Small benchmark test of CST MWS 5.1.1 transient solver

[navuho]

CST MWS v5.1.1 benchmark

Hi ALL !

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=

 

[wave-maniac]

CST MWS v5.1.1 benchmark

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

 

[navuho]

Re: CST MWS v5.1.1 benchmark

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.

 

[wave-maniac]

Re: CST MWS v5.1.1 benchmark

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

 

[navuho]

Re: CST MWS v5.1.1 benchmark

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

 

[wave-maniac]

CST MWS v5.1.1 benchmark

I sent you the log file.