In offload mode, GQA expansion buffers (_k_expanded, _v_expanded) are not
needed since compute_chunked_prefill() handles GQA inline. Previously,
these buffers were always allocated based on max_model_len, causing OOM
on 24GB GPUs (e.g., RTX 3090) when max_model_len=1M (16GB buffer).
Changes:
- Add enable_cpu_offload parameter to alloc_policy_metadata() in base class
- Skip GQA buffer allocation when enable_cpu_offload=True in XAttentionBSAPolicy
- Pass enable_cpu_offload from model_runner to policy
Memory savings: ~16GB for 1M seq, ~1.1GB for 72K seq
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Document OOM issue when using XAttention BSA + CPU offload
with large models (GLM-4-9B) on 24GB GPUs.
Issue: 8GB allocation for k_expanded buffer fails due to
using num_heads instead of num_kv_heads in GQA models.
Root cause analysis and proposed fix included.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Document the difference between compute density (BSA block level)
and communication density (CPU block level).
Key finding: Even with 37% compute density, comm density can be 100%
due to any() aggregation across heads/Q-positions spreading sparse
blocks across all CPU blocks.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Document the CUDA stream synchronization bug in XAttention BSA
- Include root cause analysis with stream timing diagrams
- Add test commands and verification results (100% accuracy)
- Update CLAUDE.md documentation index
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Rename doc to "Density Alignment Test Results" (covers both modes)
- Add Offload mode test results (3.7K-64.9K tokens, all passed)
- Add Layer 5 GPU-only test results (threshold=0.9, density=6.24%)
- Enhance test script to support both GPU-only and Offload data formats
- Add batch testing commands for all data files
- Update CLAUDE.md index
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
Implement three-phase KV chunking for sparse attention estimation:
1. softmax_compute_partial_stats: compute (m, l) per KV chunk
2. merge_softmax_stats: merge partial stats on host
3. softmax_normalize_and_block_sum: normalize with global stats
This allows computing sparse attention masks without storing full
raw attention scores in GPU memory, reducing peak memory usage
from O(q_len * k_full_len) to O(q_len * k_chunk_len).
Key changes:
- Add softmax_partial_stats_kernel with causal mask support
- Add softmax_normalize_block_sum_kernel with kv_offset parameter
- Add Python wrappers for new kernels
- Update test script to validate KV chunking alignment
- Add documentation for the new kernels
Test results show perfect alignment with xattn_estimate API:
- Density difference: 0.000000
- Mask difference: 0.0044%
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
Document the verification test for XAttention Triton kernel KV chunking:
- 32K and 64K test results with threshold 0.9/0.95/1.0
- Key finding: threshold=1.0 achieves alignment (~0% diff)
- threshold<1.0 shows 10-13% difference due to per-chunk threshold application
- Conclusion: softmax normalization is correct, issue is threshold accumulation
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
- Add DensityObserver class to track per-layer density statistics
- Integrate DensityObserver into compute_prefill for GPU-only mode
- Fix stride parameter not being passed to xattn_estimate
- Add density statistics output to test_ruler.py for XATTN_BSA
- Add comprehensive density benchmark documentation
Key changes:
- nanovllm/utils/density_observer.py: New Observer for density tracking
- xattn_bsa.py: Add stride param to xattn_estimate, integrate DensityObserver
- test_ruler.py: Enable DensityObserver and print summary for XATTN_BSA
- docs/xattn_density_benchmark.md: Benchmark results for 4K-32K contexts
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Summarizes lessons learned from GLM-4 integration:
- Config field mapping (multi_query_group_num, kv_channels, etc.)
- RoPE variants (interleaved vs half, partial vs full rotation)
- EOS token handling for multi-EOS models
- Weight name conversion patterns
- Verification checklist
Also updates CLAUDE.md to reflect GLM-4 support.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Use FlashInfer's optimized merge_state kernel for attention output merging
in chunked prefill. End-to-end improvement: +0.8% (32K) to +2.4% (64K).
