abstractions for input and model (preparing for distributed impl.) #21
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
Helps remove unnecessary bloat from model file. Also required in distributed memory mode, where input data will be too large. Inputs can be loaded back when used after restore.
- clear input and model of unnecessary data before save
- results in smaller saved model
- but requires input files to be read while using model
- localize shared memory entities before save, as they can not be deserialized
tanmaykm
force-pushed
the
distributed
branch
2 times, most recently
from
aa215e2 to
e9d10a5
Compare
|
travis build passes now |
|
@andreasnoack FYI. This is another large-scale parallel problem that we are attempting with the yahoo dataset. |
Makes things more efficient. Also essential in larger distributed memory conditions where such mappings must be avoided.
- modernize travis.yml - disable Logging - provide empty `@threads` macro on Julia v0.4
check for null case before sharing
|
@abhijithch bump |
tanmaykm
added a commit
that referenced
this pull request
Jan 28, 2016
abstractions for input and model (preparing for distributed impl.)
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Outdated suggestions cannot be applied.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.
In preparation for implementing distributed memory computation, I have created abstractions for
InputsandModeltypes.The current shared memory and thread computation use
SharedMemoryInputsandSharedMemoryModelas their respective concrete implementations. Under the hood, the shared memory implementation usesSparseMatrixCSCandMatrixas before. Distributed memory computation would replace them with something else appropriate.I find no performance regressions with the abstractions in place, and these would be useful for trying out different computation models. It would be good to merge this separately. Will follow this up with another PR for a distributed memory implementation.