Matt,<br><br>One of the people in our group profiled CP2K with the TAU to look for memory leaks.<br><a href="http://www.cs.uoregon.edu/research/tau/home.php">http://www.cs.uoregon.edu/research/tau/home.php</a><br><br>A number of them were found and some of them might be related to the parallel<br>
distribution algorithm. I can send you the .ppk so you can read the profile file in<br>more detail. The viewer is free.<br><br><div dir="ltr" align="left"><span class="691130120-13032008"><font color="#0000ff" face="Courier">The attached profile shows some memory leaks, specifically in
distribution_methods.f90 and memory_utilities.f90, but there are
others.</font></span></div>
<div dir="ltr" align="left"><span class="691130120-13032008"></span> </div>
<div dir="ltr" align="left"><span class="691130120-13032008"><font color="#0000ff" face="Courier">MEMORY LEAK! malloc size <file=cube_utils.pp.f90,
variable=info%sphere_bounds(i)%p, line=278> : PW_ENV_METHODS::PW_ENV_REBUILD
[{pw_env_methods.pp.f90} {177,12}] => CUBE_UTILS::INIT_CUBE_INFO
[{cube_utils.pp.f90} {206,14}] </font></span></div><span class="691130120-13032008"><font color="#0000ff" face="Courier">
</font><div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=distribution_methods.pp.f90, variable=nsgf, line=885> :
DISTRIBUTION_METHODS::DISTRIBUTE_MOLECULES_2D [{distribution_methods.pp.f90}
{472,14}] => DISTRIBUTION_METHODS::DISTRIBUTE_2D_COST_MODEL
[{distribution_methods.pp.f90} {809,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=distribution_optimize.pp.f90, variable=cost_cpu, line=165> :
DISTRIBUTION_METHODS::DISTRIBUTE_MOLECULES_2D [{distribution_methods.pp.f90}
{472,14}] => DISTRIBUTION_OPTIMIZE::DISTRIBUTE_2D_MONTE_CARLO
[{distribution_optimize.pp.f90} {101,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=distribution_optimize.pp.f90, variable=max_cost_para, line=271> :
DISTRIBUTION_METHODS::DISTRIBUTE_MOLECULES_2D [{distribution_methods.pp.f90}
{472,14}] => DISTRIBUTION_OPTIMIZE::DISTRIBUTE_2D_MONTE_CARLO
[{distribution_optimize.pp.f90} {101,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=f77_interface.pp.f90, variable=f_envs, line=249> : CP2K
[{cp2k.pp.f90} {38,9}] => F77_INTERFACE::INIT_CP2K
[{f77_interface.pp.f90} {208,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=input_section_types.pp.f90, variable=section%subsections, line=248>
: INPUT_CP2K::CREATE_CP2K_ROOT_SECTION [{input_cp2k.pp.f90}
{142,14}] => INPUT_SECTION_TYPES::SECTION_CREATE
[{input_section_types.pp.f90} {186,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=memory_utilities.pp.f90, variable=p, line=291> :
ATOMS_INPUT::READ_ATOMS_INPUT [{atoms_input.pp.f90} {99,14}] =>
MEMORY_UTILITIES::REALLOCATE_I1 [{memory_utilities.pp.f90} {265,14}]
</font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=memory_utilities.pp.f90, variable=p, line=704> :
BASIS_SET_TYPES::READ_GTO_BASIS_SET [{basis_set_types.pp.f90}
{800,14}] => MEMORY_UTILITIES::REALLOCATE_R3
[{memory_utilities.pp.f90} {672,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=qs_neighbor_lists.pp.f90, variable=aux_present, line=335> :
QS_ENERGY::QS_ENERGIES [{qs_energy.pp.f90} {103,14}] =>
QS_NEIGHBOR_LISTS::BUILD_QS_NEIGHBOR_LISTS [{qs_neighbor_lists.pp.f90}
{172,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=realspace_grid_types.pp.f90, variable=bounds , line=1136> :
REALSPACE_GRID_TYPES::RS_PW_TRANSFER [{realspace_grid_types.pp.f90}
{550,14}] => REALSPACE_GRID_TYPES::RS_PW_TRANSFER_DISTRIBUTED
[{realspace_grid_types.pp.f90} {749,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=realspace_grid_types.pp.f90, variable=bounds , line=956> :
REALSPACE_GRID_TYPES::RS_PW_TRANSFER [{realspace_grid_types.pp.f90}
{550,14}] => REALSPACE_GRID_TYPES::RS_PW_TRANSFER_DISTRIBUTED
