Skip to content

Merge q,kv #1199

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
hiworldwzj merged 4 commits into from
Feb 3, 2026
Merged

Merge q,kv #1199

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
228b3b7
autotune
WANDY666 Jan 26, 2026
c52089c
Merge branch 'main' of https://github.com/ModelTC/LightLLM into dev_q…
WANDY666 Jan 29, 2026
3c1313c
merge qkv
WANDY666 Jan 29, 2026
45ce36a
fix assert.
Feb 3, 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
1 change: 1 addition & 0 deletions lightllm/common/basemodel/layer_weights/meta_weights/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,7 @@
ROWMMWeight,
KVROWNMMWeight,
ROWBMMWeight,
QKVROWNMMWeight,
COLMMWeight,
)
from .norm_weight import TpRMSNormWeight, RMSNormWeight, LayerNormWeight, NoTpGEMMANormWeight, QKRMSNORMWeight
Expand Down
2 changes: 1 addition & 1 deletion lightllm/common/basemodel/layer_weights/meta_weights/mm_weight/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
from .mm_weight import (
MMWeightTpl,
)
from .rowmm_weight import ROWMMWeight, KVROWNMMWeight, ROWBMMWeight
from .rowmm_weight import ROWMMWeight, KVROWNMMWeight, ROWBMMWeight, QKVROWNMMWeight
from .colmm_weight import COLMMWeight
44 changes: 44 additions & 0 deletions lightllm/common/basemodel/layer_weights/meta_weights/mm_weight/rowmm_weight.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -92,6 +92,50 @@ def _get_tp_padded_head_num(self, head_num: int):
)


class QKVROWNMMWeight(MMWeightTpl):
def __init__(
self,
in_dim: int,
q_head_num: int,
kv_head_num: int,
head_dim: int,
weight_names: Union[str, List[str]],
data_type: torch.dtype,
bias_names: Optional[Union[str, List[str]]] = None,
quant_method: QuantizationMethod = None,
tp_rank: int = None,
tp_world_size: int = None,
) -> None:
self.tp_rank_ = tp_rank if tp_rank is not None else get_current_rank_in_dp()
self.tp_world_size_ = tp_world_size if tp_world_size is not None else get_dp_world_size()
self.repeat_times = 1
assert q_head_num % self.tp_world_size_ == 0, (
f"q_head_num must be divisible by tp_world_size_, " f"but found: {q_head_num} % {self.tp_world_size_}"
)
assert kv_head_num % self.tp_world_size_ == 0, (
f"kv_head_num must be divisible by tp_world_size_" f"but found: {kv_head_num} % {self.tp_world_size_}"
)
q_hidden_size = (q_head_num // self.tp_world_size_) * head_dim
kv_hidden_size = (kv_head_num // self.tp_world_size_) * head_dim
out_dims = [q_hidden_size, kv_hidden_size, kv_hidden_size]
super().__init__(
in_dim=in_dim,
out_dims=out_dims,
weight_names=weight_names,
data_type=data_type,
bias_names=bias_names,
quant_method=quant_method,
tp_rank=self.tp_rank_,
tp_world_size=self.tp_world_size_,
)
self.param_slicer = get_row_slice_mixin(
self.quant_method.method_name,
tp_rank=self.tp_rank_,
tp_world_size=self.tp_world_size_,
repeat_times=self.repeat_times,
)


class ROWBMMWeight(BMMWeightTpl):
def __init__(
self,
Expand Down
83 changes: 83 additions & 0 deletions ...uted_weight=true,out_dtype=torch.bfloat16,topk_num=1,use_fp8_w8a8=false}_NVIDIA_H200.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,83 @@
{
"1024": {
"BLOCK_SIZE_K": 64,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 64,
"NEED_TRANS": false,
"num_stages": 2,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 64,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 16,
"NEED_TRANS": false,
"num_stages": 3,
"num_warps": 4
},
"2048": {
"BLOCK_SIZE_K": 32,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 16,
"NEED_TRANS": false,
"num_stages": 3,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 64,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"NEED_TRANS": false,
"num_stages": 2,
"num_warps": 4
},
"512": {
"BLOCK_SIZE_K": 64,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"NEED_TRANS": false,
"num_stages": 4,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 64,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"NEED_TRANS": false,
"num_stages": 2,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 32,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"NEED_TRANS": false,
"num_stages": 2,
"num_warps": 4
},
"800": {
"BLOCK_SIZE_K": 64,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"NEED_TRANS": false,
"num_stages": 2,
"num_warps": 4
},
"8192": {
"BLOCK_SIZE_K": 64,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"NEED_TRANS": false,
"num_stages": 2,
"num_warps": 4
}
}
Comment on lines +1 to +83
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 2, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

The keys in this JSON configuration file are not sorted, which makes it difficult to read and maintain. For better readability, it's recommended to sort the keys numerically. This is especially helpful when manually inspecting or debugging the autotuning configurations.

{
  "8": {
    "BLOCK_SIZE_K": 32,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 64,
    "GROUP_SIZE_M": 1,
    "NEED_TRANS": false,
    "num_stages": 2,
    "num_warps": 4
  },
  "64": {
    "BLOCK_SIZE_K": 64,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 1,
    "NEED_TRANS": false,
    "num_stages": 2,
    "num_warps": 4
  },
  "128": {
    "BLOCK_SIZE_K": 64,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 16,
    "NEED_TRANS": false,
    "num_stages": 3,
    "num_warps": 4
  },
  "256": {
    "BLOCK_SIZE_K": 64,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 1,
    "NEED_TRANS": false,
    "num_stages": 2,
    "num_warps": 4
  },
  "512": {
    "BLOCK_SIZE_K": 64,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 1,
    "NEED_TRANS": false,
    "num_stages": 4,
    "num_warps": 4
  },
  "800": {
    "BLOCK_SIZE_K": 64,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 32,
    "NEED_TRANS": false,
    "num_stages": 2,
    "num_warps": 4
  },
  "1024": {
    "BLOCK_SIZE_K": 64,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 64,
    "NEED_TRANS": false,
    "num_stages": 2,
    "num_warps": 4
  },
  "2048": {
    "BLOCK_SIZE_K": 32,
    "BLOCK_SIZE_M": 32,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 16,
    "NEED_TRANS": false,
    "num_stages": 3,
    "num_warps": 4
  },
  "8192": {
    "BLOCK_SIZE_K": 64,
    "BLOCK_SIZE_M": 64,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 32,
    "NEED_TRANS": false,
    "num_stages": 2,
    "num_warps": 4
  }
}

83 changes: 83 additions & 0 deletions ...ted_weight=false,out_dtype=torch.bfloat16,topk_num=8,use_fp8_w8a8=false}_NVIDIA_H200.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,83 @@
{
"1": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 1,
"NEED_TRANS": false,
"num_stages": 4,
"num_warps": 4
},
"100": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"NEED_TRANS": false,
"num_stages": 3,
"num_warps": 4
},
"1024": {
"BLOCK_SIZE_K": 32,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 64,
"NEED_TRANS": false,
"num_stages": 3,
"num_warps": 4
},
"128": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"NEED_TRANS": false,
"num_stages": 2,
"num_warps": 8
},
"16": {
"BLOCK_SIZE_K": 64,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 1,
"NEED_TRANS": false,
"num_stages": 3,
"num_warps": 4
},
"256": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 16,
"NEED_TRANS": false,
"num_stages": 2,
"num_warps": 4
},
"32": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"GROUP_SIZE_M": 16,
"NEED_TRANS": false,
"num_stages": 3,
"num_warps": 4
},
"64": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"NEED_TRANS": false,
"num_stages": 2,
"num_warps": 4
},
"8": {
"BLOCK_SIZE_K": 128,
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 128,
"GROUP_SIZE_M": 32,
"NEED_TRANS": false,
"num_stages": 3,
"num_warps": 8
}
}
Comment on lines +1 to +83
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 2, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

The keys in this JSON configuration file are not sorted, which makes it difficult to read and maintain. For better readability, it's recommended to sort the keys numerically. This is especially helpful when manually inspecting or debugging the autotuning configurations.

{
  "1": {
    "BLOCK_SIZE_K": 128,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 64,
    "GROUP_SIZE_M": 1,
    "NEED_TRANS": false,
    "num_stages": 4,
    "num_warps": 4
  },
  "8": {
    "BLOCK_SIZE_K": 128,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 32,
    "NEED_TRANS": false,
    "num_stages": 3,
    "num_warps": 8
  },
  "16": {
    "BLOCK_SIZE_K": 64,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 1,
    "NEED_TRANS": false,
    "num_stages": 3,
    "num_warps": 4
  },
  "32": {
    "BLOCK_SIZE_K": 128,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 64,
    "GROUP_SIZE_M": 16,
    "NEED_TRANS": false,
    "num_stages": 3,
    "num_warps": 4
  },
  "64": {
    "BLOCK_SIZE_K": 128,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 32,
    "NEED_TRANS": false,
    "num_stages": 2,
    "num_warps": 4
  },
  "100": {
    "BLOCK_SIZE_K": 128,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 1,
    "NEED_TRANS": false,
    "num_stages": 3,
    "num_warps": 4
  },
  "128": {
    "BLOCK_SIZE_K": 128,
    "BLOCK_SIZE_M": 16,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 32,
    "NEED_TRANS": false,
    "num_stages": 2,
    "num_warps": 8
  },
  "256": {
    "BLOCK_SIZE_K": 128,
    "BLOCK_SIZE_M": 32,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 16,
    "NEED_TRANS": false,
    "num_stages": 2,
    "num_warps": 4
  },
  "1024": {
    "BLOCK_SIZE_K": 32,
    "BLOCK_SIZE_M": 64,
    "BLOCK_SIZE_N": 128,
    "GROUP_SIZE_M": 64,
    "NEED_TRANS": false,
    "num_stages": 3,
    "num_warps": 4
  }
}

38 changes: 38 additions & 0 deletions ..._kernel_configs/triton_3.5.1/NVIDIA_H200/moe_align_fused:v1/{topk_num=8}_NVIDIA_H200.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,38 @@
{
"1": {
"BLOCK_SIZE": 256,
"num_warps": 4
},
"100": {
"BLOCK_SIZE": 128,
"num_warps": 8
},
"1024": {
"BLOCK_SIZE": 256,
"num_warps": 4
},
"128": {
"BLOCK_SIZE": 256,
"num_warps": 8
},
"16": {
"BLOCK_SIZE": 128,
"num_warps": 8
},
"256": {
"BLOCK_SIZE": 128,
"num_warps": 8
},
"32": {
"BLOCK_SIZE": 128,
"num_warps": 8
},
"64": {
"BLOCK_SIZE": 128,
"num_warps": 8
},
"8": {
"BLOCK_SIZE": 128,
"num_warps": 8
}
}
56 changes: 56 additions & 0 deletions .../moe_sum_reduce:v1/{hidden_dim=2048,out_dtype=torch.bfloat16,topk_num=8}_NVIDIA_H200.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,56 @@
{
"1": {
"BLOCK_DIM": 256,
"BLOCK_M": 2,
"NUM_STAGE": 2,
"num_warps": 8
},
"100": {
"BLOCK_DIM": 1024,
"BLOCK_M": 1,
"NUM_STAGE": 1,
"num_warps": 8
},
"1024": {
"BLOCK_DIM": 1024,
"BLOCK_M": 1,
"NUM_STAGE": 4,
"num_warps": 1
},
"128": {
"BLOCK_DIM": 1024,
"BLOCK_M": 1,
"NUM_STAGE": 1,
"num_warps": 16
},
"16": {
"BLOCK_DIM": 128,
"BLOCK_M": 1,
"NUM_STAGE": 1,
"num_warps": 2
},
"256": {
"BLOCK_DIM": 1024,
"BLOCK_M": 1,
"NUM_STAGE": 4,
"num_warps": 2
},
"32": {
"BLOCK_DIM": 128,
"BLOCK_M": 1,
"NUM_STAGE": 4,
"num_warps": 4
},
"64": {
"BLOCK_DIM": 128,
"BLOCK_M": 1,
"NUM_STAGE": 4,
"num_warps": 4
},
"8": {
"BLOCK_DIM": 1024,
"BLOCK_M": 1,
"NUM_STAGE": 1,
"num_warps": 16
}
}
Comment on lines +1 to +56
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 2, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

The keys in this JSON configuration file are not sorted, which makes it difficult to read and maintain. For better readability, it's recommended to sort the keys numerically. This is especially helpful when manually inspecting or debugging the autotuning configurations.

