Skip to content

Crash using LINCOA on LUMI-G #316

Open

Description

@asenzz

Below is a stack trace; the binary and libprima are built using amdclang++ from ROCm 6.4.4. The crash doesn't occur when using COBYLA.

where full
#0  0x0000000000000000 in ?? ()

#1  0x000015554c7fa986 in lincoa_c::calfun () at /scratch/project_465003024/repo/prima/c/lincoa_c.f90:181

#2  0x00001555450ec4d3 in evaluate_mod::evaluatef () at /scratch/project_465003024/repo/prima/fortran/common/evaluate.f90:130

#3  0x0000155545177b67 in initialize_lincoa_mod::initxf () at /scratch/project_465003024/repo/prima/fortran/lincoa/initialize.f90:224

#4  0x0000155545190b19 in lincob_mod::lincob () at /scratch/project_465003024/repo/prima/fortran/lincoa/lincob.f90:256
        trtol = 0.01
#5  0x000015554519f9cd in lincoa_mod::lincoa () at /scratch/project_465003024/repo/prima/fortran/lincoa/lincoa.f90:529

#6  0x000015554c7fa509 in lincoa_c_mod::lincoa_c () at /scratch/project_465003024/repo/prima/c/lincoa_c.f90:136
        rhoend_loc = 0x1536db7fff90
        rhobeg_loc = 0x1536db7ff700
        maxfun_loc = 0x1536db7ff970
        npt_loc = 0x1536db7fef00
#7  0x000015554c7ff2fe in prima_minimize () from /project/project_465003024/lib64/libprimac.so

#8  0x000015554b46a4f6 in _ZN3svr9optimizer6pprune14pprune_biteoptEjRKSt8functionIFvPKdPdEEddRKN4arma3MatIdEEbj.omp_outlined_debug__.187(const uint32_t &, std::multimap<double, std::pair<unsigned int, unsigned int>, std::less<double>, std::allocator<std::pair<double const, std::pair<unsigned int, unsigned int> > > > &, svr::optimizer::pprune *, std::shared_ptr<std::deque<svr::optimizer::t_calfun_data*, std::allocator<svr::optimizer::t_calfun_data*> > > &, std::deque<std::deque<svr::optimizer::t_biteopt_prune, std::allocator<svr::optimizer::t_biteopt_prune> >, std::allocator<std::deque<svr::optimizer::t_biteopt_prune, std::allocator<svr::optimizer::t_biteopt_prune> > > > &, const unsigned long &, const int &, double &, double &, const uint32_t &, svr::t_omp_lock &, unsigned int) const (.global_tid.=0x154e1ffe2260, .bound_tid.=0x154e1ffe2258, prima_particles=@0x7ffffffdaef4: 200, scores=..., this=0x7ffffffde1b0, p_particles=..., 
    biteopt_particle=..., D_size=@0x7ffffffdb010: 32568, npt=@0x7ffffffdaf98: 26, rhobeg=@0x7ffffffdb150: 0.050000019552431545, rhoend=@0x7ffffffdb148: 1.000000391048631e-10, prima_iter=@0x7ffffffdaf9c: 27, res_l=..., .capture_expr.=200)
    at /flash/project_465003024/repo/tempus-core/SVRRoot/OnlineSVR/src/pprune.cpp:517
        i = @0x154e1ffe1ff8: 238
        particle_i = 238
        problem = {
          n = 4071,
          calfun = 0x15554b45c270 <svr::optimizer::pprune::calfun(double const*, double*, svr::optimizer::t_calfun_data*)>,
          calcfc = 0x0,
          x0 = 0x1536db801740,
          xl = 0x60748bc0,
          xu = 0x60750af8,
          m_ineq = 0,
          Aineq = 0x0,
          bineq = 0x0,
          m_eq = 0,
          Aeq = 0x0,
          beq = 0x0,
          m_nlcon = 0,
          f0 = nan(0x8000000000000),
          nlconstr0 = 0x0
        }
        biteopt = @0x60763690: {
          <std::__shared_ptr<svr::optimizer::t_biteopt_cost, (__gnu_cxx::_Lock_policy)2>> = {
            <std::__shared_ptr_access<svr::optimizer::t_biteopt_cost, (__gnu_cxx::_Lock_policy)2, false, false>> = {<No data fields>}, 
            members of std::__shared_ptr<svr::optimizer::t_biteopt_cost, (__gnu_cxx::_Lock_policy)2>:
            _M_ptr = 0x15382ca6ced0,
            _M_refcount = {
              _M_pi = 0x15382ca6cec0
            }
          }, <No data fields>}
        this_x0 = 0x1536db801740
        d = @0x154e1ffe1ffc: 0
        prima_options = {
          rhobeg = 0.050000019552431545,
          rhoend = 1.000000391048631e-10,
          maxfun = 27,
          iprint = 1,
          ftarget = -inf,
          npt = 26,
          ctol = 1.4901161193847656e-08,
          data = 0x15383b7fbf80,
          callback = 0x15554b460a60 <svr::optimizer::prima_progress_callback(int, double const*, double, int, int, double, int, double const*, bool*)>
        }
        prima_result = {
          x = 0x1536db809690,
          f = nan(0x8000000000000),
          cstrv = nan(0x8000000000000),
          nlconstr = 0x1536db7ff310,
          nf = 0,
          status = PRIMA_RESULT_INITIALIZED,
          success = false,
          message = 0x0
        }
        rc = 5461
        .omp.iv = 70
        .capture_expr. = 200
        it = 0
        .omp.lb = 70
        .omp.ub = 71
        .omp.stride = 200
        .omp.is_last = 0
#9  0x000015554b46ab82 in _ZN3svr9optimizer6pprune14pprune_biteoptEjRKSt8functionIFvPKdPdEEddRKN4arma3MatIdEEbj.omp_outlined.193(void) const (.global_tid.=0x154e1ffe2260, .bound_tid.=0x154e1ffe2258, prima_particles=@0x7ffffffdaef4: 200, 
    scores=..., this=0x7ffffffde1b0, p_particles=..., biteopt_particle=..., D_size=@0x7ffffffdb010: 32568, npt=@0x7ffffffdaf98: 26, rhobeg=@0x7ffffffdb150: 0.050000019552431545, rhoend=@0x7ffffffdb148: 1.000000391048631e-10, 
    prima_iter=@0x7ffffffdaf9c: 27, res_l=..., .capture_expr.=140733193388034) at /flash/project_465003024/repo/tempus-core/SVRRoot/OnlineSVR/src/pprune.cpp:488

#10 0x000015554a042349 in __kmp_invoke_microtask () from /appl/lumi/SW/LUMI-25.09/G/EB/rocm/6.4.4/lib/llvm/lib/libomp.so

#11 0x0000155549fc132f in __kmp_invoke_task_func () from /appl/lumi/SW/LUMI-25.09/G/EB/rocm/6.4.4/lib/llvm/lib/libomp.so

#12 0x0000155549fbff8c in __kmp_launch_thread () from /appl/lumi/SW/LUMI-25.09/G/EB/rocm/6.4.4/lib/llvm/lib/libomp.so

#13 0x000015554a022728 in __kmp_launch_worker(void*) () from /appl/lumi/SW/LUMI-25.09/G/EB/rocm/6.4.4/lib/llvm/lib/libomp.so

#14 0x0000155547212f6c in start_thread () from /lib64/libc.so.6

#15 0x000015554729a388 in clone3 () from /lib64/libc.so.6

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions