Skip to content

[CK][CK_TILE] Fix FMHA codegen group mode dispatcher for fwd & bwd #3730

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
arthurliu1998 wants to merge 2 commits into
base: develop
Choose a base branch
from
Open

[CK][CK_TILE] Fix FMHA codegen group mode dispatcher for fwd & bwd #3730

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
03b7f58
Fix FMHA codegne group mode dispatch
arthurliu1998 Apr 22, 2026
c0d65e7
Merge branch 'develop' into users/ArthurLiu/ck_fmha_codegen
arthurliu1998 Apr 22, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 9 additions & 3 deletions example/ck_tile/01_fmha/codegen/ops/fmha_bwd.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -827,14 +827,20 @@ def dvcheck(self) -> str:
@property
def max_seq_q_cond(self) -> str:
if self.tile.max_seq_q != 0:
return f" && (t.seqlen_q <= {self.tile.max_seq_q})"
if self.mode == "group":
return f" && (t.max_seqlen_q <= {self.tile.max_seq_q})"
else:
return f" && (t.seqlen_q <= {self.tile.max_seq_q})"
else:
return ""

@property
def extra_cond(self) -> str:
if self.tr_load == "t" and self.tile.max_seq_q == 0 and self.tile.F_bn0 == 128 and self.tile.F_bhdq == 128:
return " && (t.seqlen_k <= 256)"
if self.mode == "group":
return " && (t.max_seqlen_k <= 256)"
else:
return " && (t.seqlen_k <= 256)"
else:
return ""

Expand Down Expand Up @@ -1057,7 +1063,7 @@ def get_bwd_blobs(
hdim = tile.F_bhdq
if (mode == "group") and (spad1d == "f"):
continue
if (mode == "group" or ("no" not in mask)) and tile.max_seq_q != 0:
if ("no" not in mask) and tile.max_seq_q != 0:
continue
if (bias == "no" or bias == "alibi") and dbias == "t":
continue
Expand Down
4 changes: 4 additions & 0 deletions example/ck_tile/01_fmha/codegen/ops/fmha_fwd.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -328,6 +328,8 @@ def seqtune(self, max_bm0: int) -> str:
if self.bm0 == max_bm0 or self.bm0 == 64:
return "true/*fall back to largest tile*/"
else:
if self.mode == "group":
return f"a.max_seqlen_q <= {self.bm0}"
return f"a.seqlen_q <= {self.bm0}"

@property
Expand Down Expand Up @@ -1136,6 +1138,8 @@ def get_pipelines(
):
pipelines.append(FmhaFwdPipeline("qr_async_trload", "row", "f", "f", "f", "f", logits, bias, lse, dropout, qscale, mask, skip, "t", sink)) # fmt: skip
pipelines.append(FmhaFwdPipeline("qr_async_trload", "row", "f", "f", "t", "t", logits, bias, lse, dropout, qscale, mask, skip, "t", sink)) # fmt: skip
pipelines.append(FmhaFwdPipeline("qr_async_trload", "row", "t", "t", "f", "f", logits, bias, lse, dropout, qscale, mask, skip, "t", sink)) # fmt: skip # group mode spad
pipelines.append(FmhaFwdPipeline("qr_async_trload", "row", "t", "t", "t", "t", logits, bias, lse, dropout, qscale, mask, skip, "t", sink)) # fmt: skip # group mode spad+dpad

# # qr_async_trload_v3 bf16/fp16 not ready
# if (hdim, hdim_v) == (128, 128):
Expand Down