Change BFLOAT16 data type/API support naming and prefix #2771
Conversation
This file contains hidden or 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
Existing BFLOAT16 support is with HALF or sh, which is not good as half precision usually points to FP16 according to IEEE definition. We'd specify this as BFLOAT16 instead of HALF. According to the discussion in issue OpenMathLib#2767, update related naming and prefix as BFLOAT16 and sb. Details as: 1. Change the generic control flag from BUILD_HALF to BUILD_BFLOAT16; 2. Change the prefix of BFLOAT16 API from shxxxx to sbxxxx; 3. Change related file names and other related with above prefix also;
Guobing-Chen
force-pushed
the
BF16_data_type_naming
branch
from
4ce6bba to
93e748d
Compare
|
Closing due to objection raised in #2767 (naming/API already agreed on with the netlib team) |
|
Reopened after further discussions in #2767, though most files will have been altered in the meantime unfortunately. (Opening a new PR may be easier than updating this - please do whatever seems more practical) |
|
superseding with #2886 due to the number of code changes since the original PR |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
Existing BFLOAT16 support is named with HALF or sh, which is not good as half
precision usually points to FP16 according to IEEE definition. We'd
specify this as BFLOAT16 instead of HALF. According to the discussion in
issue #2767, update related naming and prefix as BFLOAT16 and sb.
Details as: