add 4090D config

WANDY666 · WANDY666 · commit fa1b888aad68 · 2025-11-26T05:06:51.000Z
diff --git a/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=16,out_dtype=torch.bfloat16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=16,out_dtype=torch.bfloat16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1 @@
+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 1}, "32": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 1}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}}}
diff --git a/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=16,out_dtype=torch.float16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=16,out_dtype=torch.float16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1 @@
+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 1}, "32": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 1}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}}}
diff --git a/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=2,out_dtype=torch.bfloat16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=2,out_dtype=torch.bfloat16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1 @@
+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 5}, "32": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 4}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 4}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 2}, "32": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 9}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}}}
diff --git a/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=2,out_dtype=torch.float16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=2,out_dtype=torch.float16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1 @@
+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 5}, "32": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 4}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 4}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 2}, "32": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 9}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}}}
diff --git a/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=4,out_dtype=torch.bfloat16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=4,out_dtype=torch.bfloat16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1 @@
+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 11}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 5}, "32": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 2}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}}}
diff --git a/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=4,out_dtype=torch.float16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=4,out_dtype=torch.float16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1 @@
+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 11}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 5}, "32": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 2}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}}}
diff --git a/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=5,out_dtype=torch.bfloat16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=5,out_dtype=torch.bfloat16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1 @@
+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 1}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 4}}}
diff --git a/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=5,out_dtype=torch.float16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=5,out_dtype=torch.float16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1 @@
+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 1}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 4}}}
diff --git a/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=8,out_dtype=torch.bfloat16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=8,out_dtype=torch.bfloat16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1 @@
+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 10}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 4}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 5}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}}}
diff --git a/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=8,out_dtype=torch.float16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/all_kernel_configs/_fwd_kernel_flash_decode_diverse_stage1:v1/{block_seq=256,gqa_group_size=8,out_dtype=torch.float16,q_head_dim=128}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1 @@
+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 10}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 4}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 5}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 3}}}
diff --git a/lightllm/common/triton_utils/autotune_kernel_configs/triton_3.5.1/NVIDIA_GeForce_RTX_4090_D/fp8_scaled_mm_per_token:v3/{K=14336,N=5120,out_dtype=torch.bfloat16}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/triton_utils/autotune_kernel_configs/triton_3.5.1/NVIDIA_GeForce_RTX_4090_D/fp8_scaled_mm_per_token:v3/{K=14336,N=5120,out_dtype=torch.bfloat16}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1,82 @@
+{
+  "1": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 8,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 5,
+    "num_warps": 8
+  },
+  "100": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 64,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 4
+  },
+  "1024": {
+    "BLOCK_K": 64,
+    "BLOCK_M": 64,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 4
+  },
+  "128": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 64,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 8
+  },
+  "16": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 16,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 5,
+    "num_warps": 4
+  },
+  "2048": {
+    "BLOCK_K": 128,
+    "BLOCK_M": 64,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 4
+  },
+  "256": {
+    "BLOCK_K": 128,
+    "BLOCK_M": 64,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 8
+  },
+  "32": {
+    "BLOCK_K": 128,
+    "BLOCK_M": 32,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 6,
+    "num_warps": 4
+  },
+  "64": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 64,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 4
+  },
+  "8": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 8,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 6,
+    "num_warps": 4
+  }
+}
diff --git a/lightllm/common/triton_utils/autotune_kernel_configs/triton_3.5.1/NVIDIA_GeForce_RTX_4090_D/fp8_scaled_mm_per_token:v3/{K=4096,N=5120,out_dtype=torch.bfloat16}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/triton_utils/autotune_kernel_configs/triton_3.5.1/NVIDIA_GeForce_RTX_4090_D/fp8_scaled_mm_per_token:v3/{K=4096,N=5120,out_dtype=torch.bfloat16}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1,82 @@
+{
+  "1": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 8,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 5,
+    "num_warps": 8
+  },
+  "100": {
+    "BLOCK_K": 64,
+    "BLOCK_M": 32,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 8
+  },
+  "1024": {
+    "BLOCK_K": 64,
+    "BLOCK_M": 64,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 2,
+    "num_warps": 2
+  },
+  "128": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 64,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 8
+  },
+  "16": {
+    "BLOCK_K": 128,
+    "BLOCK_M": 16,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 4
+  },
+  "2048": {
+    "BLOCK_K": 64,
+    "BLOCK_M": 64,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 4
+  },
+  "256": {
+    "BLOCK_K": 128,
+    "BLOCK_M": 64,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 8
+  },
+  "32": {
+    "BLOCK_K": 128,
+    "BLOCK_M": 32,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 5,
+    "num_warps": 4
+  },
+  "64": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 64,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 8
+  },
+  "8": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 8,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 5,
+    "num_warps": 8
+  }
+}
diff --git a/lightllm/common/triton_utils/autotune_kernel_configs/triton_3.5.1/NVIDIA_GeForce_RTX_4090_D/fp8_scaled_mm_per_token:v3/{K=5120,N=2048,out_dtype=torch.bfloat16}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/triton_utils/autotune_kernel_configs/triton_3.5.1/NVIDIA_GeForce_RTX_4090_D/fp8_scaled_mm_per_token:v3/{K=5120,N=2048,out_dtype=torch.bfloat16}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1,82 @@
+{
+  "1": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 8,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 8
+  },
+  "100": {
+    "BLOCK_K": 128,
+    "BLOCK_M": 16,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 6,
+    "num_warps": 8
+  },
+  "1024": {
+    "BLOCK_K": 64,
+    "BLOCK_M": 64,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 8
+  },
+  "128": {
+    "BLOCK_K": 64,
+    "BLOCK_M": 32,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 5,
+    "num_warps": 2
+  },
+  "16": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 8,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 5,
+    "num_warps": 8
+  },
+  "2048": {
+    "BLOCK_K": 64,
+    "BLOCK_M": 64,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 4
+  },
+  "256": {
+    "BLOCK_K": 64,
+    "BLOCK_M": 32,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 4
+  },
+  "32": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 16,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 8
+  },
+  "64": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 32,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 8
+  },
+  "8": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 8,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 5,
+    "num_warps": 8
+  }
+}
diff --git a/lightllm/common/triton_utils/autotune_kernel_configs/triton_3.5.1/NVIDIA_GeForce_RTX_4090_D/fp8_scaled_mm_per_token:v3/{K=5120,N=28672,out_dtype=torch.bfloat16}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/triton_utils/autotune_kernel_configs/triton_3.5.1/NVIDIA_GeForce_RTX_4090_D/fp8_scaled_mm_per_token:v3/{K=5120,N=28672,out_dtype=torch.bfloat16}_NVIDIA_GeForce_RTX_4090_D.json
@@ -0,0 +1,82 @@
+{
+  "1": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 8,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 4
+  },
+  "100": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 64,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 8
+  },
+  "1024": {
+    "BLOCK_K": 64,
+    "BLOCK_M": 64,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 4
+  },
+  "128": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 64,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 8
+  },
+  "16": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 16,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 4
+  },
+  "2048": {
+    "BLOCK_K": 64,
+    "BLOCK_M": 64,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 4
+  },
+  "256": {
+    "BLOCK_K": 128,
+    "BLOCK_M": 64,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 3,
+    "num_warps": 8
+  },
+  "32": {
+    "BLOCK_K": 128,
+    "BLOCK_M": 16,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 5,
+    "num_warps": 4
+  },
+  "64": {
+    "BLOCK_K": 128,
+    "BLOCK_M": 16,
+    "BLOCK_N": 128,
+    "GROUP_M": 8,
+    "num_stages": 5,
+    "num_warps": 4
+  },
+  "8": {
+    "BLOCK_K": 256,
+    "BLOCK_M": 16,
+    "BLOCK_N": 64,
+    "GROUP_M": 8,
+    "num_stages": 4,
+    "num_warps": 8
+  }
+}
diff --git a/lightllm/common/triton_utils/autotune_kernel_configs/triton_3.5.1/NVIDIA_GeForce_RTX_4090_D/fp8_scaled_mm_per_token:v3/{K=5120,N=4096,out_dtype=torch.bfloat16}_NVIDIA_GeForce_RTX_4090_D.json b/lightllm/common/triton_utils/autotune_kernel_configs/triton_3.5.1/NVIDIA_GeForce_RTX_4090_D/fp8_scaled_mm_per_token:v3/{K=5120,N=4096,out_dtype=torch.bfloat16}_NVIDIA_GeForce_RTX_4090_D.json

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1 @@`
	`1`	`+{"4096": {"8": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 1}, "32": {"BLOCK_N": 16, "num_warps": 4, "num_stages": 1}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}}, "8192": {"8": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "32": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "128": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}, "256": {"BLOCK_N": 16, "num_warps": 2, "num_stages": 1}}}`