -std=f2008 -finit-real=snan -ffpe-trap=invalid,overflow,zero -fbacktrace -fcheck=all -finline-limit=0 -fopenmp
OpenMP Threads : 4
MPI Processors : 1
Initialization
==============
Adiabatic Discriminant Scan
===========================
Eval: 0.3131878342057633E-02 0.8036328907305127 *******
Eval: 0.4346280148161612E-02 0.6812890713969033 *******
Eval: 0.5560681954265593E-02 0.6329828542347788 *******
Eval: 0.6775083760369572E-02 -0.9401582529654947 *******
Eval: 0.7989485566473553E-02 -0.5091710930353177 *******
Eval: 0.9203887372577532E-02 0.5280913616712931 *******
Eval: 0.1041828917868151E-01 0.8744848766761222 *******
Eval: 0.1163269098478549E-01 -0.6928473628782639 *******
Eval: 0.1284709279088947E-01 0.6447920563594917 *******
Eval: 0.1406149459699345E-01 -0.5046082960659918 *******
Eval: 0.1527589640309743E-01 0.9429821060372726 *******
Eval: 0.1649029820920141E-01 -0.7971433313448842 *******
Eval: 0.1770470001530539E-01 -0.8465616844903481 *******
Eval: 0.1891910182140937E-01 -0.5810388542603568 *******
Eval: 0.2013350362751335E-01 -0.7787356150611640 *******
Eval: 0.2134790543361733E-01 -0.6860596863476789 *******
Eval: 0.2256230723972131E-01 0.6962068820680440 *******
Eval: 0.2377670904582529E-01 -0.6786169935879860 *******
Eval: 0.2499111085192927E-01 0.8660259309484725 *******
Eval: 0.2620551265803326E-01 -0.5859003751012539 *******
Eval: 0.2741991446413723E-01 -0.5704082817984736 *******
Eval: 0.2863431627024121E-01 -0.9345582016293542 *******
Eval: 0.2984871807634519E-01 -0.5664484635280953 *******
Eval: 0.3106311988244917E-01 0.9937988383110834 *******
Eval: 0.3227752168855316E-01 0.7937762898333336 *******
Eval: 0.3349192349465713E-01 0.7994339058056762 *******
Eval: 0.3470632530076111E-01 -0.7577926431829237 *******
Eval: 0.3592072710686509E-01 -0.8259363887649024 *******
Eval: 0.3713512891296907E-01 0.9946045128363186 *******
Eval: 0.3834953071907305E-01 0.9021729193668240 *******
Eval: 0.3956393252517704E-01 0.6431651181173414 *******
Eval: 0.4077833433128102E-01 0.6964665390665398 *******
Eval: 0.4199273613738499E-01 -0.6507743237089173 *******
Eval: 0.4320713794348897E-01 -0.9708037952238013 *******
Eval: 0.4442153974959296E-01 0.9470787332230565 *******
Eval: 0.4563594155569693E-01 0.7028941030257136 *******
Eval: 0.4685034336180091E-01 0.8954390945458315 *******
Eval: 0.4806474516790489E-01 -0.8464855836306908 *******
Eval: 0.4927914697400888E-01 0.7637499710500957 *******
Eval: 0.5049354878011286E-01 -0.7204893432037157 *******
Eval: 0.5170795058621683E-01 0.5739490260789323 *******
Eval: 0.5292235239232081E-01 -0.8593934540315005 *******
Eval: 0.5413675419842479E-01 -0.5098263720620654 *******
Eval: 0.5535115600452877E-01 -0.9804259437127127 *******
Eval: 0.5656555781063275E-01 0.6988333111896263 *******
Eval: 0.5777995961673674E-01 0.7546709437642066 *******
Eval: 0.5899436142284071E-01 0.7665838125787900 *******
Eval: 0.6020876322894469E-01 -0.8004370150826557 *******
Eval: 0.6142316503504867E-01 0.8052569903805094 *******
Eval: 0.6263756684115265E-01 -0.7610164797046051 *******
Completed ad scan; time elapsed: 0.397 s
Adiabatic Mode Finding
======================
INFO at line 430 <gyre_sysmtx:null_vector_banded>:
Smallest element not in final block
ASSERT 'i > 0 .AND. i < this%n' failed at line 255 <core_spline:interp_v>:
Out-of-bounds interpolation
I get a similar error (see below) when trying to compute the frequencies in the attached model / input file "gyre.in.tar.gz". Honestly I would not exactly mind having the error, but it also freezes up my terminal, so I have to manually interrupt the process before proceeding.
