CST MWS v5.1.1 benchmark
I did a small benchmark test of CST MWS transient solver.
The task I was running is TE11 circular pin polarizer (lossfree) with meshsize ~ 1e6 sells.
Here are results :
------ OS--------------------CPU------------------------- Memory-------------------TIME
1. Win2003 ----------Xeon-3GHz x 2-------------------4GB, PC-2100---------90 min
============= (Intel-E7505, 533MHz)======133 MHz, CL2.5=
2. WXP-64bit------- Athlon64-3800+ (2.4 GHz)-------2GB, PC-2700---------116 min
============= (NForce3-Ultra, 1000MHz)====200MHz, CL2.5=
3. WXP--------------P4-3.36 GHz-----------------------4GB, PC-3200---------126 min
=============(i865P, 840 MHz)==========210 MHz, CL3.0=
4. WXP-------------- Athlon64-3800+ (2.4 GHz)-------2GB, PC-2700---------132 min
============= (NForce3-Ultra, 1000MHz)====200MHz, CL2.5=
5. Win2000-------- AthlonMP-1900+ x2 (1.6 GHz)-----2GB, PC-2100--------134 min
============= (AMD-762, 266MHz)========133MHz, CL2.0=
6. WXP--------------P4-3.2 GHz-------------------------2GB, PC-3200---------151 min
=============(i865P, 800 MHz)==========200 MHz, CL3.0=
Dear navuho,
Thanks for the info. I really appreciate it.
1) Could you please upload either the 5.1 version of simulation project or the SAT file along with the necessary info to setup the simulation project for V5.1. I'm going to run the same benchmark on P4 3.8 GHz.
2) Would you please mention the type, size and speed of the RAM and also the bus speed for each case?
Best regards
wave-maniac
wave-maniac
1) Could you please upload
OK. See attached file.
Just run transient solver and see a log-file
2) Would you please mention the type, size and speed of the RAM
OK. Updated.
Dear navuho,
Thanks for the file and updating the benchmark. I ran the simulation twice without changing anything. Here is the data:
------ OS----------------------------CPU---------------------------------------------- Memory--------------------------------------TIME
WinXP Pro ------------P4/3.8GHz EM64T,1MB/800---------------2GB, dual-channel DDR2 533 ECC -------------35min and 47 sec
I would appreciate similar benchmarks for HFSS.
Best regards
wave-maniac
wave-maniac
It's very strange result. Too fast :)
Did you run Transient Solver ?
Check the mesh size (1e6) and accuracy (-40 dB)
If all is ok, please send me solver log-file
I sent you the log file.