rcthomas / es

Elementary Supernova Spectrum Synthesis
http://c3.lbl.gov/es
Other
12 stars 8 forks source link

problem installing synapps #17

Open svalenti opened 6 years ago

svalenti commented 6 years ago

I have Ubuntu 14.04.5 LTS and I installed syn++ and synapps. I was able to run syn++, but not synapps. this is the error I get:

$mpirun -num-cores 2 synapps manual_synapps.yaml ERROR: Program requires at least 2 processors.

but I should have 32 CPU. did I install incorrectly appspack?

This is what I get at the end of the appspack configuration: configure: ======================== Build Configuration ============================= configure: Serial C Compiler : gcc configure: Serial C++ Compiler : g++ configure: MPI C++ Compiler : mpic++ configure: C++ Compile flags : -O3 configure: BLAS : -lcblas -lf77blas -latlas configure: LAPACK : -llapack configure: =====================================================================

and this is the end of the configuration for es: configure: ======================== Build Configuration ============================= configure: C++ Compiler : g++ configure: MPI C++ Compiler : mpic++ configure: C++ Compile flags : -O3 configure: OpenMP : Detected (-fopenmp) configure: CFITSIO : -lcfitsio configure: BLAS : -lcblas -lf77blas -latlas configure: LAPACK : -llapack configure: APPSPACK : -lappspack configure: =======================================================================

rcthomas commented 6 years ago

Looks like something's wrong with your MPI/APPSPACK setup, that's awfully hard for me to debug remotely. Can you run test MPI/APPSPACK problems successfully?

rcthomas commented 6 years ago

BTW should it be mpirun -np?

svalenti commented 6 years ago

I think i fixed the problem with appspack (examples now works) but I still get an error? Below is the error. Any idea what is still going wrong ? thanks s.

mpirun -np 5 synapps manual_synapps.yaml APPSPACK::Constraints::assertFeasible - Infeasible point terminate called after throwing an instance of 'char const*' [dark:04190] Process received signal [dark:04190] Signal: Aborted (6) [dark:04190] Signal code: (-6) [dark:04190] [ 0] /lib/x86_64-linux-gnu/libc.so.6(+0x36cb0) [0x7fb15b575cb0] [dark:04190] [ 1] /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x37) [0x7fb15b575c37] [dark:04190] [ 2] /lib/x86_64-linux-gnu/libc.so.6(abort+0x148) [0x7fb15b579028] [dark:04190] [ 3] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(_ZN9__gnu_cxx27verbose_terminate_handlerEv+0x155) [0x7fb15c093535] [dark:04190] [ 4] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0x5e6d6) [0x7fb15c0916d6] [dark:04190] [ 5] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0x5e703) [0x7fb15c091703] [dark:04190] [ 6] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0x5e922) [0x7fb15c091922] [dark:04190] [ 7] /usr/local/lib/libappspack.so.0(ZNK8APPSPACK11Constraints6Linear5errorERKSsS3+0xb5) [0x7fb15d69cff5] [dark:04190] [ 8] /usr/local/lib/libappspack.so.0(_ZNK8APPSPACK11Constraints6Linear14assertFeasibleERKNS_6VectorE+0x60) [0x7fb15d6a21f0] [dark:04190] [ 9] /usr/local/lib/libappspack.so.0(_ZN8APPSPACK6Solver22initializeBestPointPtrERNS_9Parameter4ListERKNS_11Constraints6LinearERNS_8Combiner7GenericE+0x1c8) [0x7fb15d6b4898] [dark:04190] [10] /usr/local/lib/libappspack.so.0(_ZN8APPSPACK6SolverC1ERKNS_9Parameter4ListERNS_8Executor9InterfaceERKNS_11Constraints6LinearE+0x66) [0x7fb15d6b6126] [dark:04190] [11] synapps() [0x40f940] [dark:04190] [12] /lib/x86_64-linux-gnu/libc.so.6(libc_start_main+0xf5) [0x7fb15b560f45] [dark:04190] [13] synapps() [0x42330e] [dark:04190] End of error message

mpirun noticed that process rank 0 with PID 4190 on node dark exited on signal 6 (Aborted).

rcthomas commented 6 years ago

The error is infeasible point:

https://c3.lbl.gov/es/#why-does-synapps-crash-right-away-with-an-infeasible-point-message

Probably you're specifying a combination of parameters that are out of bounds, etc. Check to make sure your parameter starting points are in your specified bounds, etc. I can probably look at it if you email the yaml to me.

Only 5 processors isn't really going to help much unless you're just interested in trying to fit one ion, BTW!