I'm trying to compute frequencies for a whole range of models automatically, checking consistency only after computation. I could handle an exit code different from zero (I could then afterwards check what went wrong with these models), but if it doesn't exit at all, I can't try the next model.
So you might find that the model/parameters are rubbish, but still it would be nice if the terminal exits. As a matter of fact, it's a pre-main sequence model, so perhaps the conditions to run GYRE are not satisfied. I did not check e.g. conditions for hydrostatic equilibrium.
If I try the model right before it (profile1000 in attachment), GYRE also fails, but at least it exits nicely, so I'm not worried about this one. The error I get from that one is also appended below.
Setup: Kubuntu 13.04.
Output:
gyre_ad [1.0.1]
===============
Compler : GCC version 4.9.0 20130404 (experimental)
Compler options : -I /usr/include/x86_64-linux-gnu -I /home/pieterd/software/gyre/madsdk/include -mtune=generic -march=x86-64 -g -Wall -Wno-unused-dummy-argument -std=f2008 -finit-real=snan -ffpe-trap=invalid,overflow,zero -fbacktrace -fcheck=all -finline-limit=0 -fopenmp
OpenMP Threads : 4
MPI Processors : 1
Initialization
==============
Adiabatic Discriminant Scan
===========================
Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.
Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.
Backtrace for this error:
Backtrace for this error:
#0 0x7F2FA70EC4C7
#1 0x7F2FA70ECACE
#2 0x7F2FA61BF0AF
#0 0x7F2FA70EC4C7
#1 0x7F2FA70ECACE
#2 0x7F2FA61BF0AF
#3 0x63BC10 in __gyre_ext_arith_MOD_exp_r at gyre_ext_arith.f90:530
#4 0x63BADD in __gyre_ext_arith_MOD_exp_c at gyre_ext_arith.f90:553
#5 0x4D1112 in solve_magnus at gyre_ivp_magnus.f90:271
#6 0x4D7B6D in __gyre_ivp_magnus_MOD_solve_magnus_gl2 at gyre_ivp_magnus.f90:126
#7 0x4648E7 in __gyre_ivp_MOD_solve at gyre_ivp.f90:95
#8 0x461F59 in __gyre_ad_shooter_MOD_shoot._omp_fn.1 at gyre_ad_shooter.f90:138
#9 0x7F2FA6BC7479
#10 0x7F2FA6557F8D
#11 0x7F2FA6281E1C
#12 0xFFFFFFFFFFFFFFFF
#3 0x63BC10 in __gyre_ext_arith_MOD_exp_r at gyre_ext_arith.f90:530
#4 0x63BADD in __gyre_ext_arith_MOD_exp_c at gyre_ext_arith.f90:553
#5 0x4D1112 in solve_magnus at gyre_ivp_magnus.f90:271
#6 0x4D7B6D in __gyre_ivp_magnus_MOD_solve_magnus_gl2 at gyre_ivp_magnus.f90:126
#7 0x4648E7 in __gyre_ivp_MOD_solve at gyre_ivp.f90:95
#8 0x461F59 in __gyre_ad_shooter_MOD_shoot._omp_fn.1 at gyre_ad_shooter.f90:138
#9 0x7F2FA6BC7479
#10 0x7F2FA6557F8D
#11 0x7F2FA6281E1C
#12 0xFFFFFFFFFFFFFFFF
^C
Output for profile1000
gyre_ad [1.0.1]
===============
Compler : GCC version 4.9.0 20130404 (experimental)
Compler options : -I /usr/include/x86_64-linux-gnu -I /home/pieterd/software/gyre/madsdk/include -mtune=generic -march=x86-64 -g -Wall -Wno-unused-dummy-argument -std=f2008 -finit-real=snan -ffpe-trap=invalid,overflow,zero -fbacktrace -fcheck=all -finline-limit=0 -fopenmp
OpenMP Threads : 4
MPI Processors : 1
Initialization
==============
Adiabatic Discriminant Scan
===========================
Eval: 0.3131878342057633E-02 0.8036328907305127 *******
Eval: 0.4346280148161612E-02 0.6812890713969033 *******
Eval: 0.5560681954265593E-02 0.6329828542347788 *******
