feat: kvpress runner 侧按 num_kv_tokens 计算 KV 写入位置

vllm/v1/worker/gpu_model_runner.py

@@ -146,6 +146,13 @@ class GPUModelRunnerBase(LoRAModelRunnerMixin):
         self.attention_chunk_size = model_config.attention_chunk_size
 
         self.cascade_attn_enabled = not self.model_config.disable_cascade_attn
+        if envs.VLLM_ENABLE_KV_COMPRESSION:
+            # KV compression changes the effective KV sequence layout and
+            # invalidates cascade attention assumptions (common-prefix blocks).
+            self.cascade_attn_enabled = False
+        # Whether the current step needs KV compaction work (score/topk/dst).
+        # This is set per-step in `_prepare_inputs`.
+        self.kv_compression_needs_compaction: bool = False
 
         # Multi-modal data support
         self.mm_registry = MULTIMODAL_REGISTRY
@@ -673,6 +680,17 @@ class GPUModelRunnerBase(LoRAModelRunnerMixin):
         np.add(self.input_batch.num_computed_tokens_cpu[req_indices],
                arange,
                out=positions_np)
+        # KV positions (where the KV for each scheduled token is temporarily
+        # written). When KV compression is enabled, KV positions are decoupled
+        # from logical positions.
+        use_kv_compression = envs.VLLM_ENABLE_KV_COMPRESSION
+        if use_kv_compression:
+            kv_positions_np = self.kv_positions_np[:total_num_scheduled_tokens]
+            np.add(self.input_batch.num_kv_tokens_cpu[req_indices],
+                   arange,
+                   out=kv_positions_np)
+        else:
+            kv_positions_np = None
 
         # Calculate M-RoPE positions.
         # Only relevant for models using M-RoPE (e.g, Qwen2-VL)
@@ -700,6 +718,8 @@ class GPUModelRunnerBase(LoRAModelRunnerMixin):
             block_size = kv_cache_group_spec.kv_cache_spec.block_size
             block_table: BlockTable = self.input_batch.block_table[
                 kv_cache_group_id]
+            slot_positions_np = (kv_positions_np
+                                 if use_kv_compression else positions_np)
             # E.g., [0, 1, 0, 1, 2, 3, 4, 0, 1, 2]
             # -> [0, 0, K, K, K + 1, K + 1, K + 2, 2 * K, 2 * K, 2 * K + 1]
             # where K is the max_num_blocks_per_req and the block size is 2.
@@ -708,11 +728,11 @@ class GPUModelRunnerBase(LoRAModelRunnerMixin):
             # block_size.
             block_table_indices = (
                 req_indices * block_table.max_num_blocks_per_req +
-                positions_np // block_size)
+                slot_positions_np // block_size)
             block_table_cpu = block_table.get_cpu_tensor()
             block_numbers = block_table_cpu.flatten(
             )[block_table_indices].numpy()
-            block_offsets = positions_np % block_size
+            block_offsets = slot_positions_np % block_size
             np.add(
                 block_numbers * block_size,
                 block_offsets,
@@ -722,9 +742,14 @@ class GPUModelRunnerBase(LoRAModelRunnerMixin):
         self.query_start_loc_np[0] = 0
         self.query_start_loc_np[1:num_reqs + 1] = cu_num_tokens
 
-        self.seq_lens_np[:num_reqs] = (
-            self.input_batch.num_computed_tokens_cpu[:num_reqs] +
-            num_scheduled_tokens)
+        if use_kv_compression:
+            self.seq_lens_np[:num_reqs] = (
+                self.input_batch.num_kv_tokens_cpu[:num_reqs] +
+                num_scheduled_tokens)
+        else:
+            self.seq_lens_np[:num_reqs] = (
+                self.input_batch.num_computed_tokens_cpu[:num_reqs] +
+                num_scheduled_tokens)
 
         # Copy the tensors to the GPU.
         self.input_ids[:total_num_scheduled_tokens].copy_(
@@ -2547,6 +2572,10 @@ class GPUModelRunnerBase(LoRAModelRunnerMixin):
         assert len(self.attn_backends) == 0 and len(
             self.attn_metadata_builders
         ) == 0, "Attention backends are already initialized"
+        if envs.VLLM_ENABLE_KV_COMPRESSION and self.full_cuda_graph:
+            raise ValueError(
+                "KV compression is currently incompatible with full CUDA "
+                "graph mode.")
         for i, kv_cache_group_spec in enumerate(
                 kv_cache_config.kv_cache_groups):
             kv_cache_spec = kv_cache_group_spec.kv_cache_spec
@@ -2570,7 +2599,16 @@ class GPUModelRunnerBase(LoRAModelRunnerMixin):
                     raise NotImplementedError(
                         "Non-Attention backend is not supported by V1 "
                         "GPUModelRunner.")
+                if (envs.VLLM_ENABLE_KV_COMPRESSION
+                        and attn_backend_i.get_name() != "FLASH_ATTN_VLLM_V1"):
+                    raise ValueError(
+                        "KV compression currently requires "
+                        "VLLM_ATTENTION_BACKEND=FLASH_ATTN_VLLM_V1.")
             elif isinstance(kv_cache_spec, MambaSpec):
+                if envs.VLLM_ENABLE_KV_COMPRESSION:
+                    raise ValueError(
+                        "KV compression is currently only supported for "
+                        "Transformer attention layers.")
                 attn_backend_i = Mamba2AttentionBackend
             else:
                 raise ValueError(