Key changes:
- Add merge_attention_outputs_flashinfer() with LSE format conversion
- FlashInfer uses log2, flash_attn uses ln: convert via LOG2_E/LN_2
- Keep original Triton kernel for fallback
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add comprehensive list of 1M+ context models from Hugging Face
- Categorize by type: text-only LLM vs vision-language models
- Separate ≤10B (practical) from >10B (resource-intensive) models
- Include Qwen, GLM, InternLM, Llama, MiniMax, Gradient AI series
- Add VRAM requirements and technical comparison table
- Update CLAUDE.md documentation index
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Document the performance impact of block_size on softmax_fuse_block_sum:
- Current 4096 (reshaped 512) is the WORST point: 95ms
- Optimal 1024 (reshaped 128): 6ms - 15x faster
- Performance follows U-shaped curve
Add tests/bench_estimate_block_size.py for benchmarking and propose
hierarchical block sum approach for optimization.
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
Add load_k_only_to_slot_layer() to OffloadEngine for estimate phase:
- Only load K (not K+V) during block selection in select_blocks()
- Reduces H2D transfer by 50% in estimate phase
- 64K context: XAttn/Full ratio drops from 1.48x to 0.99x
- 32K context: XAttn/Full ratio drops from 1.67x to 1.20x
The estimate phase uses flat_group_gemm_fuse_reshape(Q, K) which
only requires K for attention score computation. V is unused.
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
Add comprehensive performance analysis for XAttention:
- NVTX marker locations and usage
- Block size impact on offload mode (4096 vs 1024)
- Detailed timing breakdown for estimate vs compute phases
- softmax_fuse_block_sum_kernel analysis
- Optimization recommendations
Key findings:
- block_size=4096 is 2x faster than 1024 for 64K context
- find_blocks_chunked is bottleneck (40%) at block_size=4096
- estimate_gemm becomes bottleneck (24%) at block_size=1024
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
Merge GPU-only sparse attention support from tzj/minference-exp branch:
**GPU-only mode additions:**
- Add compute_prefill/compute_decode methods to SparsePolicy base class
- Add GPU-only attention routing in attention.py
- Add alloc_policy_metadata() for pre-allocating GQA buffers
- Add XAttention + BSA sparse attention for GPU-only prefill
- Add kvcache_manager to set_context() for policy access
**bench.py enhancements:**
- Add --model argument for configurable model path
- Add --policy argument (full, xattn) for sparse policy selection
- Add --enable-policy flag for FullAttentionPolicy routing
- Add --enforce-eager option to disable CUDA graphs
- Add --gpu-util option for GPU memory utilization
**Documentation:**
- Add gpu_only_xattn_guide.md with performance analysis
- Add gpu_only_sparse_integration.md baseline document
- Add gpu-vram-requirement.md rule for GPU-only mode
Both CPU offload and GPU-only paths are preserved and functional.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Document baseline performance before integrating sparse attention
to GPU-only mode:
- GPU-only Full Attention: 4869 tok/s (32K prefill)
- CPU Offload Full Attention: 1500 tok/s (3.2x slower)
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
- Add --model parameter (default: Llama-3.1-8B-Instruct)
- Add --enable-xattn flag for XAttention BSA sparse prefill
- Add --xattn-threshold and --xattn-stride parameters
- Change default num-gpu-blocks from 6 to 4
- Add benchmark results doc with Full vs XAttn comparison (32K/128K)
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
- Document kernel gap analysis showing 77-81% CPU scheduling overhead
- Identify GPU utilization at 12.8% with potential to reach 39.5%
- Outline optimization directions: CUDA Graph, Triton fusion, C++ extension
- Add documentation index entry in CLAUDE.md
Generated with [Claude Code](https://claude.ai/code)
via [Happy](https://happy.engineering)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Happy <yesreply@happy.engineering>
- Document ring buffer pipeline triggering nsys timestamp bug
- Update profile_offload.sh to use test_ruler.py with options
- Add reference to new doc in CLAUDE.md
