Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Monitor memory usage of VT components #1009

Closed
lifflander opened this issue Aug 31, 2020 · 1 comment · Fixed by #1013
Closed

Monitor memory usage of VT components #1009

lifflander opened this issue Aug 31, 2020 · 1 comment · Fixed by #1013
Assignees
@cz4rs
Labels
pri: high type: task

Comments

@lifflander
Copy link
Collaborator

What Needs to be Done?

We've been trying to solve some memory growth in EMPIRE. However, it's difficult to tell where memory increase might occur. Thus, the idea is to add a virtual component method for reporting memory usage across the all the VT components. To do this, we should add serializers (that count the size at the least---doesn't have to serialize properly) to the VT components and then invoke checkpoint::getSize on the component.
cz4rs added a commit that referenced this issue Sep 1, 2020
@cz4rs
#1009 add serialize method for ArgConfig component
6a34657
@cz4rs cz4rs mentioned this issue Sep 1, 2020
Merged
cz4rs added a commit that referenced this issue Sep 8, 2020
@cz4rs
#1009 add getMemoryFootprint method for ArgConfig component
e5b91fe
cz4rs added a commit that referenced this issue Sep 9, 2020
@cz4rs
#1009 add serialize method for ArgConfig component
2d3b91a
cz4rs added a commit that referenced this issue Sep 10, 2020
@cz4rs
#1009 add serialize method for ArgConfig component
2a2c766
cz4rs added a commit that referenced this issue Sep 10, 2020
@cz4rs
#1009 add serialize method for theConfig component
6b1aee1
cz4rs added a commit that referenced this issue Sep 10, 2020
@cz4rs
#1009 add serialize method for theTimeTrigger component
1d5dddc
cz4rs added a commit that referenced this issue Sep 11, 2020
@cz4rs
#1009 add serialize method for theLBManager component
e80cfe9 
problematic types (see members not serialized):
- LBType (ambiguous call)
- LoadModel (pure abstract interface)
cz4rs added a commit that referenced this issue Sep 11, 2020
@cz4rs
#1009 add serialize method for theStatsReader component
950ce1e
@PhilMiller
Copy link
Member

Following on this, we'll want to expand on our test suite to

  • Run the serialization sanitizer over all of the components
  • Add tests that exercise essentially everything over many iterations, to hunt for growth

cz4rs added a commit that referenced this issue Sep 11, 2020
@cz4rs
#1009 add serialize method for theNodeStats component
b3f87dd
cz4rs added a commit that referenced this issue Sep 12, 2020
@cz4rs
#1009 add serialize method for theMemUsage component
62c2ac0
cz4rs added a commit that referenced this issue Sep 12, 2020
@cz4rs
#1009 add serialize method for theHandleRDMA component
f589662
cz4rs added a commit that referenced this issue Sep 12, 2020
@cz4rs
#1009 add serialize method for theCollective component
8b4b86a
cz4rs added a commit that referenced this issue Sep 15, 2020
@cz4rs
#1009 add serialize method for theConfig component
f5bce90
cz4rs added a commit that referenced this issue Sep 15, 2020
@cz4rs
#1009 add serialize method for theTimeTrigger component
0e419d5
cz4rs added a commit that referenced this issue Sep 15, 2020
@cz4rs
#1009 add serialize method for theLBManager component
7791182 
problematic types (see members not serialized):
- LBType (ambiguous call)
- LoadModel (pure abstract interface)
cz4rs added a commit that referenced this issue Sep 15, 2020
@cz4rs
#1009 add serialize method for theStatsReader component
7056275
cz4rs added a commit that referenced this issue Sep 15, 2020
@cz4rs
#1009 add serialize method for theNodeStats component
d95eac5
cz4rs added a commit that referenced this issue Sep 15, 2020
@cz4rs
#1009 add serialize method for theMemUsage component
61221ee
cz4rs added a commit that referenced this issue Sep 15, 2020
@cz4rs
#1009 add serialize method for theHandleRDMA component
6803a2e
cz4rs added a commit that referenced this issue Sep 15, 2020
@cz4rs
#1009 add serialize method for theCollective component
69c1637
cz4rs added a commit that referenced this issue Sep 16, 2020
@cz4rs
#1009 add serialize method for theConfig component
5342621
cz4rs added a commit that referenced this issue Sep 16, 2020
@cz4rs
#1009 add serialize method for theTimeTrigger component
ed34573
cz4rs added a commit that referenced this issue Sep 16, 2020
@cz4rs
#1009 add serialize method for theLBManager component
a7f78ac 
problematic types (see members not serialized):
- LBType (ambiguous call)
- LoadModel (pure abstract interface)
cz4rs added a commit that referenced this issue Sep 16, 2020
@cz4rs
#1009 add serialize method for theStatsReader component
cf048de
cz4rs added a commit that referenced this issue Sep 16, 2020
@cz4rs
#1009 add serialize method for theNodeStats component
8f173e6
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: complete support for RDMA Manager
c5142b2 
Serialize static templated container using type-erased lambdas.
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: remove virtual serialization
9b22874
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: print for actual type of live component
2141c8f
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: show components' footprint on startup and…
ac089a7 
… shutdown
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: add command-line argument
2e8ef95
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: remove unused serializers
830cc3a 
After removing virtual serialization, these serializers are no longer
used.
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: test RDMA footprint
61383e7
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: clear out removed variables after rebase
a7d548b
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: add support for component::Diagnostic
bb25f07
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: improve code coverage
505a057
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: remove unnnecessary includes
171083d 
Remove unnecessary checkpoint headers and enabe_ifs (virtual serialization
leftovers).
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: enable printing by default
a0a1384 
Verify no errors in CI with memory footprinting turned on by default.
Note: switch back to disabled by default before merging.
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: disable printing by default
2e699ca
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: formatting corrections
101a365 
- use standard formatting
- remove superfluous blank lines
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: display warning for unsupported components
79728ce
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: improve formatting and units
92db8e8
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: improve test coverage
9cbdd9a
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: improve MsgSharedPtr support
52e25f8
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: limit MsgSharedPtr serializer to footprin…
24ee3af 
…ting
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: add missing members to serializers
809f826 
Add any missing members to existing serializers and add serializers for
classes involved.
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: add serializer for DiagnosticStatsPack
df266a4 
Add serializer for DiagnosticStatsPack and use it in derived classes.
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: use branch for checkpoint
a292ff4
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: use | operator for MPI types
fccc102
cz4rs added a commit that referenced this issue Nov 21, 2020
@cz4rs
#1009: memory footprinting: use | operator for std::list iterators
951f236
cz4rs added a commit that referenced this issue Nov 24, 2020
@cz4rs
#1009: memory footprinting: use | operator for std::list iterators
852a74e
cz4rs added a commit that referenced this issue Nov 24, 2020
@cz4rs
#1009: memory footprinting: use develop branch for checkpoint
5faf0bd 
This reverts commit a292ff4.
cz4rs added a commit that referenced this issue Nov 24, 2020
@cz4rs
#1009: memory footprinting: do not use countBytes when | operator is …
f424afb 
…available
cz4rs added a commit that referenced this issue Nov 25, 2020
@cz4rs
Merge pull request #1013 from DARMA-tasking/1009-components-memory-usage
65408f6 
#1009 add memory footprinting support for VT components
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@cz4rs cz4rs

Labels
pri: high type: task
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

#1009 add memory footprinting support for VT components DARMA-tasking/vt
3 participants
@PhilMiller @lifflander @cz4rs