I 算法迭代提示 “10 -1.#IND00e+000”
请问各位大大这是什么情况,仿真目标尺寸不大,但是因为是介质目标,所以占的内存比较大,以前仿真时从来没出现过这种情况,按我的理解那个位置应该是迭代计算的精度的,可是现在成乱码了
贴上solver logfile
Number of threads used: 16.
--------------------------------------------------------------------------------
Units settings:
Dimensions: mm
Frequency: GHz
Time: ns
--------------------------------------------------------------------------------
Boundary conditions:
YZsymmetry: none
XZsymmetry: none
XYsymmetry: none
Xmin: open
Xmax: open
Ymin: open
Ymax: open
Zmin: open
Zmax: open
--------------------------------------------------------------------------------
Mesh time : 0 s
--------------------------------------------------------------------------------
================================================================================
Integral Equation Solver calculation cycle started
--------------------------------------------------------------------------------
Solver started at: 10:16 PM Wednesday, 20. November 2013
Module: Frequency Domain: Integral Equation Solver (Win 64)
Version: 2012.4 Release from 2012-07-07 (change 216796)
--------------------------------------------------------------------------------
Using CFIE for PEC solids.
--------------------------------------------------------------------------------
Calculation 1 of 1 (Frequency: 37.5 GHz)
--------------------------------------------------------------------------------
The iterative solver (MLFMM) will be used. Preconditioner: Auto.
--------------------------------------------------------------------------------
Number of threads used: 16.
--------------------------------------------------------------------------------
Increased memory requirements by parallelization: 1.41 MB
--------------------------------------------------------------------------------
Approximate near field memory: 0.013307 GB
--------------------------------------------------------------------------------
Approximate far field memory: 5.910814 GB
--------------------------------------------------------------------------------
Stimulation with plane wave
--------------------------------------------------------------------------------
Matrix setup time : 1924 s
Degrees of freedom : 11772
Floating-point format : single
Number of surfaces : 3924
Solver order : 1st
Stimulation with : Plane wave
Accuracy : -1.#IND
Iterations : 499
Solver time : 29877 s
Far field memory : 5910.81 MB
Near field memory : 13.4008 MB
Memory for solve/precond.: 99.3807 MB
Number of Level : [ 5, 5 ]
--------------------------------------------------------------------------------
*** Error ***
Could not solve equation system.
--------------------------------------------------------------------------------
Calculation finished.
--------------------------------------------------------------------------------
Total Solver Time : 0 s
================================================================================
--------------------------------------------------------------------------------
----------------------------------------------------------------------------
Peak memory used (kB) Free physical memory (kB)
Physical Virtual At begin Minimum
----------------------------------------------------------------------------
Solver start 27060 366204 31943500 31943140
Matrices calc. 6578744 8378236 25814012 25409080
Eq. system solve 6903396 8575248 25416388 25090168
Eq. system setup 6903396 8575248 31935972 25090168
Solver run total 6903396 8575248 31943504 25090168
Computer name NODE74
----------------------------------------------------------------------------
--------------------------------------------------------------------------------
I求解器设置
相关文章:
- 关于CST的算法问题 (05-08)
- 在CST或者FDTD等时域算法计算带孔缝腔体的问题时的问题 (05-08)
- 电磁数值分析算法的比较 (05-08)
- CTS中优化算法能否自定义 (05-08)
- cst 四个优化算法中各有什么优点? (05-08)
- CST的主要算法 (05-08)