Hi,
I was quite disappointed when comparing optimization times of my 7-port design with ADS 2004A and ser IV LIBRA ver. 6.6.
Designs are the same in both cases, having 601 points, 18 optimization goals, 25 rounds (gradient opt.) and only 6 optimization variables (just to compare running times). ADS execution time 742.23 seconds versus 242 seconds with ser IV. So the difference is enormous. Can you give me a good explanation ? Why does it take os long time to optimize the same circuit with ADS ?
Erkki
I was quite disappointed when comparing optimization times of my 7-port design with ADS 2004A and ser IV LIBRA ver. 6.6.
Designs are the same in both cases, having 601 points, 18 optimization goals, 25 rounds (gradient opt.) and only 6 optimization variables (just to compare running times). ADS execution time 742.23 seconds versus 242 seconds with ser IV. So the difference is enormous. Can you give me a good explanation ? Why does it take os long time to optimize the same circuit with ADS ?
Erkki
Another possibility is the tolerance. Are you running with the same tolerance in both softwares? In ADS by default all the tolerance are set to a strict setting. This means that:
V_RelTol=1e-6
V_AbsTol=1e-6 V
I_RelTol=1e-6
I_AbsTol=1e-12 A
I do not recall what these use to be in Series IV.