AMBER Archive (2005)

Subject: AMBER: The effect appling paches

From: Yagi Toru (yagi_at_bestsystems.co.jp)
Date: Tue Apr 19 2005 - 08:11:54 CDT


Dear all,

I'm writing about a current patch, bugfix.all (bugfix1 - 42)

After applied the patch, bugfix.all,
the results of some tests changes from original results.
Some of them do not seem within numerical errors.
Are these differences fair as a result of appling the patch?
How can I check the reasonability?

The system I used is as followings
CPU: Xeon 3.06 GHz (not EM64T)
OS : RedHat WS 3.0
Compiler: Intel 8.0 or PGI 5.2-4

For example, I show a part of $AMBERHOME/test/noper/mdout.cap.dif file
76c76
< Etot = -3409.9453 EKtot = 2056.0343 EPtot =
-5465.9796

---
 >  Etot   =     -3404.4787  EKtot   =      2056.0343  EPtot      = 
-5460.5130
79c79,80
<  EELEC  =    -10563.1142  EHBOND  =         0.0000  RESTRAINT  = 
    0.0000
---
 >  EELEC  =    -10563.1142  EHBOND  =         0.0000  RESTRAINT  = 
     5.4666
 >  EAMBER (non-restraint)  =     -5465.9796
82c83
<  Etot   =     -3123.6124  EKtot   =      2059.0258  EPtot      = 
-5182.6382
---
 >  Etot   =     -3118.1049  EKtot   =      2059.0258  EPtot      = 
-5177.1307
85c86,87
<  EELEC  =    -10278.2360  EHBOND  =         0.0000  RESTRAINT  = 
    0.0000
 >  EELEC  =    -10278.2360  EHBOND  =         0.0000  RESTRAINT  = 
     5.5075
 >  EAMBER (non-restraint)  =     -5182.6382
....

Any comment will be appreciated.

Toru YAGI

----------------------------------------------------------------------- The AMBER Mail Reflector To post, send mail to amber_at_scripps.edu To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu