MPICH folks confirmed it's an MPICH problem and an issue is created at https://github.com/pmodels/mpich/issues/6843
--Junchao Zhang On Tue, Dec 12, 2023 at 7:53 PM Junchao Zhang <junchao.zh...@gmail.com> wrote: > I was able to reproduce it. Let me ask MPICH developers. > > --Junchao Zhang > > > On Tue, Dec 12, 2023 at 3:06 PM Randall Mackie <rlmackie...@gmail.com> > wrote: > >> It now seems to me that petsc+mpich is no longer valgrind clean, or I am >> doing something wrong. >> >> A simple program: >> >> >> Program test >> >> #include "petsc/finclude/petscsys.h" >> use petscsys >> >> PetscInt :: ierr >> >> call PetscInitialize(PETSC_NULL_CHARACTER,ierr) >> call PetscFinalize(ierr) >> >> end program test >> >> >> PETSc compiled in debug mode, complex scalars, and download-mpich, when >> run with valgrind generates errors like these: >> >> ==3997== Syscall param writev(vector[...]) points to uninitialised byte(s) >> ==3997== at 0x8C31867: writev (writev.c:26) >> ==3997== by 0x9C20DE4: MPL_large_writev (mpl_sock.c:31) >> ==3997== by 0x9BF1050: MPIDI_CH3I_Sock_writev (sock.c:2689) >> ==3997== by 0x9BF9812: MPIDI_CH3_iStartMsgv (ch3_istartmsgv.c:92) >> ==3997== by 0x9BA7790: MPIDI_CH3_EagerContigSend (ch3u_eager.c:191) >> ==3997== by 0x9BCA7EC: MPID_Send (mpid_send.c:132) >> ==3997== by 0x9BCAC64: MPID_Send_coll (mpid_send.c:206) >> ==3997== by 0x9A2AC7C: MPIC_Send (helper_fns.c:126) >> ==3997== by 0x993A645: MPIR_Bcast_intra_binomial >> (bcast_intra_binomial.c:146) >> ==3997== by 0x99FF64A: MPIR_Bcast_allcomm_auto (mpir_coll.c:323) >> ==3997== by 0x99FFC06: MPIR_Bcast_impl (mpir_coll.c:420) >> ==3997== by 0x99FCF86: MPID_Bcast (mpid_coll.h:30) >> ==3997== by 0x99FFE13: MPIR_Bcast (mpir_coll.c:465) >> ==3997== by 0x974A513: internal_Bcast (bcast.c:93) >> ==3997== by 0x974A72B: PMPI_Bcast (bcast.c:143) >> ==3997== by 0x4B8D6DB: PETScParseFortranArgs_Private (zstart.c:182) >> ==3997== by 0x4B8DDFA: PetscInitFortran_Private (zstart.c:200) >> ==3997== by 0x4B34931: PetscInitialize_Common (pinit.c:974) >> ==3997== by 0x4B8E8C7: petscinitializef_ (zstart.c:284) >> ==3997== by 0x4959434: __petscsys_MOD_petscinitializenohelp >> (petscsysmod.F90:374) >> ==3997== Address 0x1ffeffcac0 is on thread 1's stack >> ==3997== in frame #4, created by MPIDI_CH3_EagerContigSend >> (ch3u_eager.c:160) >> ==3997== Uninitialised value was created by a stack allocation >> ==3997== at 0x9BA7601: MPIDI_CH3_EagerContigSend (ch3u_eager.c:160) >> ==3997== >> >> ==3997== Syscall param write(buf) points to uninitialised byte(s) >> ==3997== at 0x8C2B697: write (write.c:26) >> ==3997== by 0x9BF0F1D: MPIDI_CH3I_Sock_write (sock.c:2614) >> ==3997== by 0x9BF7AAE: MPIDI_CH3_iStartMsg (ch3_istartmsg.c:68) >> ==3997== by 0x9BA7A27: MPIDI_CH3_EagerContigShortSend >> (ch3u_eager.c:262) >> ==3997== by 0x9BCA766: MPID_Send (mpid_send.c:119) >> ==3997== by 0x9BCAC64: MPID_Send_coll (mpid_send.c:206) >> ==3997== by 0x9A2AC7C: MPIC_Send (helper_fns.c:126) >> ==3997== by 0x993A645: MPIR_Bcast_intra_binomial >> (bcast_intra_binomial.c:146) >> ==3997== by 0x99FF64A: MPIR_Bcast_allcomm_auto (mpir_coll.c:323) >> ==3997== by 0x99FFC06: MPIR_Bcast_impl (mpir_coll.c:420) >> ==3997== by 0x99FCF86: MPID_Bcast (mpid_coll.h:30) >> ==3997== by 0x99FFE13: MPIR_Bcast (mpir_coll.c:465) >> ==3997== by 0x974A513: internal_Bcast (bcast.c:93) >> ==3997== by 0x974A72B: PMPI_Bcast (bcast.c:143) >> ==3997== by 0x4DB95A2: PetscOptionsGetenv (pdisplay.c:61) >> ==3997== by 0x4E0D745: PetscStrreplace (str.c:572) >> ==3997== by 0x4AC8DEA: PetscOptionsFilename (options.c:416) >> ==3997== by 0x4ACF0B5: PetscOptionsInsertFile (options.c:632) >> ==3997== by 0x4AD3CB5: PetscOptionsInsert (options.c:861) >> ==3997== by 0x4B8E0EF: PetscInitFortran_Private (zstart.c:206) >> ==3997== Address 0x1ffeff7998 is on thread 1's stack >> ==3997== in frame #3, created by MPIDI_CH3_EagerContigShortSend >> (ch3u_eager.c:223) >> ==3997== Uninitialised value was created by a stack allocation >> ==3997== at 0x9BA788F: MPIDI_CH3_EagerContigShortSend >> (ch3u_eager.c:223) >> ==3997== >> >> Is this a known issue or am I doing something wrong? >> >> Thanks, Randy >> >
