Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.
Sign upBug when calculating Cv and internal energy on replica exchanged systems #5
Comments
ghost
assigned
toastedcrumpets
Jun 4, 2012
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
toastedcrumpets
Jun 27, 2012
Owner
This appears to be due to DynamO forgetting to tell the output plugins about any accumulated free stream time from cell events. It looks like this needs to be removed from DynamO (it will simplify the code somewhat too).
|
This appears to be due to DynamO forgetting to tell the output plugins about any accumulated free stream time from cell events. It looks like this needs to be removed from DynamO (it will simplify the code somewhat too). |
This comment has been minimized.
Show comment
Hide comment
This comment has been minimized.
toastedcrumpets
Jun 28, 2012
Owner
This issue has been fixed. It was due to a free stream accumulator optimization which has since been removed. This commit marks the final change required to fix the issue.
e5c882b
|
This issue has been fixed. It was due to a free stream accumulator optimization which has since been removed. This commit marks the final change required to fix the issue. |
toastedcrumpets commentedJun 4, 2012
The time averaging of properties seems to be broken when performing replica exchange simulations. This is not a critical bug, as the affected properties (Cv and U) are typically calculated using the energy histograms for these simulations, and these work correctly. It is probably due to a miscounting of the time passed in each simulation after a swap.