Eval: 0.6775083760369572E-02 -0.9401582529654947 *******
Eval: 0.7989485566473553E-02 -0.5091710930353177 *******
Eval: 0.9203887372577532E-02 0.5280913616712931 *******
Eval: 0.1041828917868151E-01 0.8744848766761222 *******
Eval: 0.1163269098478549E-01 -0.6928473628782639 *******
Eval: 0.1284709279088947E-01 0.6447920563594917 *******
Eval: 0.1406149459699345E-01 -0.5046082960659918 *******
Eval: 0.1527589640309743E-01 0.9429821060372726 *******
Eval: 0.1649029820920141E-01 -0.7971433313448842 *******
Eval: 0.1770470001530539E-01 -0.8465616844903481 *******
Eval: 0.1891910182140937E-01 -0.5810388542603568 *******
Eval: 0.2013350362751335E-01 -0.7787356150611640 *******
Eval: 0.2134790543361733E-01 -0.6860596863476789 *******
Eval: 0.2256230723972131E-01 0.6962068820680440 *******
Eval: 0.2377670904582529E-01 -0.6786169935879860 *******
Eval: 0.2499111085192927E-01 0.8660259309484725 *******
Eval: 0.2620551265803326E-01 -0.5859003751012539 *******
Eval: 0.2741991446413723E-01 -0.5704082817984736 *******
Eval: 0.2863431627024121E-01 -0.9345582016293542 *******
Eval: 0.2984871807634519E-01 -0.5664484635280953 *******
Eval: 0.3106311988244917E-01 0.9937988383110834 *******
Eval: 0.3227752168855316E-01 0.7937762898333336 *******
Eval: 0.3349192349465713E-01 0.7994339058056762 *******
Eval: 0.3470632530076111E-01 -0.7577926431829237 *******
Eval: 0.3592072710686509E-01 -0.8259363887649024 *******
Eval: 0.3713512891296907E-01 0.9946045128363186 *******
Eval: 0.3834953071907305E-01 0.9021729193668240 *******
Eval: 0.3956393252517704E-01 0.6431651181173414 *******
Eval: 0.4077833433128102E-01 0.6964665390665398 *******
Eval: 0.4199273613738499E-01 -0.6507743237089173 *******
Eval: 0.4320713794348897E-01 -0.9708037952238013 *******
Eval: 0.4442153974959296E-01 0.9470787332230565 *******
Eval: 0.4563594155569693E-01 0.7028941030257136 *******
Eval: 0.4685034336180091E-01 0.8954390945458315 *******
Eval: 0.4806474516790489E-01 -0.8464855836306908 *******
Eval: 0.4927914697400888E-01 0.7637499710500957 *******
Eval: 0.5049354878011286E-01 -0.7204893432037157 *******
Eval: 0.5170795058621683E-01 0.5739490260789323 *******
Eval: 0.5292235239232081E-01 -0.8593934540315005 *******
Eval: 0.5413675419842479E-01 -0.5098263720620654 *******
Eval: 0.5535115600452877E-01 -0.9804259437127127 *******
Eval: 0.5656555781063275E-01 0.6988333111896263 *******
Eval: 0.5777995961673674E-01 0.7546709437642066 *******
Eval: 0.5899436142284071E-01 0.7665838125787900 *******
Eval: 0.6020876322894469E-01 -0.8004370150826557 *******
Eval: 0.6142316503504867E-01 0.8052569903805094 *******
Eval: 0.6263756684115265E-01 -0.7610164797046051 *******
Completed ad scan; time elapsed: 0.397 s
Adiabatic Mode Finding
======================
INFO at line 430 <gyre_sysmtx:null_vector_banded>:
Smallest element not in final block
ASSERT 'i > 0 .AND. i < this%n' failed at line 255 <core_spline:interp_v>:
Out-of-bounds interpolation
- Attachments
-
- gyre.in.tar.gz
- (226.56 KiB) Downloaded 339 times
-
rhtownsend
- Site Admin
- Posts: 397
- Joined: Sun Mar 31, 2013 4:22 pm
Re: floating point exception in example models
Oops, I didn't see this part of the message!