{
  "1": {
    "BLOCK_DIM": 256,
    "BLOCK_M": 2,
    "NUM_STAGE": 2,
    "num_warps": 8
  },
  "8": {
    "BLOCK_DIM": 1024,
    "BLOCK_M": 1,
    "NUM_STAGE": 1,
    "num_warps": 16
  },
  "16": {
    "BLOCK_DIM": 128,
    "BLOCK_M": 1,
    "NUM_STAGE": 1,
    "num_warps": 2
  },
  "32": {
    "BLOCK_DIM": 128,
    "BLOCK_M": 1,
    "NUM_STAGE": 4,
    "num_warps": 4
  },
  "64": {
    "BLOCK_DIM": 128,
    "BLOCK_M": 1,
    "NUM_STAGE": 4,
    "num_warps": 4
  },
  "100": {
    "BLOCK_DIM": 1024,
    "BLOCK_M": 1,
    "NUM_STAGE": 1,
    "num_warps": 8
  },
  "128": {
    "BLOCK_DIM": 1024,
    "BLOCK_M": 1,
    "NUM_STAGE": 1,
    "num_warps": 16
  },
  "256": {
    "BLOCK_DIM": 1024,
    "BLOCK_M": 1,
    "NUM_STAGE": 4,
    "num_warps": 2
  },
  "1024": {
    "BLOCK_DIM": 1024,
    "BLOCK_M": 1,
    "NUM_STAGE": 4,
    "num_warps": 1
  }
}

56 changes: 56 additions & 0 deletions ...n_3.5.1/NVIDIA_H200/silu_and_mul_fwd:v1/{N=192,out_dtype=torch.bfloat16}_NVIDIA_H200.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,56 @@
{
"1024": {
"BLOCK_M": 1,
"BLOCK_N": 256,
"NUM_STAGES": 2,
"num_warps": 4
},
"128": {
"BLOCK_M": 1,
"BLOCK_N": 256,
"NUM_STAGES": 1,
"num_warps": 8
},
"2048": {
"BLOCK_M": 1,
"BLOCK_N": 256,
"NUM_STAGES": 1,
"num_warps": 1
},
"256": {
"BLOCK_M": 1,
"BLOCK_N": 256,
"NUM_STAGES": 1,
"num_warps": 8
},
"512": {
"BLOCK_M": 1,
"BLOCK_N": 128,
"NUM_STAGES": 2,
"num_warps": 4
},
"64": {
"BLOCK_M": 1,
"BLOCK_N": 64,
"NUM_STAGES": 4,
"num_warps": 1
},
"8": {
"BLOCK_M": 1,
"BLOCK_N": 64,
"NUM_STAGES": 4,
"num_warps": 1
},
"800": {
"BLOCK_M": 1,
"BLOCK_N": 256,
"NUM_STAGES": 2,
"num_warps": 1
},
"8192": {
"BLOCK_M": 8,
"BLOCK_N": 256,
"NUM_STAGES": 4,
"num_warps": 1
}
}
Comment on lines +1 to +56
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Feb 2, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

The keys in this JSON configuration file are not sorted, which makes it difficult to read and maintain. For better readability, it's recommended to sort the keys numerically. This is especially helpful when manually inspecting or debugging the autotuning configurations.

{
  "8": {
    "BLOCK_M": 1,
    "BLOCK_N": 64,
    "NUM_STAGES": 4,
    "num_warps": 1
  },
  "64": {
    "BLOCK_M": 1,
    "BLOCK_N": 64,
    "NUM_STAGES": 4,
    "num_warps": 1
  },
  "128": {
    "BLOCK_M": 1,
    "BLOCK_N": 256,
    "NUM_STAGES": 1,
    "num_warps": 8
  },
  "256": {
    "BLOCK_M": 1,
    "BLOCK_N": 256,
    "NUM_STAGES": 1,
    "num_warps": 8
  },
  "512": {
    "BLOCK_M": 1,
    "BLOCK_N": 128,
    "NUM_STAGES": 2,
    "num_warps": 4
  },
  "800": {
    "BLOCK_M": 1,
    "BLOCK_N": 256,
    "NUM_STAGES": 2,
    "num_warps": 1
  },
  "1024": {
    "BLOCK_M": 1,
    "BLOCK_N": 256,
    "NUM_STAGES": 2,
    "num_warps": 4
  },
  "2048": {
    "BLOCK_M": 1,
    "BLOCK_N": 256,
    "NUM_STAGES": 1,
    "num_warps": 1
  },
  "8192": {
    "BLOCK_M": 8,
    "BLOCK_N": 256,
    "NUM_STAGES": 4,
    "num_warps": 1
  }
}

Loading