You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Debian unstable currently has libint2-2.7.2. I built 2.9.0 with the same configure option, then installed the shared library and ran the cp2k (which was built against 2.7.2 or earlier) testsuite, which segfaults in the libint wrapper:
HFX_MEM_INFO| Est. max. program size before HFX [MiB]: 164
Thread 1 "cp2k" received signal SIGSEGV, Segmentation fault.
0x0000555556ac7c5b in hfx_libint_interface::evaluate_eri_screen (libint=..., a=..., b=..., c=..., d=..., zeta_a=4.3362376435999996, zeta_b=4.3362376435999996, zeta_c=4.3362376435999996,
zeta_d=4.3362376435999996, n_a=0, n_b=1, n_c=0, n_d=1, max_val=0, potential_parameter=..., r1=1.3379568968392406, r2=1.3379568968392406, p_work=<not associated>)
at /build/reproducible-path/cp2k-2023.2/src/hfx_libint_interface.F:1359
1359 max_val = MAX(max_val, ABS(p_work(i)))
(gdb) l
1354 CALL build_quartet_data_screen(B, A, D, C, Zeta_B, Zeta_A, Zeta_D, Zeta_C, m_max, &
1355 potential_parameter, libint, R1, R2)
1356
1357 CALL cp_libint_get_eris(n_c, n_d, n_a, n_b, libint, p_work, a_mysize)
1358 DO i = 1, mysize
1359 max_val = MAX(max_val, ABS(p_work(i)))
1360 END DO
1361 CASE (5)
1362 CALL build_quartet_data_screen(C, D, A, B, Zeta_C, Zeta_D, Zeta_A, Zeta_B, m_max, &
1363 potential_parameter, libint, R1, R2)
(gdb) bt
#0 0x0000555556ac7c5b in hfx_libint_interface::evaluate_eri_screen (libint=..., a=..., b=..., c=..., d=..., zeta_a=4.3362376435999996, zeta_b=4.3362376435999996, zeta_c=4.3362376435999996,
zeta_d=4.3362376435999996, n_a=0, n_b=1, n_c=0, n_d=1, max_val=0, potential_parameter=..., r1=1.3379568968392406, r2=1.3379568968392406, p_work=<not associated>)
at /build/reproducible-path/cp2k-2023.2/src/hfx_libint_interface.F:1359
#1 0x0000555556a99562 in hfx_screening_methods::calc_screening_functions (qs_env=<optimized out>, basis_parameter=..., lib=..., potential_parameter=..., coeffs_set=<not associated>,
coeffs_kind=<not associated>, coeffs_pgf=..., radii_pgf=..., max_set=3, max_pgf=3, n_threads=1, i_thread=0, p_work=<not associated>) at /build/reproducible-path/cp2k-2023.2/src/hfx_screening_methods.F:424
#2 0x00005555560483d8 in integrate_four_center::__hfx_energy_potential_MOD_integrate_four_center._omp_fn.0 () at /build/reproducible-path/cp2k-2023.2/src/hfx_energy_potential.F:825
#3 0x00007ffff3b26866 in GOMP_parallel (fn=fn@entry=0x55555603a390 <integrate_four_center::__hfx_energy_potential_MOD_integrate_four_center._omp_fn.0>, data=data@entry=0x7fffffffac30, num_threads=1,
num_threads@entry=0, flags=flags@entry=0) at ../../../src/libgomp/parallel.c:178
#4 0x000055555604b3b7 in hfx_energy_potential::integrate_four_center (qs_env=0x55555d1b26b0, x_data=<error reading variable: value requires 1073152 bytes, which is more than max-value-size>, ks_matrix=...,
ehfx=0, rho_ao=..., hfx_section=0x55555d221f60, para_env=0x55555d3813d0, geometry_did_change=.TRUE., irep=1, distribute_fock_matrix=.TRUE., ispin=1)
at /build/reproducible-path/cp2k-2023.2/src/hfx_energy_potential.F:407
#5 0x0000555556ae0aaa in hfx_exx::calculate_exx (qs_env=0x55555d1b26b0, unit_nr=6, hfx_sections=0x55555d221f60,
