Support gfx950 layouts #692
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
zhanglx13
force-pushed
the
support_layout_gfx950
branch
2 times, most recently
from
7910a91 to
fd5c641
Compare
zhanglx13
force-pushed
the
support_layout_gfx950
branch
2 times, most recently
from
7370be5 to
6bfb674
Compare
API change: - For blocked layout, use -tensorShape, which only takes two dims as dim0,dim1 - For dot layout, use -dotShape, which takes three dims as M,N,K
Separate each layout's code into their own files
- When kWidth is large, use a smaller elemSize honrizontally to save space - Improve the labels, such as - change vec to kWidth for operands - change opA/opB to inA/inB and include operand dims - remove group dims in the operands so that they don't overlap with operand block dims - Better alignment: dot op and mfma zoomed-in pics are bottom aligned
kGroup is defined as total elements per thread / kWidth for one mfma instruction. We need kGroup = 2 only for the newly added mfma_f32_16x16x128_f8f6f4 and mfma_f32_32x32x64_f8f6f4 with f8 input type on MI350.
And print mfma instruction name accordingly. For now, mixed precision mfma between 8-bit and 4- or 6-bit is not supported yet.
- Support data types - Support both 32 and 64 banks - Still working on LDS accesses
- Fixed the issue with maxPhase computation. Need to submit a PR to fix it in the triton compiler - For ds_read_b64 with 64 banks, there are bank conflicts. We need to figure out a different swizzling pattern to avoid bank conflicts.
Assumed a basic global access pattern
mfma_transpose_load instructions - Elements along the M/N dim are contiguous in both global memory and LDS. Note that this is not the in-thread transpose case. - Swizzling is disabled
zhanglx13
force-pushed
the
support_layout_gfx950
branch
2 times, most recently
from
80d0c5f to
1b331cb
Compare
|
Aside from using it a bit to get a feel if there's anything hard to use, the only thing I can comment is that it might be worthwhile to break the python file down to several pieces to be more modular. The tex files are now broken into several ones already so it makes sense to have python follow the same thing, for readability and future-proof. The |
jtang10
approved these changes
Jan 24, 2025
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.
No description provided.