Repository navigation
The speed of running openfast in multi-thread parallel mode using python is too slow #3460
Unanswered
laitiandeng-maker
asked this question in
Q&A
Replies: 1 comment 5 replies
|
Which version of OpenFAST are you using? Is it compiled with OpenMP enabled? |
5 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Hello OpenFAST community,




When using multithreaded parallelism to call OpenFAST from Python, I encountered the problem of excessively long runtimes. When I ran a single-threaded 500 s simulation (with a turbulent wind field also 500 s) of the OC4 jacket, it took only 30 minutes (as shown in Figure 1). However, when I ran 18 threads concurrently, it took 4 hours and 49 minutes to finish 18 cases (as shown in Figure 2). Theoretically, when memory is sufficient, if multithreaded concurrent simulation is used, all 18 cases should finish in about 30 minutes. The parallel Python code that calls openfast.exe is shown in Figure 3, and the computer configuration is shown in Figure 4. I have already tried different parallel codes, but I still cannot achieve true parallel processing speed. Could you please help answer my confusion and questions?
Best regards
All reactions