[claudesquad] update from 'perf_opt-1' on 07 Jan 26 05:58 CST

CLAUDE.md

@@ -46,6 +46,17 @@ python bench_offload.py
 
 ## Local Package Installation for Multi-Instance
 
+**CRITICAL**: After ANY code modification in the `nanovllm/` directory, you MUST reinstall the package before running tests or benchmarks:
+
+```bash
+pip install -e . --prefix=./.local --no-deps
+```
+
+Then run with PYTHONPATH:
+```bash
+PYTHONPATH=./.local/lib/python3.10/site-packages:$PYTHONPATH python <script.py>
+```
+
 **IMPORTANT**: When running multiple Claude instances on different worktrees, do NOT use `pip install -e .` globally as it will affect other instances. Instead, use local installation:
 
 1. **Install to worktree-local directory**:

nanovllm/engine/model_runner.py

@@ -590,14 +590,15 @@ class ModelRunner:
 
     def run_chunked_offload_decode(self, seqs: list[Sequence]) -> list[int]:
         """
-        Run decode with ring buffer (CPU is primary storage).
+        Run decode with cross-layer pipeline (CPU is primary storage).
 
         All KV is on CPU. Uses decode_slot (slot[0]) to write new KV.
-        Other slots (slots[1:]) are used to load previous KV chunks via pipeline.
+        Optimized with cross-layer pipeline: Layer N's data is loaded while
-        New token's KV is written to decode_slot then offloaded to CPU only when block is full.
+        Layer N-1 computes, achieving transfer/compute overlap.
 
         Key: decode_slot is dedicated to writing new KV, never used for loading.
-        Optimization: Batch offloads - only offload when block is full, attend to all accumulated tokens.
+        Optimization: Cross-layer pipeline reduces effective latency by overlapping
+        H2D transfers with attention computation across layers.
         """
         assert len(seqs) == 1, "Ring buffer decode only supports single sequence"
         seq = seqs[0]
@@ -618,6 +619,12 @@ class ModelRunner:
         # Get decode start position for accumulated token tracking
         decode_start_pos = self.kvcache_manager.get_decode_start_pos(seq)
 
+        # Get prefilled CPU blocks for pipeline initialization
+        cpu_block_table = self.kvcache_manager.get_prefilled_cpu_blocks(seq)
+
+        # Start cross-layer pipeline (preloads Layer 0's data)
+        offload_engine.start_decode_pipeline(cpu_block_table)
+
         # Set up context for chunked decode
         set_context(
             is_prefill=False,
@@ -634,6 +641,9 @@ class ModelRunner:
         logits = self.run_model(input_ids, positions, is_prefill=False)
         reset_context()
 
+        # End cross-layer pipeline
+        offload_engine.end_decode_pipeline()
+
         # Only offload when block is full (pos_in_block == block_size - 1)
         # This avoids unnecessary offloading on every decode step
         if pos_in_block == self.block_size - 1:

nanovllm/kvcache/offload_engine.py

@@ -142,6 +142,40 @@ class OffloadEngine:
         decode_buf_mb = 2 * num_layers * block_size * num_kv_heads * head_dim * dtype.itemsize / (1024 * 1024)
         logger.info(f"  Per-layer decode buffer: {decode_buf_mb:.1f} MB")
 
