From: Jim Phillips (jim_at_ks.uiuc.edu)
Date: Fri Aug 05 2005 - 12:28:10 CDT
Those LBD messages are part of every parallel run and are quite normal.
It's just reporting how the load balancer is doing. In this case, your
maximum load is only slightly more than your average load, and there are
171 total messages being sent between processors, which looks pretty good.
The older message from Brain Bennion was a hang during load balancing,
probably due to a compiler bug (a workaround is in 2.6b1).
-Jim
On Tue, 2 Aug 2005, Jason Lee wrote:
> Hi, all:
>
> I am running a simulation for a large system. It will give following output,
> but it still runs well.
>
> LDB: LOAD: AVG 6.44591 MAX 7.94926 MSGS: TOTAL 170 MAXC 7 MAXP 16 None
> LDB: LOAD: AVG 6.44591 MAX 6.5742 MSGS: TOTAL 171 MAXC 7 MAXP 16 Refine
> PRESSURE: 1900400 1554.25 -211.193 332.088 -102.242 1944.5 -137.892 188.451
> -190.12 2021.06
> GPRESSURE: 1900400 -459.307 1.33026 358.845 106.08 -300.498 -87.3214 209.22
> -268.485 -114.336
> PRESSAVG: 1900400 -291.044 -149.405 166.55 -193.018 -128.237 -22.3247 166.934
> 66.1118 -24.2065
> GPRESSAVG: 1900400 -275.153 -137.005 164.977 -191.022 -143.709 -30.8227
> 170.748 59.8807 -27.6602
> ENERGY: 1900400 6676.8317 3999.1887 0.0000 0.0000
> -88574.7776 7693.2389 0.0000 0.0000 19183.2864
> -51022.2319 298.8598 -50711.3286 -50714.6861 299.0018
> 1839.9330 -291.3801 277848.8653 -147.8292 -148.8406
>
>
>
> I have read the following message
> http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l/0328.html
> It seems it is related with load balancer. But I still don't understand this
> issue. Will this affect the simulation accuracy?
> Thanks a lot.
> Jason
>
> _________________________________________________________________
> 与世界各地的朋友进行交流，免费下载 MSN Messenger:
> http://messenger.msn.com/cn
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:39:48 CST