I compiled AMBER 8 in both in serial and parallel mode. The machine is
dual processor (dual core) Xeon - OS - Red Hat Linux
uname -a gives
Linux localhost.localdomain 2.4.21-4.ELsmp #1 SMP Fri Oct 3 17:52:56 EDT
2003 i686 i686 i386 GNU/Linux
I applied all bug fixes (using bugfix.all).
The sander test runs were all passed.
When I tested the serial and parallel for a longer job with the following
input file:
----------------------------------------------
&cntrl
IREST = 1,
NTX = 5,
NTB = 2,
NTT = 1, TAUTP = 1.0,
TEMPI = 300.0,
TEMP0 = 300.0,
NTP = 1,
NSTLIM= 1000, DT = .002,
NTC = 2,
NTF = 2,
NTPR = 25, NTWX = 400,NTWV=1000,
&end
---------------------------------------------------
upto 400 steps the results are identical (i.e. if we consider 4 digits
after decimal - as sander prints). However, after that the two simulations
start to diverge slowly.
-------------------------------------------------------------------------
For PARALLEL
NSTEP = 400 TIME(PS) = 2337.800 TEMP(K) = 301.88 PRESS =
-146.6
Etot = -69050.9144 EKtot = 17607.3844 EPtot =
-86658.2987
NSTEP = 425 TIME(PS) = 2337.850 TEMP(K) = 299.56 PRESS =
-121.8
Etot = -69048.6159 EKtot = 17471.7576 EPtot =
-86520.3735
......................................................
..............................................
NSTEP = 1000 TIME(PS) = 2339.000 TEMP(K) = 298.45 PRESS =
-103.1
Etot = -69039.0240 EKtot = 17407.0499 EPtot =
-86446.0739
------------------------------------------------------------------------------
For SERIAL
NSTEP = 400 TIME(PS) = 2337.800 TEMP(K) = 301.88 PRESS =
-146.6
Etot = -69050.9144 EKtot = 17607.3844 EPtot =
-86658.2987
NSTEP = 425 TIME(PS) = 2337.850 TEMP(K) = 299.56 PRESS =
-121.8
Etot = -69048.6160 EKtot = 17471.7575 EPtot =
-86520.3735
..........................
.......................
NSTEP = 1000 TIME(PS) = 2339.000 TEMP(K) = 300.35 PRESS =
-33.4
Etot = -69044.6579 EKtot = 17518.1645 EPtot =
-86562.8224
----------------------------------------------------------------
Has anyone experienced the same?
Thanks,
Pradipta
-----------------------------------------------------------------------------------
This email was sent from IIT Guwahati Webmail. If you are not the intended recipient, please contact the sender by email and delete all copies; your cooperation in this regard is appreciated.
http://www.iitg.ernet.in
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Apr 05 2006 - 06:07:04 PDT
