2026-08-06 06:44:31 +00:00
|
|
|
|
#!/bin/bash
|
2026-08-05 08:24:38 +00:00
|
|
|
|
# BI-V100 engine patches for Qwen3.6-35B-A3B (Qwen3_5 architecture)
|
2026-07-30 16:06:20 +00:00
|
|
|
|
#
|
2026-08-05 08:24:38 +00:00
|
|
|
|
# All modifications are FULL FILE REPLACEMENTS — no AST patch scripts.
|
|
|
|
|
|
# Each file was read in full from the base image vllm source, modified
|
|
|
|
|
|
# with the necessary fixes, and placed here as a complete copy.
|
2026-07-30 16:06:20 +00:00
|
|
|
|
#
|
2026-08-05 08:24:38 +00:00
|
|
|
|
# Base image: git.modelhub.org.cn:9443/enginex-iluvatar/bi100-3.2.3-x86-ubuntu20.04-py3.10-poc-llm-infer:v1.2.3
|
|
|
|
|
|
# vllm install path: /usr/local/corex/lib/python3/dist-packages/vllm/
|
|
|
|
|
|
|
|
|
|
|
|
VLLM=/usr/local/corex/lib/python3/dist-packages/vllm
|
|
|
|
|
|
VLLM64=/usr/local/corex/lib64/python3/dist-packages/vllm
|
|
|
|
|
|
|
|
|
|
|
|
# Detect which lib path exists
|
|
|
|
|
|
if [ -d "$VLLM" ]; then
|
|
|
|
|
|
V=$VLLM
|
|
|
|
|
|
elif [ -d "$VLLM64" ]; then
|
|
|
|
|
|
V=$VLLM64
|
|
|
|
|
|
else
|
|
|
|
|
|
echo "[patch_ops] ERROR: vllm not found at lib or lib64 path"
|
|
|
|
|
|
exit 1
|
|
|
|
|
|
fi
|
|
|
|
|
|
|
|
|
|
|
|
echo "[patch_ops] vllm path: $V"
|
|
|
|
|
|
|
2026-08-05 08:38:36 +00:00
|
|
|
|
# --- _custom_ops.py: SMEM 48KB fix + hardware ops bindings -------------------
|
|
|
|
|
|
# Base image returns 32KB (32768) for get_max_shared_memory_per_block, but
|
|
|
|
|
|
# BI-V100 actually has 48KB (49152) confirmed via ixsmi. This limits Triton
|
|
|
|
|
|
# tile sizes and ixformer internal allocations if not corrected.
|
|
|
|
|
|
# CCCL GridEvenShare test (catch2_test_grid_even_share.cu) validates that
|
|
|
|
|
|
# work distribution depends on correct hardware parameters — wrong SMEM
|
|
|
|
|
|
# means wrong tile_size means wrong grid_size.
|
|
|
|
|
|
# FULL FILE REPLACEMENT.
|
|
|
|
|
|
cp ./_custom_ops.py $V/_custom_ops.py
|
|
|
|
|
|
echo "[patch_ops] _custom_ops.py → / (SMEM 32KB→48KB fix)"
|
|
|
|
|
|
|
2026-08-05 08:24:38 +00:00
|
|
|
|
# --- paged_attn.py: pure-PyTorch attention fallback --------------------------
|
|
|
|
|
|
# Bypasses Triton context_attention_fwd (hangs BI-V100 permanently).
|
|
|
|
|
|
# Uses K-tiling Flash Attention online softmax for prefix attention.
|
|
|
|
|
|
# Uses pure-PyTorch decode for seq_len > 32K.
|
|
|
|
|
|
# CCCL-ported: adaptive tile sizing from dispatch_reduce.cuh GridEvenShare.
|
|
|
|
|
|
cp ./paged_attn.py $V/attention/ops/paged_attn.py
|
|
|
|
|
|
echo "[patch_ops] paged_attn.py → attention/ops/"
|
2026-07-30 16:06:20 +00:00
|
|
|
|
|
2026-08-06 07:01:14 +00:00
|
|
|
|
# --- prefix_prefill.py: Triton-free prefix attention -------------------------
|
|
|
|
|
|
# On BI-V100, Triton is not installed. This file provides the context_attention_fwd
|
|
|
|
|
|
# function that paged_attn.py imports. Even though our paged_attn.py comments out
|
|
|
|
|
|
# the Triton import, the base xformers.py may still try to import it.
|
|
|
|
|
|
# Deploy it so the import doesn't crash — the function itself won't be called.
|
|
|
|
|
|
cp ./prefix_prefill.py $V/attention/ops/prefix_prefill.py
|
|
|
|
|
|
echo "[patch_ops] prefix_prefill.py → attention/ops/"
|
|
|
|
|
|
|
2026-08-05 08:24:38 +00:00
|
|
|
|
# --- model_runner.py: prefix_cache_hit fix -----------------------------------
|
|
|
|
|
|
# Bug: Case 1 (prefix_cache_len <= context_len) leaves prefix_cache_hit=True,
|
|
|
|
|
|
# causing undersized block_tables in chunked prefill chunk 2+.
|
|
|
|
|
|
# Fix: set prefix_cache_hit=False for Case 1.
|
|
|
|
|
|
# FULL FILE REPLACEMENT — no patch_model_runner.py script.
|
|
|
|
|
|
cp ./model_runner.py $V/worker/model_runner.py
|
|
|
|
|
|
echo "[patch_ops] model_runner.py → worker/"
|
2026-07-30 16:06:20 +00:00
|
|
|
|
|
2026-08-05 08:24:38 +00:00
|
|
|
|
# --- xformers.py: head_dim>128 fallback + Q-tiling --------------------------
|
|
|
|
|
|
# Injects _run_sdpa_fallback (pure matmul+softmax) for head_dim=256.
|
|
|
|
|
|
# ixformer flash attention crashes (is_causal=True) or gives wrong output.
|
|
|
|
|
|
# Also disables auto chunked-prefill (Q-tiling handles long context).
|
|
|
|
|
|
# FULL FILE REPLACEMENT — no patch_xformers_sdpa_seq.py script.
|
|
|
|
|
|
cp ./xformers.py $V/attention/backends/xformers.py
|
|
|
|
|
|
echo "[patch_ops] xformers.py → attention/backends/"
|
|
|
|
|
|
|
|
|
|
|
|
# --- arg_utils.py: disable auto chunked-prefill for 32K+ --------------------
|
|
|
|
|
|
# Q-tiling in _run_sdpa_fallback handles long-context memory.
|
|
|
|
|
|
# FULL FILE REPLACEMENT.
|
|
|
|
|
|
cp ./arg_utils.py $V/engine/arg_utils.py
|
|
|
|
|
|
echo "[patch_ops] arg_utils.py → engine/"
|
|
|
|
|
|
|
|
|
|
|
|
# --- logits_processor.py: seq_groups=None guard ------------------------------
|
|
|
|
|
|
# Prevents crash when seq_groups is None during intermediate chunked-prefill.
|
|
|
|
|
|
# FULL FILE REPLACEMENT.
|
|
|
|
|
|
cp ./logits_processor.py $V/model_executor/layers/logits_processor.py
|
|
|
|
|
|
echo "[patch_ops] logits_processor.py → model_executor/layers/"
|
2026-07-30 16:06:20 +00:00
|
|
|
|
|
2026-08-05 08:30:21 +00:00
|
|
|
|
# --- sampler.py: CCCL-ported top-k fast path for sampling --------------------
|
|
|
|
|
|
# When all sequences use top_p=1.0, skip full sort+cumsum and use torch.topk.
|
|
|
|
|
|
# CCCL partition/flagged.cu insight: radix select is O(N×bits_per_pass) vs
|
|
|
|
|
|
# full sort O(N log N). For vocab=152064: topk ~11 passes vs sort ~17 passes.
|
|
|
|
|
|
# FULL FILE REPLACEMENT.
|
|
|
|
|
|
cp ./sampler.py $V/model_executor/layers/sampler.py
|
|
|
|
|
|
echo "[patch_ops] sampler.py → model_executor/layers/"
|
|
|
|
|
|
|
2026-07-30 16:06:20 +00:00
|
|
|
|
# --- transformers: Qwen3_5 tokenizer / model files --------------------------
|
2026-08-05 08:36:45 +00:00
|
|
|
|
# NOTE: patch_transformers_qwen3_5.py is the ONLY remaining patch script.
|
|
|
|
|
|
# It modifies pip-installed transformers' configuration_auto.py and __init__.py
|
|
|
|
|
|
# to register qwen3_5/qwen3_5_moe. These files come from pip (version-specific)
|
|
|
|
|
|
# so we can't pre-copy them — the patch script inserts lines after known anchors.
|
2026-08-06 06:35:37 +00:00
|
|
|
|
pip install transformers==4.55.3 -i https://pypi.tuna.tsinghua.edu.cn/simple 2>/dev/null || \
|
|
|
|
|
|
pip install transformers==4.55.3 2>/dev/null || \
|
|
|
|
|
|
echo "[patch_ops] WARNING: pip install transformers failed, using pre-installed version"
|
2026-07-30 16:06:20 +00:00
|
|
|
|
cp -r ./qwen3_5 /usr/local/lib/python3.10/site-packages/transformers/models/
|
|
|
|
|
|
cp -r ./qwen3_5_moe /usr/local/lib/python3.10/site-packages/transformers/models/
|
|
|
|
|
|
python3 ./patch_transformers_qwen3_5.py
|
2026-08-05 08:24:38 +00:00
|
|
|
|
echo "[patch_ops] transformers Qwen3_5 models installed"
