微波EDA网,见证研发工程师的成长!
首页 > 研发问答 > 微波和射频技术 > 电磁仿真讨论 > How to set acceleration in CST?

How to set acceleration in CST?

时间:03-30 整理:3721RD 点击:
im trying to accelerate the simulation. im using i7-5820k CPU @ 3.3GHz 6 core, 128 RAM, Tesla K40c and Quadro K620 (graphics cards).

im using time domain solver in CST. i set CPU and Hardware acceleration. i thick/select for multithreading (CPU) up to 48 threads (the max) and i thick/select hardware acceleration to 8 devices (max). is that grantee maximum speed using my system? i dont think its using the hardware acceration. i feel something still missing

Check with the hardware tool from cst if your Tesla card is recognized. Inspect the task manager how many cores are used during the simulation. Read the log-file after the simulation run if if the Tesla card is used. Do you have enough token for hardware acceleration?

hi johnjoe,
thanks for the reply. the tesla card is supported by CST. i have 50 tokens. is that enough? how to make sure its accelerating?

Hi,

you can check by running the simulation with and without hardware acceleration. Just check the time stamps in the log file, depending on your simulation, there should be a difference. And second, before the transient time-domain simulation starts, it should be written something like "prepare hardware acceleration".

Greetings

hi,
how to measure the acceleration and increase if i want more?

The acceleration performance is measured by the difference in time of sim with and without hardware acc. You can increase the performance by better cpu or more and/or faster tesla cards.

Copyright © 2017-2020 微波EDA网 版权所有

网站地图

Top