TFactor: bulkerify GEMV (both MC and GPU) #1214
Draft
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
albestro
force-pushed
the
alby/tfactor-optim/bulk
branch
from
6558780 to
bf17fcc
Compare
Base automatically changed from
alby/tfactor-optim/no-gemv-divergence
to
master
November 22, 2024 11:28
albestro
force-pushed
the
alby/tfactor-optim/bulk
branch
4 times, most recently
from
37ea75b to
2845339
Compare
|
cscs-ci run just to check if there is any major problems |
rasolca
requested changes
Dec 4, 2024
Comment on lines
268
to
269
| matrix::Matrix<T, Device::GPU> ws_T({nworkspaces * nrefls_step, nrefls_step}, | ||
| {nrefls_step, nrefls_step}); |
There was a problem hiding this comment.
All Ts are allocated at scheduling. Better reuse it.
msimberg
approved these changes
Dec 5, 2024
albestro
force-pushed
the
alby/tfactor-optim/bulk
branch
from
05dc48d to
f83f317
Compare
rasolca
reviewed
Dec 6, 2024
|
|
||
| namespace dlaf::factorization::internal { | ||
|
|
||
| inline size_t getTFactorNWorkers() noexcept { |
There was a problem hiding this comment.
What about this to be snake case?
msimberg
reviewed
Dec 9, 2024
albestro
force-pushed
the
alby/tfactor-optim/bulk
branch
from
0397c9e to
d370894
Compare
rasolca
approved these changes
Dec 10, 2024
msimberg
approved these changes
Dec 10, 2024
|
cscs-ci run |
|
Converted to draft in order to prevent merging, since I'm still doing some checks. |
albestro
commented
Dec 11, 2024
Co-authored-by: Mikael Simberg <mikael.simberg@iki.fi>
taus are set for gemv, but then full tile is reduced, hence also the diag with taus.
tile_t might get moved by first worker before any other work is able to use its size
albestro
force-pushed
the
alby/tfactor-optim/bulk
branch
from
f0638e1 to
5997e52
Compare
in case of odd division of tiles, it might end up being selected but not used.
|
cscs-ci run |
|
Codecov ReportAttention: Patch coverage is
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. Additional details and impacted files@@ Coverage Diff @@
## master #1214 +/- ##
==========================================
- Coverage 94.38% 94.02% -0.37%
==========================================
Files 154 155 +1
Lines 9374 9461 +87
Branches 1160 1163 +3
==========================================
+ Hits 8848 8896 +48
- Misses 320 340 +20
- Partials 206 225 +19 ☔ View full report in Codecov by Sentry. |
add assert also for gpu code that calls add manually
1 task
albestro
force-pushed
the
alby/tfactor-optim/bulk
branch
from
2f90f23 to
d199b88
Compare
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 this PR we try to increase parallelisation of GEMV step of TFactor, which is by construction serial (all results goes into the same block tile), by using workspaces for storing partial results and then reducing them before the final TRMV step.
Algorithmically, the main change is that
stepGEMVloop has been replaced with a singlestepGEMVAll, and the concept has been applied in a similar way for both backends, MC and GPU:pika::bulksplits input tiles over multiple task, each one stores the partial results in their workspace and just one at the end does the reduction.In order to implement this solution, workspaces for intermediate results have been added (there is another option which does not require any additional workspace nor reduction, and that we might explore for MC in another PR).
TODO:
Close #798.
EDIT: since this is conceptually similar to #798 and it is going to be closed as soon as this gets merged, I migrated here the doc fixes happened there.