Page 1 of 1

Successful but run/tests fails

Posted: Fri May 16, 2025 2:20 pm
by tpaudel01

The vasp (6.5.1) build with compiler/intel/20 openmpi/4.1 intel-mkl/20
keeps on crashing with a segmentation fault. (see error message below).
The build is successful with any of the makefile.include
(makefile.include.intel, makefile.include.intel_omp,
makefile.include.intel_ompi_mkl_omp) and the error is the same for each
case.

What puzzles me is that the build with the same include file works just
fine for 6.4.3. I would appreciate any help/direction to solve the issue.

I found a similar issue reported earlier, but it was for a different Intel
system and is not applicable to me.

Thank you in advance for helping me.

c2110:1948274:0:1948274] Caught signal 11 (Segmentation fault: address not
mapped to object at address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
Caught signal 11 (Segmentation fault: address not mapped to object at
address 0x440000e8)
==== backtrace (tid:1948273) ====
0 /util/opt/ucx/1.12.1/gcc/8.5.0/lib/libucs.so.0(ucs_handle_error+0x2a4)
1 /util/opt/ucx/1.12.1/gcc/8.5.0/lib/libucs.so.0(+0x2ee94)
2 /util/opt/ucx/1.12.1/gcc/8.5.0/lib/libucs.so.0(+0x2f138)
3 /util/opt/openmpi/4.1.5/intel/19.1.3/lib/libmpi.so.40(MPI_Comm_size+0x51)
4 /util/comp/intel/20/mkl/lib/intel64/libmkl_blacs_intelmpi_lp64.so(MKLMPI_Comm_size+0x29)
5 /util/comp/intel/20/mkl/lib/intel64/libmkl_blacs_intelmpi_lp64.so(mkl_blacs_init+0xb1)
6 /util/comp/intel/20/mkl/lib/intel64/libmkl_blacs_intelmpi_lp64.so(Cblacs_pinfo+0x38)
7 /util/comp/intel/20/mkl/lib/intel64/libmkl_blacs_intelmpi_lp64.so(blacs_gridmap_+0x52f)
8 /util/comp/intel/20/mkl/lib/intel64/libmkl_blacs_intelmpi_lp64.so(blacs_gridinit_+0xae)
9 /home/tpaudel/tpaudel/inst_dir/vasp.6.5.1/bin/vasp_std()
10 /home/tpaudel/tpaudel/inst_dir/vasp.6.5.1/bin/vasp_std()
11 /home/tpaudel/tpaudel/inst_dir/vasp.6.5.1/bin/vasp_std()
12 /home/tpaudel/tpaudel/inst_dir/vasp.6.5.1/bin/vasp_std()
13 /home/tpaudel/tpaudel/inst_dir/vasp.6.5.1/bin/vasp_std()
14 /home/tpaudel/tpaudel/inst_dir/vasp.6.5.1/bin/vasp_std()
15 /home/tpaudel/tpaudel/inst_dir/vasp.6.5.1/bin/vasp_std()
16 /home/tpaudel/tpaudel/inst_dir/vasp.6.5.1/bin/vasp_std()
17 /lib64/libc.so.6(__libc_start_main+0xe5)


Re: Successful but run/tests fails

Posted: Sun May 18, 2025 7:15 pm
by tpaudel01

It turns out the issue was caused by using the Intel MPI BLACS library instead of the OpenMPI one. Thank you everyone for reading..


Re: Successful but run/tests fails

Posted: Mon May 19, 2025 6:49 am
by jonathan_lahnsteiner2

Dear tpaudel,

Great you were able to solve your issue on your own.

All the Best Jonathan