From: thomas lemmin (updalpe_at_gmail.com)
Date: Mon Mar 25 2013 - 04:40:07 CDT
Dear NAMD experts,
I am trying to run a 500k bead Martini simulation on 256 CPU. After about 1
million steps, it crashes due to a Bad global angle count.
If I simply restart the simulation, it runs again smoothly for 1 millions
step and crashes again due to the angle count.
Since MARTINI bonds are larger than All atom bonds, would it be possible
that even if the simulation is stable, angles might not be contained in 2 x
2 x 2 blocks of patches?
LDB: ============= START OF LOAD BALANCING ============== 13486.1
LDB: ============== END OF LOAD BALANCING =============== 13486.1
Info: useSync: 1 useProxySync: 0
LDB: =============== DONE WITH MIGRATION ================ 13486.1
LDB: ============= START OF LOAD BALANCING ============== 13486.7
LDB: Largest compute 228212 load 0.001872 is 0.4% of average load 0.495746
LDB: Average compute 0.000234 is 0.0% of average load 0.495746
LDB: TIME 13486.8 LOAD: AVG 0.495746 MAX 0.554075 PROXIES: TOTAL 22172
MAXPE 110 MAXPATCH 7 None MEM: 650.883 MB
LDB: TIME 13486.9 LOAD: AVG 0.495746 MAX 0.554075 PROXIES: TOTAL 33893
MAXPE 214 MAXPATCH 9 RefineTorusLB MEM: 650.883 MB
LDB: TIME 13488.4 LOAD: AVG 0.495746 MAX 0.50125 PROXIES: TOTAL 34023
MAXPE 207 MAXPATCH 8 RefineTorusLB MEM: 650.883 MB
LDB: ============== END OF LOAD BALANCING =============== 13488.6
Info: useSync: 1 useProxySync: 0
LDB: =============== DONE WITH MIGRATION ================ 13488.7
FATAL ERROR: Bad global angle count!
FATAL ERROR: See http://www.ks.uiuc.edu/Research/namd/bugreport.html
Thank you
Thomas Lemmin
This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:23:05 CST