> I'll also point out the huge jump in the mode identification number in the
> red giant model:
>
> Mode : -353 7 360 0.1121962284733227E+02
> 0.0000000000000000E+00 0.7264204676687335E-11 6
> Mode : -352 7 359 0.1125019601084189E+02
> 0.0000000000000000E+00 0.4763779933772715E-11 6
> Mode : -351 7 358 0.1128019668084656E+02
> 0.0000000000000000E+00 0.1243273109485961E-11 10
> *Mode : -274 10 284 0.1130705063783703E+02
> 0.0000000000000000E+00 0.1203083274905371E-09 14
> *Mode : -193 8 201 0.1132699805051222E+02
> 0.0000000000000000E+00 0.1208018410339823E-08 8
> Mode : -348 7 355 0.1135151363881633E+02
> 0.0000000000000000E+00 0.2956567722155446E-10 7
> Mode : -347 7 354 0.1138139220430958E+02
> 0.0000000000000000E+00 0.1426072489435195E-09 10
> Mode : -346 7 353 0.1141271544665557E+02
> 0.0000000000000000E+00 0.1148942633976547E-10 6
> Mode : -345 7 352 0.1144457673327924E+02
> 0.0000000000000000E+00 0.1393297851301246E-10 5
>
>
> I know the difficulty with labelling rules in evolved models and I
> admittedly didn't check out the eigenfunctions for the modes before,
> during, and after the jump in the labelling, but, I thought I'd point it
> out just in case.
I've known about this for sometime; it looks pretty nasty, but comes about because for some modes the gravitational potential perturbation "lifts" the whole displacement eigenfunction above/below the zero line --- causing many nodes to disappear. The fix is to use Masao Takata's classification scheme for dipole modes; I hope to get this coded up sometime soon (or someone else could, maybe!).
> I'll also point out the huge jump in the mode identification number in the
> red giant model:
>
> Mode : -353 7 360 0.1121962284733227E+02
> 0.0000000000000000E+00 0.7264204676687335E-11 6
> Mode : -352 7 359 0.1125019601084189E+02
> 0.0000000000000000E+00 0.4763779933772715E-11 6
> Mode : -351 7 358 0.1128019668084656E+02
> 0.0000000000000000E+00 0.1243273109485961E-11 10
> *Mode : -274 10 284 0.1130705063783703E+02
> 0.0000000000000000E+00 0.1203083274905371E-09 14
> *Mode : -193 8 201 0.1132699805051222E+02
> 0.0000000000000000E+00 0.1208018410339823E-08 8
> Mode : -348 7 355 0.1135151363881633E+02
> 0.0000000000000000E+00 0.2956567722155446E-10 7
> Mode : -347 7 354 0.1138139220430958E+02
> 0.0000000000000000E+00 0.1426072489435195E-09 10
> Mode : -346 7 353 0.1141271544665557E+02
> 0.0000000000000000E+00 0.1148942633976547E-10 6
> Mode : -345 7 352 0.1144457673327924E+02
> 0.0000000000000000E+00 0.1393297851301246E-10 5
>
>
> I know the difficulty with labelling rules in evolved models and I
> admittedly didn't check out the eigenfunctions for the modes before,
> during, and after the jump in the labelling, but, I thought I'd point it
> out just in case.
I've known about this for sometime; it looks pretty nasty, but comes about because for some modes the gravitational potential perturbation "lifts" the whole displacement eigenfunction above/below the zero line --- causing many nodes to disappear. The fix is to use Masao Takata's classification scheme for dipole modes; I hope to get this coded up sometime soon (or someone else could, maybe!).
-
rhtownsend
- Site Admin
- Posts: 397
- Joined: Sun Mar 31, 2013 4:22 pm
Re: floating point exception in example models
pieterd wrote:
> Hi Rich,
>
> I get a similar error (see below) when trying to compute the frequencies in the
> attached model / input file "gyre.in.tar.gz". Honestly I would not exactly
> mind having the error, but it also freezes up my terminal, so I have to manually
> interrupt the process before proceeding.
>
Unfortunately, there's probably not much I can do about the terminal freezing; this is tied up in the gfortran runtime, and hence out of my control.
However, I can of course attempt to fix the bug which is causing the error in the first place!
> I'm trying to compute frequencies for a whole range of models automatically, checking
> consistency only after computation. I could handle an exit code different from zero
> (I could then afterwards check what went wrong with these models), but if it doesn't
> exit at all, I can't try the next model.
>
> So you might find that the model/parameters are rubbish, but still it would be nice
> if the terminal exits. As a matter of fact, it's a pre-main sequence model, so
> perhaps the conditions to run GYRE are not satisfied. I did not check e.g. conditions
> for hydrostatic equilibrium.
>
> If I try the model right before it (profile1000 in attachment), GYRE also fails, but
> at least it exits nicely, so I'm not worried about this one. The error I get from
> that one is also appended below.
>
It seems that these models are bogus; the w parameter (third column of the input files) has a huge value at the first datapoint, and hence is (grossly) non-monotonic. This is something you should take up with Bill; but in the meantime, I'm going to put out a minor release of GYRE which checks for non-monotonicity in MESA files.
Thanks for the report!
cheers,
Rich
> Hi Rich,
>
> I get a similar error (see below) when trying to compute the frequencies in the
> attached model / input file "gyre.in.tar.gz". Honestly I would not exactly
> mind having the error, but it also freezes up my terminal, so I have to manually
> interrupt the process before proceeding.
>
Unfortunately, there's probably not much I can do about the terminal freezing; this is tied up in the gfortran runtime, and hence out of my control.
However, I can of course attempt to fix the bug which is causing the error in the first place!
> I'm trying to compute frequencies for a whole range of models automatically, checking
> consistency only after computation. I could handle an exit code different from zero
> (I could then afterwards check what went wrong with these models), but if it doesn't
> exit at all, I can't try the next model.
>
> So you might find that the model/parameters are rubbish, but still it would be nice
> if the terminal exits. As a matter of fact, it's a pre-main sequence model, so
> perhaps the conditions to run GYRE are not satisfied. I did not check e.g. conditions
> for hydrostatic equilibrium.
>
> If I try the model right before it (profile1000 in attachment), GYRE also fails, but
> at least it exits nicely, so I'm not worried about this one. The error I get from
> that one is also appended below.
>
It seems that these models are bogus; the w parameter (third column of the input files) has a huge value at the first datapoint, and hence is (grossly) non-monotonic. This is something you should take up with Bill; but in the meantime, I'm going to put out a minor release of GYRE which checks for non-monotonicity in MESA files.
Thanks for the report!
cheers,
Rich
Re: floating point exception in example models
I'm posting this report here because I suspect it's a related issue. I might be totally wrong though.
When I try to run the non-adiabatic Beta Cep test case with the sixth-order solver, I get floating point errors in the proximity search and a crash.
On Scientific Linux 6.4 (Carbon) compiled with MESA SDK (April 2013).
W
When I try to run the non-adiabatic Beta Cep test case with the sixth-order solver, I get floating point errors in the proximity search and a crash.
Code: Select all
gyre_nad [1.1.2]
================
Compiler : GCC version 4.9.0 20130408 (experimental)
Compiler options : -I /usr/users/wball/mesa/sdk/include -march=corei7-avx -mcx16 -msahf -mno-movbe -maes -mpclmul -mpopcnt -mno-abm -mno-lwp -mno-fma -mno-fma4 -mno-xop -mno-bmi -mno-tbm -mno-avx2 -msse4.2 -msse4.1 -mno-lzcnt -mno-rtm -mno-hle -mno-rdrnd -mno-f16c -mno-fsgsbase -mno-rdseed -mno-prfchw -mno-adx -mfxsr -mxsave -mxsaveopt --param l1-cache-size=32 --param l1-cache-line-size=64 --param l2-cache-size=20480 -mtune=corei7-avx -mno-avx -mno-bmi2 -O2 -std=f2008 -finit-real=snan -ffpe-trap=invalid,overflow,zero -fbacktrace -fopenmp
OpenMP Threads : 32
MPI Processors : 1
Initialization
==============
Reading from MESA file bcep.mesa
Detected new-variant file
Adding central point
Reading from MESA file bcep.mesa
Detected new-variant file
Adding central point
Discriminant Scan
=================
Completed scan; time elapsed: 0.524 s
Scan Mode Search
================
l n_pg n_p n_g Re(omega) Im(omega) |D| n_iter
