[Pw_forum] Question on required computing resources for pwcond.x

[Yangchuan Li]

Dear all,

When I try to run pwcond.x on a functionalized CNT model (200 atoms). I set
ewind = 4.0, epsproj = 1.d-9,  nz1 = 11. And I got a error message like the
following.

******** relevant output lines *********
 ngper, shell number =         5337         534
 ngper, n2d =         5337        1054
---  E-Ef =    0.0000000  k =    0.0000000   0.0000000
---  ie =          1  ik =          1
*** Error in `pwcond.x': free(): invalid next size (normal):
0x0000000003f80d90 ***

Program received signal SIGABRT: Process abort signal.

Backtrace for this error:
#0  0x7F36C70B1777
#1  0x7F36C70B1D7E
#2  0x7F36C65CFD3F
#3  0x7F36C65CFCC9
#4  0x7F36C65D30D7
#5  0x7F36C660C393
#6  0x7F36C661866D
#7  0x42444B in jbloch_ at jbloch.f90:216
#8  0x408EA1 in compbs_ at compbs.f90:245
#9  0x4129D1 in do_cond_ at do_cond.f90:521
#10  0x403853 in pwcond at condmain.f90:22
#11  0x7F36C65BAEC4
--------------------------------------------------------------------------
mpirun noticed that process rank 1 with PID 8321 on node east exited on
signal 6 (Aborted).
--------------------------------------------------------------------------
********* end of relevant output **********

Later I run the same code on another workstation and it gives me the
following error message.

forrtl: severe (41): insufficient virtual memory

I think this may means the model is too big for the computing resource. So,
My question is how many computing resources approximately is needed for a
200 atoms system when running pwcond.x? Does change ewind and epsproj will
improve the situation?
Any comments will be appreciated. Thanks!

Best,
Yangchuan Li
Graduate Research Assistant
UT Austin | ME
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20160415/b70935d3/attachment.html>