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
There is a small performance degradation between 2023.04 and the 2024.01-alpha2.
For example, for ESM4.2, Main loop clocks are almost the same between 2024.01-alpha2 and 2023.04. The Initialization clock has increased a bit from 96secs to 125secs.
This is related to #1388. nearest_index is a bit slower than the indp (this was used in horiz_interp in 2023.04). Particularly the part of the code that checks that the array passed in is monotonically increasing/decreasing is not very efficient
To Reproduce
Compare initialization times between 2023.04 and 2024.01-alpha2.
Expected behavior
Performance of FMS should be as best as possible.
System Environment
Describe the system environment, include:
This happens in any system, but I saw this in the amd box with the data_override unit tests and in GAEA.
Additional context
N/A
The text was updated successfully, but these errors were encountered:
Describe the bug
There is a small performance degradation between 2023.04 and the 2024.01-alpha2.
For example, for ESM4.2, Main loop clocks are almost the same between 2024.01-alpha2 and 2023.04. The Initialization clock has increased a bit from 96secs to 125secs.
This is related to #1388.
nearest_index
is a bit slower than the indp (this was used in horiz_interp in 2023.04). Particularly the part of the code that checks that the array passed in is monotonically increasing/decreasing is not very efficientTo Reproduce
Compare initialization times between 2023.04 and 2024.01-alpha2.
Expected behavior
Performance of FMS should be as best as possible.
System Environment
Describe the system environment, include:
This happens in any system, but I saw this in the amd box with the data_override unit tests and in GAEA.
Additional context
N/A
The text was updated successfully, but these errors were encountered: