Avoid zero-initialization of packing buffers #70
Merged
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
robertknight
force-pushed
the
gemm-pack-uninit-2
branch
from
99b4ef4 to
d20a369
Compare
Avoid overhead from zeroing memory when prepacking matrix inputs by allowing the packing functions to write to uninitialized memory. In a TTS demo using Tacotron2 [^1], which uses an LSTM-based decoder, this reduces the "real-time factor" (degree of slowdown over real-time generation) from ~1.9x to ~1.65x. [^1]: https://github.com/robertknight/xd-tts
robertknight
force-pushed
the
gemm-pack-uninit-2
branch
from
d20a369 to
87c0171
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Outdated suggestions cannot be applied.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.
Avoid overhead from zeroing memory when prepacking matrix inputs by allowing the packing functions to write to uninitialized memory. The packing functions guarantee that the entire buffer passed to them is initialized when they return.
In a TTS demo using Tacotron2 1, which uses an LSTM-based decoder, this reduces the "real-time factor" (degree of slowdown over real-time generation) from ~1.9x to ~1.65x.
TODO:
Vec::spare_capacity_mutcan in theory return a larger capacity than we asked for.Footnotes
https://github.com/robertknight/xd-tts ↩