syncing with rccl and addressing fp8_fnuz issues #112
+257
−379
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
mberenjk
requested review from
AtlantaPepsi,
JhaShweta1,
KawtharShafie,
PedramAlizadeh,
akolliasAMD,
corey-derochie-amd,
edgargabriel,
gilbertlee-amd,
haripriyaayyalasomayajula,
mustafabar,
nileshnegi,
nusislam,
thananon and
wenkaidu
as code owners
Contributor
There was a problem hiding this comment.
Why were these added to the header?
Comment on lines
+567
to
+568
| case ncclFloat8e4m3: fp8_e4m3 = ncclVerifiablePremulScalar<rccl_float8>(rank); break; | ||
| case ncclFloat8e5m2: fp8_e5m2 = ncclVerifiablePremulScalar<rccl_bfloat8>(rank); break; |
Contributor
There was a problem hiding this comment.
Missing if/else. Likely will need to add fp8_e4m3_fnuz and fp8_e5m2_fnuz fields to the union.
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Outdated suggestions cannot be applied.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.
Details
Do not mention proprietary info or link to internal work items in this PR.
Work item: "Internal", or link to GitHub issue (if applicable).
[LWPCOMMLIBS-577]
What were the changes?
Switched to hip_fp8.h for proper FP8 support.
Ensured alignment with NCCL synchronization.
Resolved fnuz incompatibility to enhance stability.
Why were the changes made?
RCCL-Tests needs to switch over to HIP FP8 header file which has become available in recent ROCm releases.
How was the outcome achieved?
rccl-tests have been tested and passed with datatype=fp8_e4m3/fp8_e5m2.
Additional Documentation:
What else should the reviewer know?