[{realspace_grid_types.pp.f90} {749,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=reference_manager.pp.f90, variable=indx, line=210> :
ENVIRONMENT::CP2K_FINALIZE [{environment.pp.f90} {764,14}] =>
REFERENCE_MANAGER::PRINT_ALL_REFERENCES [{reference_manager.pp.f90}
{190,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=reference_manager.pp.f90, variable=irank, line=209> :
ENVIRONMENT::CP2K_FINALIZE [{environment.pp.f90} {764,14}] =>
REFERENCE_MANAGER::PRINT_ALL_REFERENCES [{reference_manager.pp.f90}
{190,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=reference_manager.pp.f90, variable=ival, line=208> :
ENVIRONMENT::CP2K_FINALIZE [{environment.pp.f90} {764,14}] =>
REFERENCE_MANAGER::PRINT_ALL_REFERENCES [{reference_manager.pp.f90}
{190,14}] </font></div>
<div dir="ltr" align="left"><font color="#0000ff" face="Courier"><br>MEMORY LEAK! malloc size
<file=topology_generate_util.pp.f90, variable=h_list, line=519> :
TOPOLOGY::CONNECTIVITY_CONTROL [{topology.pp.f90} {312,14}] =>
TOPOLOGY_GENERATE_UTIL::TOPOLOGY_GENERATE_BOND [{topology_generate_util.pp.f90}
{437,14}] </font></div></span><br><div class="gmail_quote">On Tue, Mar 11, 2008 at 5:18 PM, Matt W <<a href="mailto:MattWa...@gmail.com">MattWa...@gmail.com</a>> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br>
No, thankfully it's not related. Looks like another overflow<br>
though...thanks for pushing things!<br>
<br>
Matt<br>
<div class="Ih2E3d"><br>
On Mar 11, 9:03 pm, "Nichols A. Romero" <<a href="mailto:naro...@gmail.com">naro...@gmail.com</a>> wrote:<br>
> Guys,<br>
><br>
> This might be related.<br>
><br>
> I just tried H2O-8192.inp on 512 & 1024 cores and I received the error<br>
> below.<br>
><br>
> BTW, I don't have this problem with the H2O-4096.inp<br>
><br>
> *<br>
> *** ERROR in distribute_molecules_2d (MODULE distribution_methods) ***<br>
> *<br>
><br>
> *** The memory allocation for the data object <cost_ele_pair> failed. The<br>
> ***<br>
> *** requested memory size is -1835008 Kbytes<br>
> ***<br>
><br>
> *** Program stopped at line number 570 of MODULE distribution_methods ***<br>
><br>
> *<br>
> *** ERROR in distribute_molecules_2d (MODULE distribution_methods) ***<br>
> *<br>
><br>
> *** The memory allocation for the data object <cost_ele_pair> failed. The<br>
> ***<br>
> *** requested memory size is -1835008 Kbytes<br>
> ***<br>
><br>
</div>> On Tue, Mar 11, 2008 at 10:14 AM, Nichols A. Romero <<a href="mailto:naro...@gmail.com">naro...@gmail.com</a>><br>
<div class="Ih2E3d">> wrote:<br>
><br>
><br>
><br>
> > So can I. I am running that and some larger systems.<br>
><br>
> > Many thanks to Matt!<br>
><br>
</div><div class="Ih2E3d">> > On 3/11/08, Teodoro Laino <<a href="mailto:teodor...@gmail.com">teodor...@gmail.com</a>> wrote:<br>
><br>
> > > I can confirm that with this bug-fix H2O-2048 is running nicely.<br>
><br>
> > > teo<br>
><br>
> > > On 11 Mar 2008, at 12:09, Matt W wrote:<br>
><br>
> > > (not well tested though).<br>
><br>
> > --<br>
</div><div class="Ih2E3d">> > Nichols A. Romero, Ph.D.<br>
> > DoD User Productivity Enhancement and Technology Transfer (PET) Group<br>
> > High Performance Technologies, Inc.<br>
> > Reston, VA<br>
> > 443-567-8328 (C)<br>
> > 410-278-2692 (O)<br>
><br>
> --<br>
> Nichols A. Romero, Ph.D.<br>
> DoD User Productivity Enhancement and Technology Transfer (PET) Group<br>
> High Performance Technologies, Inc.<br>
> Reston, VA<br>
> 443-567-8328 (C)<br>
> 410-278-2692 (O)<br>
</div><div><div></div><div class="Wj3C7c"><br>
</div></div></blockquote></div><br><br clear="all"><br>-- <br>Nichols A. Romero, Ph.D.<br>DoD User Productivity Enhancement and Technology Transfer (PET) Group<br>High Performance Technologies, Inc.<br>Reston, VA<br>443-567-8328 (C)<br>
410-278-2692 (O)