0 1 1 0 0.3274215704757669E+01 0.0000000000000000E+00 0.1394059854505166E-14 7
0 2 2 0 0.4234434532862081E+01 0.0000000000000000E+00 0.4805567002666517E-14 7
0 3 3 0 0.4944000282247599E+01 0.0000000000000000E+00 0.2982628660902119E-13 6
0 4 4 0 0.5822368119053219E+01 0.0000000000000000E+00 0.1719662209106620E-13 6
0 5 5 0 0.6955913757695812E+01 0.0000000000000000E+00 0.1700653582465684E-13 6
0 6 6 0 0.8136061463467337E+01 0.0000000000000000E+00 0.1639127751691792E-13 6
0 7 7 0 0.9335003105315893E+01 0.0000000000000000E+00 0.4604800659766906E-14 6
0 8 8 0 0.1047305038822267E+02 0.0000000000000000E+00 0.1987987621948230E-13 6
0 9 9 0 0.1155646895674471E+02 0.0000000000000000E+00 0.1204817434890942E-13 6
0 10 10 0 0.1265178176576728E+02 0.0000000000000000E+00 0.1171182718214501E-13 5
0 11 11 0 0.1379450925987289E+02 0.0000000000000000E+00 0.1142958723337299E-13 5
0 12 12 0 0.1496858597858989E+02 0.0000000000000000E+00 0.1385654726724147E-13 6
0 13 13 0 0.1615025551288126E+02 0.0000000000000000E+00 0.1078663130867263E-13 6
0 14 14 0 0.1733228072622317E+02 0.0000000000000000E+00 0.5714191030585764E-14 5
0 15 15 0 0.1851852167051163E+02 0.0000000000000000E+00 0.2969928060449353E-13 5
0 16 16 0 0.1971333272039996E+02 0.0000000000000000E+00 0.4911655227970691E-14 5
Completed mode search; time elapsed: 0.274 s
Proximity Mode Search
=====================
l n_pg n_p n_g Re(omega) Im(omega) |D| n_it
Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.
Backtrace for this error:
Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.
Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.
Backtrace for this error:
Backtrace for this error:
Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.
Backtrace for this error:
#0 0x2AAAAB6464C7
#1 0x2AAAAB646ACE
#2 0x2AAAAC48791F
#3 0x4B143D in __gyre_ext_arith_MOD_exp_r
#4 0x4B1B09 in __gyre_ext_arith_MOD_exp_c
#0 0x2AAAAB6464C7
#0 0x2AAAAB6464C7#0
0x2AAAAB6464C7
#1 0x2AAAAB646ACE
#1 0x2AAAAB646ACE
#2 0x#5 2AAAAC48791F0x
#1 458C70#2 in __gyre_ivp_magnus_MOD_solve_magnus.constprop.0 at gyre_ivp_magnus.f90:0
#6 0x45D138#3 in 0x__gyre_ivp_magnus_MOD_solve_magnus_gl64B143D
in __gyre_ext_arith_MOD_exp_r
0x2AAAAC48791F
0x2AAAAB646ACE
#4 0x4B1B09#7 in 0x__gyre_ext_arith_MOD_exp_c438FE8
in __gyre_ivp_MOD_solve#2
0x2AAAAC48791F
#5 0x#3 458C700x in 4B143D__gyre_ivp_magnus_MOD_solve_magnus.constprop.0 in at #8 __gyre_ext_arith_MOD_exp_r0x
414E19gyre_ivp_magnus.f90:0
in __gyre_nad_shooter_MOD_shoot._omp_fn.2 at gyre_nad_shooter.f90:0
#3 0x4B143D in __gyre_ext_arith_MOD_exp_r
#9 0x2AAAABBDE479
#6 0x45D138 in __gyre_ivp_magnus_MOD_solve_magnus_gl6
#4 0x4B1B09 in __gyre_ext_arith_MOD_exp_c#4
0x4B1B09 in __gyre_ext_arith_MOD_exp_c
#10 0x2AAAAC23F850
#7 0x438FE8 in __gyre_ivp_MOD_solve
#11 0x2AAAAC53D90C
#5 0x458C70#5 in 0x__gyre_ivp_magnus_MOD_solve_magnus.constprop.0458C70#12 in 0x__gyre_ivp_magnus_MOD_solve_magnus.constprop.0 at FFFFFFFFFFFFFFFFgyre_ivp_magnus.f90:0
#8 at 0xgyre_ivp_magnus.f90:0
414E19 in __gyre_nad_shooter_MOD_shoot._omp_fn.2 at gyre_nad_shooter.f90:0
#9 0x2AAAABBDE479
#6 0x45D138 in __gyre_ivp_magnus_MOD_solve_magnus_gl6
#6 0x45D138 in __gyre_ivp_magnus_MOD_solve_magnus_gl6
#10 0x2AAAAC23F850
#11 0x2AAAAC53D90C
#7 0x438FE8 in __gyre_ivp_MOD_solve
#12 0xFFFFFFFFFFFFFFFF
#7 0x438FE8 in __gyre_ivp_MOD_solve
#8 0x414E19 in __gyre_nad_shooter_MOD_shoot._omp_fn.2 at gyre_nad_shooter.f90:0
#8 0x414E19 in __gyre_nad_shooter_MOD_shoot._omp_fn.2 at #9 gyre_nad_shooter.f90:0
0x2AAAABBDE479
#9 0x2AAAABBDE479
#10 0x2AAAAC23F850
#10 0x2AAAAC23F850
#11 0x2AAAAC53D90C
#11 0x2AAAAC53D90C
#12 0xFFFFFFFFFFFFFFFF
#12 0xFFFFFFFFFFFFFFFF
W
-
rhtownsend
- Site Admin
- Posts: 397
- Joined: Sun Mar 31, 2013 4:22 pm
Re: floating point exception in example models
Hi Warrick --warrick wrote:I'm posting this report here because I suspect it's a related issue. I might be totally wrong though.
When I try to run the non-adiabatic Beta Cep test case with the sixth-order solver, I get floating point errors in the proximity search and a crash.
On Scientific Linux 6.4 (Carbon) compiled with MESA SDK (April 2013).Code: Select all
gyre_nad [1.1.2] ================ Compiler : GCC version 4.9.0 20130408 (experimental) Compiler options : -I /usr/users/wball/mesa/sdk/include -march=corei7-avx -mcx16 -msahf -mno-movbe -maes -mpclmul -mpopcnt -mno-abm -mno-lwp -mno-fma -mno-fma4 -mno-xop -mno-bmi -mno-tbm -mno-avx2 -msse4.2 -msse4.1 -mno-lzcnt -mno-rtm -mno-hle -mno-rdrnd -mno-f16c -mno-fsgsbase -mno-rdseed -mno-prfchw -mno-adx -mfxsr -mxsave -mxsaveopt --param l1-cache-size=32 --param l1-cache-line-size=64 --param l2-cache-size=20480 -mtune=corei7-avx -mno-avx -mno-bmi2 -O2 -std=f2008 -finit-real=snan -ffpe-trap=invalid,overflow,zero -fbacktrace -fopenmp OpenMP Threads : 32 MPI Processors : 1 Initialization ============== Reading from MESA file bcep.mesa Detected new-variant file Adding central point Reading from MESA file bcep.mesa Detected new-variant file Adding central point Discriminant Scan ================= Completed scan; time elapsed: 0.524 s Scan Mode Search ================ l n_pg n_p n_g Re(omega) Im(omega) |D| n_iter 0 1 1 0 0.3274215704757669E+01 0.0000000000000000E+00 0.1394059854505166E-14 7 0 2 2 0 0.4234434532862081E+01 0.0000000000000000E+00 0.4805567002666517E-14 7 0 3 3 0 0.4944000282247599E+01 0.0000000000000000E+00 0.2982628660902119E-13 6 0 4 4 0 0.5822368119053219E+01 0.0000000000000000E+00 0.1719662209106620E-13 6 0 5 5 0 0.6955913757695812E+01 0.0000000000000000E+00 0.1700653582465684E-13 6 0 6 6 0 0.8136061463467337E+01 0.0000000000000000E+00 0.1639127751691792E-13 6 0 7 7 0 0.9335003105315893E+01 0.0000000000000000E+00 0.4604800659766906E-14 6 0 8 8 0 0.1047305038822267E+02 0.0000000000000000E+00 0.1987987621948230E-13 6 0 9 9 0 0.1155646895674471E+02 0.0000000000000000E+00 0.1204817434890942E-13 6 0 10 10 0 0.1265178176576728E+02 0.0000000000000000E+00 0.1171182718214501E-13 5 0 11 11 0 0.1379450925987289E+02 0.0000000000000000E+00 0.1142958723337299E-13 5 0 12 12 0 0.1496858597858989E+02 0.0000000000000000E+00 0.1385654726724147E-13 6 0 13 13 0 0.1615025551288126E+02 0.0000000000000000E+00 0.1078663130867263E-13 6 0 14 14 0 0.1733228072622317E+02 0.0000000000000000E+00 0.5714191030585764E-14 5 0 15 15 0 0.1851852167051163E+02 0.0000000000000000E+00 0.2969928060449353E-13 5 0 16 16 0 0.1971333272039996E+02 0.0000000000000000E+00 0.4911655227970691E-14 5 Completed mode search; time elapsed: 0.274 s Proximity Mode Search ===================== l n_pg n_p n_g Re(omega) Im(omega) |D| n_it Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation. Backtrace for this error: Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation. Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation. Backtrace for this error: Backtrace for this error: Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation. Backtrace for this error: #0 0x2AAAAB6464C7 #1 0x2AAAAB646ACE #2 0x2AAAAC48791F #3 0x4B143D in __gyre_ext_arith_MOD_exp_r #4 0x4B1B09 in __gyre_ext_arith_MOD_exp_c #0 0x2AAAAB6464C7 #0 0x2AAAAB6464C7#0 0x2AAAAB6464C7 #1 0x2AAAAB646ACE #1 0x2AAAAB646ACE #2 0x#5 2AAAAC48791F0x #1 458C70#2 in __gyre_ivp_magnus_MOD_solve_magnus.constprop.0 at gyre_ivp_magnus.f90:0 #6 0x45D138#3 in 0x__gyre_ivp_magnus_MOD_solve_magnus_gl64B143D in __gyre_ext_arith_MOD_exp_r 0x2AAAAC48791F 0x2AAAAB646ACE #4 0x4B1B09#7 in 0x__gyre_ext_arith_MOD_exp_c438FE8 in __gyre_ivp_MOD_solve#2 0x2AAAAC48791F #5 0x#3 458C700x in 4B143D__gyre_ivp_magnus_MOD_solve_magnus.constprop.0 in at #8 __gyre_ext_arith_MOD_exp_r0x 414E19gyre_ivp_magnus.f90:0 in __gyre_nad_shooter_MOD_shoot._omp_fn.2 at gyre_nad_shooter.f90:0 #3 0x4B143D in __gyre_ext_arith_MOD_exp_r #9 0x2AAAABBDE479 #6 0x45D138 in __gyre_ivp_magnus_MOD_solve_magnus_gl6 #4 0x4B1B09 in __gyre_ext_arith_MOD_exp_c#4 0x4B1B09 in __gyre_ext_arith_MOD_exp_c #10 0x2AAAAC23F850 #7 0x438FE8 in __gyre_ivp_MOD_solve #11 0x2AAAAC53D90C #5 0x458C70#5 in 0x__gyre_ivp_magnus_MOD_solve_magnus.constprop.0458C70#12 in 0x__gyre_ivp_magnus_MOD_solve_magnus.constprop.0 at FFFFFFFFFFFFFFFFgyre_ivp_magnus.f90:0 #8 at 0xgyre_ivp_magnus.f90:0 414E19 in __gyre_nad_shooter_MOD_shoot._omp_fn.2 at gyre_nad_shooter.f90:0 #9 0x2AAAABBDE479 #6 0x45D138 in __gyre_ivp_magnus_MOD_solve_magnus_gl6 #6 0x45D138 in __gyre_ivp_magnus_MOD_solve_magnus_gl6 #10 0x2AAAAC23F850 #11 0x2AAAAC53D90C #7 0x438FE8 in __gyre_ivp_MOD_solve #12 0xFFFFFFFFFFFFFFFF #7 0x438FE8 in __gyre_ivp_MOD_solve #8 0x414E19 in __gyre_nad_shooter_MOD_shoot._omp_fn.2 at gyre_nad_shooter.f90:0 #8 0x414E19 in __gyre_nad_shooter_MOD_shoot._omp_fn.2 at #9 gyre_nad_shooter.f90:0 0x2AAAABBDE479 #9 0x2AAAABBDE479 #10 0x2AAAAC23F850 #10 0x2AAAAC23F850 #11 0x2AAAAC53D90C #11 0x2AAAAC53D90C #12 0xFFFFFFFFFFFFFFFF #12 0xFFFFFFFFFFFFFFFF
W
The MAGNUS_GL6 solver is likely the problem -- it can have problems doing non-adiabatic calculations (specifically, floating-point overflows) unless the grid is very fine in the region of the star where the thermal timescale is long. You might want to try adding this shoot_grid namelist to the input file (after the other shoot_grid namelists):
(and try changing alpha_thm; larger values will result in finer grids). Let me know how you get on!Code: Select all
&shoot_grid op_type = 'RESAMP_THERMAL' alpha_thm = 0.1 /
cheers,
Rich