Skip to content

MOE: add AITER_MOE_FORCE_BF16_ACT to force bf16 activations (opt-in) #3593

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

Merged
zufayu merged 4 commits into from
Jul 14, 2026
Merged

MOE: add AITER_MOE_FORCE_BF16_ACT to force bf16 activations (opt-in) #3593

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
b0752af
moe: add AITER_MOE_FORCE_BF16_ACT to force bf16 activations (opt-in)
Jun 8, 2026
028a9f6
Merge branch 'main' into main
sphinx07 Jun 8, 2026
aedc6fd
style: apply black formatting
Jul 13, 2026
af81aa0
Merge branch 'main' into main
zufayu Jul 13, 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
10 changes: 8 additions & 2 deletions aiter/fused_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -596,6 +596,7 @@ def fused_moe_(
):
q_dtype_a = dtypes.fp8
bf16_fp8_bound = int(os.environ.get("AITER_BF16_FP8_MOE_BOUND", "256"))
force_bf16_act = os.environ.get("AITER_MOE_FORCE_BF16_ACT", "0") == "1"
if quant_type == QuantType.per_1x32 and q_dtype_w == dtypes.i4x2:
# a16wi4: bf16 activations, int4 weights with groupwise scale
q_dtype_a = dtypes.bf16
Expand All @@ -604,9 +605,14 @@ def fused_moe_(
q_dtype_a = dtypes.fp8
elif quant_type == QuantType.per_1x32:
if activation == ActivationType.Swiglu and gate_mode == GateMode.SEPARATED:
q_dtype_a = dtypes.bf16 if M < _SWIGLU_MXFP4_BF16_BOUND else dtypes.fp4x2
if force_bf16_act:
q_dtype_a = dtypes.bf16
else:
q_dtype_a = (
dtypes.bf16 if M < _SWIGLU_MXFP4_BF16_BOUND else dtypes.fp4x2
)
elif activation == ActivationType.Swiglu or gate_mode == GateMode.INTERLEAVE:
if get_gfx() != "gfx950" or M < bf16_fp8_bound:
if force_bf16_act or get_gfx() != "gfx950" or M < bf16_fp8_bound:
q_dtype_a = dtypes.bf16
else:
q_dtype_a = dtypes.fp8
Expand Down
Loading