common: MemoryModel: performance improvements and API changes#58115
Merged
common: MemoryModel: performance improvements and API changes#58115
Conversation
Signed-off-by: Samuel Just <sjust@redhat.com> (cherry picked from commit 720ce78)
as 'snap' as a struct name does not follow the naming convention, and 'snap' is way overloaded in the codebase. Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
Allowing clients to hold on to the MemoryModel object for multiple "snapshots", saving (per measurements) ~40% of the sampling run time. Also: separating the parsing of /proc/maps (which consumes most of the remaining execution time) into a separate function. This will allow us to create a "no need for precise heap" fast mode (in a future commit). Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
We're going to add this functionality to other components, simplify this one first to match. Signed-off-by: Samuel Just <sjust@redhat.com> Signed-off-by: Shreekara <sshreeka@redhat.com> (cherry picked from commit f337bda)
Also - stopping the parsing of /proc/status once all the needed fields have been found, The resulting code is measurably faster than the previous version. Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
leave error reporting to the caller. This change achieves two targets: - no need for a context when creating the MemoryModel object; and - (more importantly) the error messages are now reported in the context of the caller, and can be made to appear only once in the logs. The existing client of the MemoryModel is modified to reflect the changes to its API. Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
Main points: - using charconv as a faster alternative to strtoll functions; - only parsing the addresses for relevant lines; - lines ending with '[heap]' are OK Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
in the object itself Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
922c353 to
d699e23
Compare
Contributor
Author
|
jenkins test make check |
Contributor
Author
|
@vshankar - can you review? thanks |
Contributor
Ofcourse 👍 (I will plan to have a look in the next few days) |
mchangir
reviewed
Jul 15, 2024
Contributor
mchangir
left a comment
There was a problem hiding this comment.
just a small query towards improving performance
Comment on lines
+123
to
+127
| if (cmp_against(ln, "VmSize:", s.size) || | ||
| cmp_against(ln, "VmRSS:", s.rss) || cmp_against(ln, "VmHWM:", s.hwm) || | ||
| cmp_against(ln, "VmLib:", s.lib) || | ||
| cmp_against(ln, "VmPeak:", s.peak) || | ||
| cmp_against(ln, "VmData:", s.data)) { |
Contributor
There was a problem hiding this comment.
Would it be too much to push all required items in a hash table and test for the presence of the prefix of the current line in the hash table instead ?
Contributor
Author
There was a problem hiding this comment.
Would it be too much to push all required items in a hash table and test for the presence of the prefix of the current line in the hash table instead ?
I do not see a way to do have better performance this way (I had originally measured multiple ideas I could think of). For example - note that we have sub-strings of at least two lengths here.
vshankar
reviewed
Jul 15, 2024
Contributor
vshankar
left a comment
There was a problem hiding this comment.
MDCache change look fine.
Contributor
|
This PR is under test in https://tracker.ceph.com/issues/67020. |
vshankar
approved these changes
Aug 5, 2024
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
Performance: per my off-line (code snippets) benchmarks
(see benchtest):
The API changes:
The PR includes work by @athanatos & @shreekarSS,
picked from #54698