Root cause: 4-slot ring buffer pipeline (4 transfer streams +
1 compute stream) triggers event ordering bug in nsys < 2024.2
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add new sections to xattn_bsa_policy_design.md:
- Performance benchmarks: 128K context comparison (Full vs XAttn BSA)
- Density trend analysis across chunks
- Memory leak issue and fix (64GB -> 4GB reduction)
- Memory monitoring guide with gpu-monitor agent
- Density statistics API documentation
- Known issues and optimization directions
Update CLAUDE.md description to reflect new content.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Document flat_group_gemm_fuse_reshape and softmax_fuse_block_sum kernels
- Explain anti-diagonal sum principle and stride sampling
- Add GPU-specific BLOCK_M/N constraints (RTX 3090 vs A100)
- Show Q/K can have different lengths (chunked prefill support)
- Update CLAUDE.md with doc reference
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add xattn_estimate_chunked function ported from COMPASS
- Support chunked prefill with q_start_pos parameter
- Ensure 100% consistency with standard xattn_estimate when
using matching chunk_size parameter
- Add test and documentation
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Document CUDA Graph memory behavior based on actual testing:
- Memory overhead at each stage (model, cache, warmup, capture, replay)
- StaticCache is the main overhead (~144MB for 1K tokens)
- Graph capture adds minimal overhead (~8MB)
- Graph replay requires zero additional allocation
- Performance improvement: ~2.8x decode throughput
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Root Cause:
- OffloadEngine.reset() cleared GPU buffers but NOT CPU cache
- Previous request's KV cache data persisted in CPU memory, contaminating subsequent requests
Fixes:
- Add k_cache_cpu.zero_() and v_cache_cpu.zero_() to OffloadEngine.reset()
- Add clear_decode_tracking(seq) call in HybridKVCacheManager.deallocate()
Results:
- niah_single_1 accuracy improved from ~80% to 94% (+14%)
- Remaining ~6% errors are model limitations, not state leakage
Also:
- Update docs/ruler_32k_chunked_offload_issue.md with fix details
- Remove debug planning files (findings.md, progress.md, task_plan.md)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add comprehensive documentation analyzing the 32K chunked offload
accuracy issues with proposed solutions covering LSE precision,
ring buffer state management, and position encoding validation.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add docs/block_sparse_attn_interface.md with BSA function signatures
- Update CLAUDE.md documentation index
- Remove obsolete DEBUG_SUMMARY.md and test_report_sparse_policy_refactor.md
- Add notes.md to .gitignore
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Create docs/sparse_policy_implementation_guide.md with comprehensive guide
- Rewrite .claude/rules/sparse-policy.md with mandatory base class requirements
- Add new doc reference to CLAUDE.md documentation index
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Document accuracy degradation issue in 32K context with chunked offload
- Add detailed hypothesis analysis and debugging approach
- Include 4-slot ring buffer experiment results
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Remove cross-layer pipeline from OffloadEngine (saves ~1GB GPU memory for long sequences)
- Delete layer_k/v_buffer_a/b double buffers
- Remove start_decode_pipeline, get_decode_layer_kv, end_decode_pipeline methods
- Remove pipeline state tracking variables
- Simplify decode to use ring buffer pipeline only (more efficient for long sequences)
- Rename compute_chunked_attention → compute_chunked_prefill for clarity
- Add mandatory needle test requirements: --enable-offload --input-len 32768
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add comprehensive documentation for the SparsePolicy abstraction:
- SparsePolicy base class and abstract methods
- FullAttentionPolicy prefill/decode flow
- Ring buffer and cross-layer pipeline modes
- Code conventions and testing guidelines
Update CLAUDE.md documentation index with reference.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