x_data=<error reading variable: value requires 1073152 bytes, which is more than max-value-size>, do_gw=.FALSE., do_admm=.FALSE., calc_forces=.FALSE., reuse_hfx=.FALSE., do_im_time=.FALSE.,
e_ex_from_gw=4.6355780447858831e-310, e_admm_from_gw=..., t3=0) at /build/reproducible-path/cp2k-2023.2/src/hfx_exx.F:229
#6 0x00005555569795b3 in mp2::mp2_main (qs_env=0x55555d1b26b0, calc_forces=.FALSE.) at /build/reproducible-path/cp2k-2023.2/src/mp2.F:743
#7 0x0000555556851d81 in qs_energy::qs_energies_mp2 (calc_forces=.FALSE., qs_env=0x55555d1b26b0) at /build/reproducible-path/cp2k-2023.2/src/qs_energy.F:170
#8 qs_energy::qs_energies (qs_env=0x55555d1b26b0, consistent_energies=<optimized out>, calc_forces=.FALSE.) at /build/reproducible-path/cp2k-2023.2/src/qs_energy.F:121
#9 0x0000555556842d1c in qs_force::qs_calc_energy_force (qs_env=<optimized out>, calc_force=.FALSE., consistent_energies=.FALSE., linres=.FALSE.) at /build/reproducible-path/cp2k-2023.2/src/qs_force.F:112
#10 0x00005555560d623a in force_env_methods::force_env_calc_energy_force (force_env=0x55555d420e00, calc_force=.FALSE., consistent_energies=<error reading variable: Cannot access memory at address 0x0>,
skip_external_control=<error reading variable: Cannot access memory at address 0x0>, eval_energy_forces=<error reading variable: Cannot access memory at address 0x0>,
require_consistent_energy_force=<error reading variable: Cannot access memory at address 0x0>, linres=<error reading variable: Cannot access memory at address 0x0>,
calc_stress_tensor=<error reading variable: Cannot access memory at address 0x0>) at /build/reproducible-path/cp2k-2023.2/src/force_env_methods.F:256
#11 0x000055555558f027 in cp2k_runs::cp2k_run (input_declaration=0x555558ea5e80, input_file_name=..., output_unit=6, mpi_comm=..., initial_variables=..., _input_file_name=_input_file_name@entry=1024,
_initial_variables=1024, _initial_variables=1024) at /build/reproducible-path/cp2k-2023.2/src/start/cp2k_runs.F:355
#12 0x0000555555591a79 in cp2k_runs::run_input (input_declaration=0x555558ea5e80, input_file_path=..., output_file_path=..., initial_variables=...,
mpi_comm=<error reading variable: Cannot access memory at address 0x0>, _input_file_path=_input_file_path@entry=1024, _output_file_path=1024, _initial_variables=1024)
at /build/reproducible-path/cp2k-2023.2/src/start/cp2k_runs.F:997
#13 0x000055555557db16 in cp2k () at /build/reproducible-path/cp2k-2023.2/src/start/cp2k.F:379
(gdb)
Looks like the API or ABI changed between 2.7.2 and 2.9.0 but the SOVERSION stayed at 2?
AFAICT, p_work is not associated after the call to libint2_build_eri, when I step into it I see this function being called:
Debian unstable currently has libint2-2.7.2. I built 2.9.0 with the same configure option, then installed the shared library and ran the cp2k (which was built against 2.7.2 or earlier) testsuite, which segfaults in the libint wrapper:
Looks like the API or ABI changed between 2.7.2 and 2.9.0 but the
SOVERSION
stayed at 2?AFAICT,
p_work
is not associated after the call tolibint2_build_eri
, when I step into it I see this function being called:The text was updated successfully, but these errors were encountered: