[Pw_forum] Erroneous CPU Time Report

[Paul M. Grant]

Lorenzo, I've pasted below the info you requested.

BTW, I'm new to the forum, and hope this is the way to respond (add) to a
thread.

More info.  I recompiled Espresso using the latest Intel compiler and still
zero or small cpu times are reported.  I suspect this behavior may involve
the fact that I have a dual processor MB.  Also, from the recent post by
Paolo Giannozzi, I gather my situation emerges rather often!

-Paul 

PWSCF        :     0.00s CPU time,   48.80s wall time

     init_run     :     0.00s CPU
     electrons    :     0.00s CPU


     electrons    :     0.00s CPU
     c_bands      :     0.00s CPU (      31 calls,   0.000 s avg)
     sum_band     :     0.00s CPU (      31 calls,   0.000 s avg)
     v_of_rho     :     0.00s CPU (      31 calls,   0.000 s avg)
     v_h          :     0.00s CPU (      31 calls,   0.000 s avg)
     v_xc         :     0.00s CPU (      31 calls,   0.000 s avg)
     newd         :     0.00s CPU (      31 calls,   0.000 s avg)
     mix_rho      :     0.00s CPU (      31 calls,   0.000 s avg)

     c_bands      :     0.00s CPU (      31 calls,   0.000 s avg)
     init_us_2    :     0.00s CPU (      63 calls,   0.000 s avg)
     cegterg      :     0.00s CPU (      31 calls,   0.000 s avg)

     sum_band     :     0.00s CPU (      31 calls,   0.000 s avg)
     becsum       :     0.00s CPU (      31 calls,   0.000 s avg)
     addusdens    :     0.00s CPU (      31 calls,   0.000 s avg)

     cegterg      :     0.00s CPU (      31 calls,   0.000 s avg)
     h_psi        :     0.00s CPU (     112 calls,   0.000 s avg)
     g_psi        :     0.00s CPU (      80 calls,   0.000 s avg)
     diaghg       :     0.00s CPU (     110 calls,   0.000 s avg)
     update       :     0.00s CPU (      80 calls,   0.000 s avg)
     last         :     0.00s CPU (      31 calls,   0.000 s avg)

     h_psi        :     0.00s CPU (     112 calls,   0.000 s avg)
     init         :     0.00s CPU (     112 calls,   0.000 s avg)
     add_vuspsi   :     0.00s CPU (     112 calls,   0.000 s avg)
     s_psi        :     0.00s CPU (     112 calls,   0.000 s avg)

     General routines
     ccalbec      :     0.00s CPU (      31 calls,   0.000 s avg)
     cft3         :     0.00s CPU (     373 calls,   0.000 s avg)
     cft3s        :     0.00s CPU (    1858 calls,   0.000 s avg)
     davcio       :     0.00s CPU (      31 calls,   0.000 s avg)


 

 


-----Original Message-----
From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf
Of Lorenzo Paulatto
Sent: Tuesday, February 06, 2007 12:23 AM
To: pw_forum at pwscf.org
Subject: Re: [Pw_forum] Erroneous CPU Time Report

> While running the complete set of Espresso 3.2 PWscf examples
> (successfully), I noted that the elapsed CPU time reported is either zero
> or
> some unrealistically small value such as 0.02s.  The "wall time" is fine.
> My motherboard is a dual Xeon processor Supermicro, OS is SuSE 10.1, and
> fortran compiler G95.  Any suggestions?  Thanks.
>

Could you post an example of the time part of your output, like this one:
     PWSCF        :  1m45.98s CPU time

     init_run     :     5.84s CPU
     electrons    :    99.58s CPU
     forces       :     0.50s CPU

     electrons    :    99.58s CPU
     c_bands      :    84.10s CPU (       5 calls,  16.819 s avg)
     sum_band     :    15.35s CPU (       5 calls,   3.070 s avg)
     v_of_rho     :     0.08s CPU (       5 calls,   0.015 s avg)
     mix_rho      :     0.01s CPU (       5 calls,   0.002 s avg)

     c_bands      :    84.10s CPU (       5 calls,  16.819 s avg)
     init_us_2    :     0.77s CPU (     720 calls,   0.001 s avg)
     cegterg      :    83.09s CPU (     300 calls,   0.277 s avg)

     sum_band     :    15.35s CPU (       5 calls,   3.070 s avg)

     wfcrot       :     5.31s CPU (      60 calls,   0.089 s avg)
     cegterg      :    83.09s CPU (     300 calls,   0.277 s avg)
     h_psi        :    70.91s CPU (    1269 calls,   0.056 s avg)
     g_psi        :     0.30s CPU (     909 calls,   0.000 s avg)
     overlap      :     4.58s CPU (     909 calls,   0.005 s avg)
     cdiaghg      :     2.60s CPU (    1209 calls,   0.002 s avg)
     update       :     3.80s CPU (     909 calls,   0.004 s avg)
     last         :     2.45s CPU (     360 calls,   0.007 s avg)

     h_psi        :    70.91s CPU (    1269 calls,   0.056 s avg)
     init         :     0.61s CPU (    1269 calls,   0.000 s avg)
     firstfft     :    30.13s CPU (   13485 calls,   0.002 s avg)
     secondfft    :    27.76s CPU (   13485 calls,   0.002 s avg)
     add_vuspsi   :     0.94s CPU (    1269 calls,   0.001 s avg)

     General routines
     ccalbec      :     1.43s CPU (    1329 calls,   0.001 s avg)
     cft3         :     0.08s CPU (      23 calls,   0.003 s avg)
     cft3s        :    59.82s CPU (   31770 calls,   0.002 s avg)
     davcio       :     0.86s CPU (    1020 calls,   0.001 s avg)

thanks, LP


----------------------------------------------------------------
  SISSA Webmail https://webmail.sissa.it/
  Powered by SquirrelMail http://www.squirrelmail.org/

_______________________________________________
Pw_forum mailing list
Pw_forum at pwscf.org
http://www.democritos.it/mailman/listinfo/pw_forum