Improvements towards figuring out #118 #120
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
It will be useful in some places of the code to measure memory usage but not necessarily to display it as a normal "report" in the logs. Thus, splitting metrics collection from reporting will help us reuse the collection bits in other places of the code. While doing this I also took the chance of simplifying how these memory collection metrics are performed. For starters, we don't need to know about that many fields, as we're usually only interested on VM/RSS curernt/peak usage. Additionally we can use a single file as the source of the information, which should help speeding up this data collection as well. Signed-off-by: Rodrigo Tobar <[email protected]>
Pretty names are unnecessarily long: they don't really give any useful information, specially because we already have the location in the file where the report is generated Signed-off-by: Rodrigo Tobar <[email protected]>
Memory reporting doesn't really need to use the Options object anymore (it was previously used to keep track of peak memory usage, which the kernel already does for us); additionally we already have the file/line location showing up on the log statement when we generate a memory report, so there's no need to have the information twice in the same line. Signed-off-by: Rodrigo Tobar <[email protected]>
These were not used in reality, or were previously not used correctly, so there's no need to keep them around. If we need some memory tracking functionality in the future, we can reinstate them in a different form. Signed-off-by: Rodrigo Tobar <[email protected]>
VR's main function didn't have a high-level try/catch block to handle uncaught exceptions rising from the code, leaving the behavior of the program to the fate of whatever default the C++ runtime library provides. This commit adds such a block. When an uncaught exception is received, an error is logged, and then either an MPI_Abort is issued in COMM_WORLD, or a simpler exit is invoked on non-MPI builds. Signed-off-by: Rodrigo Tobar <[email protected]>
This might help us understand what exactly is causing #118 (if logs are correctly flushed and don't get too scrambled). Signed-off-by: Rodrigo Tobar <[email protected]>
3d-runRelevant configuration parameters full file:
R_200crit v/s Mass_200crit: R_200mean v/s Mass_200mean: Mass_200crit histogram: Mass_200mean histogram: R_200crit histogram: R_200mean histogram: 6d-runRelevant configuration parameters full file:
R_200crit v/s Mass_200crit: R_200mean v/s Mass_200mean: Mass_200crit histogram: Mass_200mean histogram: R_200crit histogram: R_200mean histogram: |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
These are simple changes that, while not necessarily fixing #118, are self-contained, and might help figuring out where the memory is blowing exactly.