Merging the new op: nlp_kv_cache_unpad_to_sharded into main #8459
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
caixunshiren
requested review from
TT-BrianLiu,
tt-aho and
mywoodstock
as code owners
|
Perhaps a more generic name |
I think this is a great name! @cglagovichTT what do you think? |
...ry/nlp_tms/kernels/dataflow/reader_unary_unpad_dims_interleaved_start_id_shard_optimized.cpp
Show resolved
Hide resolved
cglagovichTT
requested changes
May 14, 2024
...ry/nlp_tms/kernels/dataflow/reader_unary_unpad_dims_interleaved_start_id_shard_optimized.cpp
Outdated
Show resolved
Hide resolved
tt_eager/tt_dnn/op_library/nlp_tms/nlp_kv_cache_unpad_to_sharded.cpp
Outdated
Show resolved
Hide resolved
caixunshiren
force-pushed
the
nlp-unpad2sharded
branch
from
56333e8
to
6e96e86
Compare
cglagovichTT
requested changes
May 14, 2024
tests/tt_eager/python_api_testing/unit_testing/misc/test_nlp_kv_cache_load_slice.py
Outdated
Show resolved
Hide resolved
cglagovichTT
approved these changes
May 14, 2024
|
Looks good! |
|
All Post commit run passed. Good to merge. Would be good if you can approve it? @TT-BrianLiu @tt-aho |
caixunshiren
force-pushed
the
nlp-unpad2sharded
branch
from
7d0a8d7
to
945104e
Compare
caixunshiren
force-pushed
the
nlp-unpad2sharded
branch
from
945104e
to
0c3704d
Compare
TT-BrianLiu
approved these changes
May 15, 2024
tests/tt_eager/python_api_testing/unit_testing/misc/test_nlp_kv_cache_load_slice.py
Outdated
Show resolved
Hide resolved
tests/tt_eager/python_api_testing/unit_testing/misc/test_nlp_kv_cache_load_slice.py
Outdated
Show resolved
Hide resolved
tests/tt_eager/python_api_testing/unit_testing/misc/test_nlp_kv_cache_load_slice.py
Outdated
Show resolved
Hide resolved
…ng on long seqlen; removed debug printout; added bfp8 test cases
…unecessary variables and conditions in host code
…leaned up host side code
caixunshiren
force-pushed
the
nlp-unpad2sharded
branch
from
677b273
to
17ae8f4
Compare
|
All post commit tests passed. Rebased to main and merged |
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.
Issue: #7379
This new op fused the unpadding and following i2s op together, with optimized noc_async_read_barrier placement, to speed up llama3 model's kv cache unpadding by 5-6x. The same op can be applied to many other transformer models.