|
2026-07-30 16:06:20 +00:00
|
|
|
|
|
2026-08-05 08:24:38 +00:00
|
|
|
|
# --- vllm model: Qwen3.6 (Qwen3_5 arch) ------------------------------------
|
2026-08-05 08:36:45 +00:00
|
|
|
|
# FULL FILE REPLACEMENT of registry.py with Qwen3_5 entries pre-added.
|
|
|
|
|
|
# No more patch_vllm_qwen3_5.py script.
|
2026-08-05 08:24:38 +00:00
|
|
|
|
cp ./mamba_cache.py $V/model_executor/models/
|
|
|
|
|
|
cp ./qwen3_5.py $V/model_executor/models/qwen3_5.py
|
2026-08-05 08:36:45 +00:00
|
|
|
|
cp ./registry.py $V/model_executor/models/registry.py
|
|
|
|
|
|
echo "[patch_ops] qwen3_5.py + registry.py deployed"
|
2026-07-30 16:06:20 +00:00
|
|
|
|
|
2026-08-06 07:01:14 +00:00
|
|
|
|
# --- paged_attention_v2_pytorch.py: PyTorch V2 attention fallback ------------
|
|
|
|
|
|
# _custom_ops.py imports this from the parent of its own directory.
|
|
|
|
|
|
# In docker, vllm lives at $V/, so we place it one level up AND next to _custom_ops.
|
|
|
|
|
|
# Belt-and-suspenders: also copy to /workspace/ where _custom_ops.py's _repo_root points.
|
|
|
|
|
|
cp ./paged_attention_v2_pytorch.py $V/paged_attention_v2_pytorch.py
|
|
|
|
|
|
cp ./paged_attention_v2_pytorch.py /workspace/paged_attention_v2_pytorch.py
|
|
|
|
|
|
echo "[patch_ops] paged_attention_v2_pytorch.py → $V/ + /workspace/"
|
|
|
|
|
|
|
2026-08-05 08:24:38 +00:00
|
|
|
|
# --- sequence.py: fix completion_tokens inflation ----------------------------
|
|
|
|
|
|
cp ./sequence.py $V/sequence.py
|
|
|
|
|
|
echo "[patch_ops] sequence.py → /"
|
2026-07-30 16:06:20 +00:00
|
|
|
|
|
2026-08-05 08:24:38 +00:00
|
|
|
|
# --- scheduler.py: record num_cached_tokens ---------------------------------
|
|
|
|
|
|
cp ./scheduler.py $V/core/scheduler.py
|
|
|
|
|
|
echo "[patch_ops] scheduler.py → core/"
|
2026-07-30 16:06:20 +00:00
|
|
|
|
|
2026-08-05 08:24:38 +00:00
|
|
|
|
# --- tool parser: Qwen3 XML tool call format --------------------------------
|
2026-08-05 08:36:45 +00:00
|
|
|
|
# FULL FILE REPLACEMENT of __init__.py with Qwen3CoderToolParser pre-added.
|
|
|
|
|
|
# No more patch_vllm_tool_parser.py script.
|
2026-08-05 08:24:38 +00:00
|
|
|
|
cp ./qwen3coder_tool_parser.py $V/entrypoints/openai/tool_parsers/
|
2026-08-05 08:36:45 +00:00
|
|
|
|
cp ./tool_parsers_init.py $V/entrypoints/openai/tool_parsers/__init__.py
|
|
|
|
|
|
echo "[patch_ops] qwen3_coder tool parser deployed"
|
2026-08-05 08:24:38 +00:00
|
|
|
|
|
|
|
|
|
|
# --- reasoning parser: Qwen3 <think>...</think> split -----------------------
|
|
|
|
|
|
cp -r ./reasoning $V/
|
|
|
|
|
|
cp ./protocol.py $V/entrypoints/openai/protocol.py
|
|
|
|
|
|
cp ./cli_args.py $V/entrypoints/openai/cli_args.py
|
|
|
|
|
|
cp ./serving_chat.py $V/entrypoints/openai/serving_chat.py
|
|
|
|
|
|
cp ./api_server.py $V/entrypoints/openai/api_server.py
|
|
|
|
|
|
cp ./chat_utils.py $V/entrypoints/chat_utils.py
|
|
|
|
|
|
echo "[patch_ops] reasoning parser + serving files installed"
|
2026-07-30 16:06:20 +00:00
|
|
|
|
|
2026-08-05 08:24:38 +00:00
|
|
|
|
echo "[patch_ops] DONE — all patches applied via full file replacement"
|