+        # ========== Cross-layer pipeline buffers for decode ==========
+        # Double-buffered layer cache for pipelined decode:
+        # - Buffer A: Current layer's prefilled KV being computed
+        # - Buffer B: Next layer's prefilled KV being loaded
+        # Shape: [max_prefill_blocks, block_size, kv_heads, head_dim]
+        # Memory: 2 * max_prefill_blocks * block_size * kv_heads * head_dim * dtype_size
+        max_prefill_blocks = num_cpu_blocks  # Can hold all prefill blocks
+        self.layer_k_buffer_a = torch.zeros(
+            max_prefill_blocks, block_size, num_kv_heads, head_dim,
+            dtype=dtype, device="cuda"
+        )
+        self.layer_v_buffer_a = torch.zeros(
+            max_prefill_blocks, block_size, num_kv_heads, head_dim,
+            dtype=dtype, device="cuda"
+        )
+        self.layer_k_buffer_b = torch.zeros(
+            max_prefill_blocks, block_size, num_kv_heads, head_dim,
+            dtype=dtype, device="cuda"
+        )
+        self.layer_v_buffer_b = torch.zeros(
+            max_prefill_blocks, block_size, num_kv_heads, head_dim,
+            dtype=dtype, device="cuda"
+        )
+        layer_buf_mb = 4 * max_prefill_blocks * block_size * num_kv_heads * head_dim * dtype.itemsize / (1024 * 1024)
+        logger.info(f"  Cross-layer pipeline buffers: {layer_buf_mb:.1f} MB ({max_prefill_blocks} blocks × 2)")
+
+        # Pipeline state tracking
+        self._pipeline_active = False
+        self._pipeline_current_buffer = 0  # 0 = buffer A, 1 = buffer B
+        self._pipeline_next_layer_event = torch.cuda.Event()
+        self._pipeline_cpu_blocks: list = []  # CPU block IDs to load
+        self._pipeline_num_blocks = 0
+        self._pipeline_layer_stream = torch.cuda.Stream()  # Dedicated stream for layer loading
+
         # ========== Fixed-address CPU KV cache (pinned memory) ==========
         self.k_cache_cpu = torch.zeros(
             num_layers, num_cpu_blocks, block_size, num_kv_heads, head_dim,
@@ -1064,3 +1098,119 @@ class OffloadEngine:
                 if e.__class__.__name__ == 'BdbQuit':
                     raise
                 logger.warning(f"Debug hook error: {e}")
+
+    # ========== Cross-layer Pipeline Methods for Decode ==========
+
+    def start_decode_pipeline(self, cpu_block_ids: List[int]) -> None:
+        """
+        Start cross-layer pipeline for decode.
+
+        Called at the beginning of a decode step to initialize the pipeline.
+        Preloads Layer 0's data into buffer A.
+
+        Args:
+            cpu_block_ids: List of CPU block IDs for prefilled blocks
+        """
+        if not cpu_block_ids:
+            self._pipeline_active = False
+            return
+
+        self._pipeline_active = True
+        self._pipeline_cpu_blocks = cpu_block_ids
+        self._pipeline_num_blocks = len(cpu_block_ids)
+        self._pipeline_current_buffer = 0
+
+        # Preload Layer 0 into buffer A
+        self._load_layer_to_buffer(0, 0)  # layer_id=0, buffer_idx=0 (A)
+
+    def get_decode_layer_kv(self, layer_id: int, num_blocks: int) -> Tuple[Tensor, Tensor]:
+        """
+        Get KV cache for a layer during decode.
+
+        If pipeline is active, returns data from the current buffer.
+        Also triggers preloading of the next layer (if not last layer).
+
+        Args:
+            layer_id: Current layer ID
+            num_blocks: Number of blocks to return
+
+        Returns:
+            (k_cache, v_cache) tensors, shape: [num_blocks, block_size, kv_heads, head_dim]
+        """
+        if not self._pipeline_active:
+            raise RuntimeError("Decode pipeline not active. Call start_decode_pipeline first.")
+
+        # Wait for current layer's data to be ready
+        self.compute_stream.wait_event(self._pipeline_next_layer_event)
+
+        # Get current buffer
+        if self._pipeline_current_buffer == 0:
+            k = self.layer_k_buffer_a[:num_blocks]
+            v = self.layer_v_buffer_a[:num_blocks]
+        else:
+            k = self.layer_k_buffer_b[:num_blocks]
+            v = self.layer_v_buffer_b[:num_blocks]
+
+        # Trigger preloading of next layer (if not last layer)
+        next_layer_id = layer_id + 1
+        if next_layer_id < self.num_layers:
+            # Use the other buffer for next layer
+            next_buffer_idx = 1 - self._pipeline_current_buffer
+            self._load_layer_to_buffer(next_layer_id, next_buffer_idx)
+            # Switch to next buffer for next layer
+            self._pipeline_current_buffer = next_buffer_idx
+
+        return k, v
+
+    def _load_layer_to_buffer(self, layer_id: int, buffer_idx: int) -> None:
+        """
+        Async load a layer's prefilled blocks to the specified buffer.
+
+        Uses sgDMA for efficient strided transfer from CPU cache.
+
+        Args:
+            layer_id: Layer index to load
+            buffer_idx: 0 for buffer A, 1 for buffer B
+        """
+        num_blocks = self._pipeline_num_blocks
+        cpu_block_ids = self._pipeline_cpu_blocks
+
+        # Select target buffer
+        if buffer_idx == 0:
+            k_buffer = self.layer_k_buffer_a
+            v_buffer = self.layer_v_buffer_a
+        else:
+            k_buffer = self.layer_k_buffer_b
+            v_buffer = self.layer_v_buffer_b
+
+        # Load all blocks for this layer using dedicated stream
+        with torch.cuda.stream(self._pipeline_layer_stream):
+            for i, cpu_block_id in enumerate(cpu_block_ids):
+                # Copy from CPU cache (has layer dimension) to GPU buffer
+                k_buffer[i].copy_(
+                    self.k_cache_cpu[layer_id, cpu_block_id],
+                    non_blocking=True
+                )
+                v_buffer[i].copy_(
+                    self.v_cache_cpu[layer_id, cpu_block_id],
+                    non_blocking=True
+                )
+            # Record event when all transfers complete
+            self._pipeline_next_layer_event.record(self._pipeline_layer_stream)
+
+    def end_decode_pipeline(self) -> None:
+        """
+        End the cross-layer pipeline.
+
+        Called at the end of a decode step to clean up pipeline state.
+        """
+        if self._pipeline_active:
+            # Ensure all transfers complete before ending
+            self._pipeline_layer_stream.synchronize()
+            self._pipeline_active = False
+            self._pipeline_cpu_blocks = []
+            self._pipeline_num_blocks = 0
+
+    def is_pipeline_active(self) -> bool:
+        """Check if decode pipeline is currently active."""
+        return self._pipeline_active

nanovllm/layers/attention.py

@@ -479,17 +479,15 @@ class Attention(nn.Module):
         context,
     ) -> torch.Tensor:
         """
-        Compute decode attention using ring buffer pipeline (same as prefill).
+        Compute decode attention using cross-layer pipeline.
 
-        Uses the same loading mechanism as _chunked_prefill_attention:
+        Optimization: Uses double-buffered layer cache to overlap H2D transfer
-        - Load one block at a time from CPU to GPU slot
+        with computation across layers:
-        - Compute attention for each block
+        - Layer N computes while Layer N+1's data is being loaded
-        - Merge results using online softmax
+        - Each layer only waits for its own data, not all layers' data
-        - Finally merge with decode buffer (accumulated decode tokens)
 
-        This approach is simpler and proven correct (prefill tests pass).
+        This reduces effective latency from O(num_layers * transfer_time) to
-        The only difference from prefill is the additional decode buffer
+        O(transfer_time + num_layers * compute_time) when transfer < compute.
-        that stores new tokens generated during decode.
         """
         from nanovllm.kvcache.chunked_attention import flash_attn_with_lse, merge_attention_outputs
 
@@ -533,9 +531,16 @@ class Attention(nn.Module):
             )
 
         offload_engine = kvcache_manager.offload_engine
-        load_slots = offload_engine.decode_load_slots  # Available slots for loading
 
-        # Use ring buffer pipeline (same as prefill) to load prefilled blocks
+        # Use cross-layer pipeline if active (initialized in model_runner)
+        if offload_engine.is_pipeline_active():
+            o_acc, lse_acc = self._decode_with_layer_pipeline(
+                q_batched, cpu_block_table, offload_engine,
+                block_size, last_block_valid_tokens
+            )
+        else:
+            # Fallback to original ring buffer pipeline
+            load_slots = offload_engine.decode_load_slots
             o_acc, lse_acc = self._decode_ring_buffer_pipeline(
                 q_batched, cpu_block_table, load_slots, offload_engine,
                 block_size, last_block_valid_tokens
@@ -652,3 +657,62 @@ class Attention(nn.Module):
                     o_acc, lse_acc = merge_attention_outputs(o_acc, lse_acc, prev_o, prev_lse)
 
         return o_acc, lse_acc
+
+    def _decode_with_layer_pipeline(
+        self,
+        q_batched: torch.Tensor,
+        cpu_block_table: list,
+        offload_engine,
+        block_size: int,
+        last_block_valid_tokens: int,
+    ):
+        """
+        Decode using cross-layer pipeline for optimized H2D transfer.
+
+        This method uses pre-loaded layer buffers instead of loading
+        blocks one by one. The pipeline loads the next layer's data
+        while the current layer computes, achieving transfer/compute overlap.
+
+        The key insight is that each layer needs the SAME blocks but from
+        different layers of CPU cache. By double-buffering and pipelining
+        across layers, we reduce total latency.
+        """
+        from nanovllm.kvcache.chunked_attention import flash_attn_with_lse, merge_attention_outputs
+
+        num_blocks = len(cpu_block_table)
+        if num_blocks == 0:
+            return None, None
+
+        compute_stream = offload_engine.compute_stream
+
+        # Get KV from pre-loaded layer buffer (triggers next layer loading)
+        prev_k, prev_v = offload_engine.get_decode_layer_kv(self.layer_id, num_blocks)
+
+        # prev_k, prev_v shape: [num_blocks, block_size, kv_heads, head_dim]
+        # Reshape to [1, num_blocks * block_size, kv_heads, head_dim]
+        total_tokens = num_blocks * block_size
+
+        # Handle partial last block
+        if last_block_valid_tokens < block_size:
+            # Only use valid tokens from last block
+            actual_tokens = (num_blocks - 1) * block_size + last_block_valid_tokens
+            # Flatten and truncate
+            prev_k_flat = prev_k.reshape(-1, prev_k.shape[-2], prev_k.shape[-1])[:actual_tokens]
+            prev_v_flat = prev_v.reshape(-1, prev_v.shape[-2], prev_v.shape[-1])[:actual_tokens]
+        else:
+            prev_k_flat = prev_k.reshape(-1, prev_k.shape[-2], prev_k.shape[-1])
+            prev_v_flat = prev_v.reshape(-1, prev_v.shape[-2], prev_v.shape[-1])
+
+        # Add batch dimension: [1, total_tokens, kv_heads, head_dim]
+        prev_k_batched = prev_k_flat.unsqueeze(0)
+        prev_v_batched = prev_v_flat.unsqueeze(0)
+
+        # Compute attention on all prefilled blocks at once
+        with torch.cuda.stream(compute_stream):
+            o_acc, lse_acc = flash_attn_with_lse(
+                q_batched, prev_k_batched, prev_v_batched,
+                softmax_scale=self.scale,
+                causal=False,
+            )
+
+        return o_acc, lse_acc