You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have a simple question about the possibility to parallize the computation. Your example and tuition is clear and i went through quite easily, thank you.
Now I try to utilize your package on my mission. My mission is quite simple but more complex than the sample math equation. My mission is about find out the best parameter for a single factor which includes 2-5 parameters. Each round od calculation takes about 2minutes, and the possibility of the combination of parameters is from 2000 to 100000. I found that the package only use one core of the server. May you give me some suggestion on parallazation.
I noticed that i could input the vector to speed up little as x or use gpu. But i would like to stick to cpu and use thread thing to achieve the parallization. Do you think this idea is good? I have few 32 cores or 24 cores severs , I would like to run 20core on optimization. I think it is beyond the vector solution, right?
Thanks
The text was updated successfully, but these errors were encountered:
Dear Guofei,
I have a simple question about the possibility to parallize the computation. Your example and tuition is clear and i went through quite easily, thank you.
Now I try to utilize your package on my mission. My mission is quite simple but more complex than the sample math equation. My mission is about find out the best parameter for a single factor which includes 2-5 parameters. Each round od calculation takes about 2minutes, and the possibility of the combination of parameters is from 2000 to 100000. I found that the package only use one core of the server. May you give me some suggestion on parallazation.
I noticed that i could input the vector to speed up little as x or use gpu. But i would like to stick to cpu and use thread thing to achieve the parallization. Do you think this idea is good? I have few 32 cores or 24 cores severs , I would like to run 20core on optimization. I think it is beyond the vector solution, right?
Thanks
The text was updated successfully, but these errors were encountered: