IntelGpu-ComfyUI-Collection

202610311328_QwenEdit2511_omnixpu加速_B70_64G_冷启+热启.txt · 2026-10-31 13:28

← 返回执行日志
冷启179.64 s
热启19.51 s
热启 vs 冷启9.21×
每步耗时1.47 s/it
峰值 staged19,582 MB

逐次执行

阶段耗时步数s/itstaged MB其他指标
冷启179.64 s61.4719,582CUTE ×3
热启19.51 s61.4719,582—
原始日志全文(已脱敏本地路径)
[INFO] got prompt
[INFO] Model storage policy: fast_disk=False paths=['<ComfyUI>/models/vae/qwen_image_vae.safetensors']
[INFO] Using pytorch attention in VAE
[INFO] Using pytorch attention in VAE
[INFO] VAE load device: xpu:0, offload device: cpu, dtype: torch.float32
[INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=5921.6MiB
[INFO] Requested to load WanVAE
[INFO] Model WanVAE prepared for dynamic VRAM loading. 483MB Staged. 0 patches attached. Force pre-loaded 60 weights: 89 KB.
[INFO] Found quantization metadata version 1
[INFO] Using MixedPrecisionOps for text encoder
[INFO] Model storage policy: fast_disk=False paths=['<本地路径>/qwen_2.5_vl_7b_int8_convrot.safetensors']
[INFO] CLIP/text encoder model load device: xpu:0, offload device: cpu, current: cpu, dtype: torch.float16
[INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=6016.1MiB
[INFO] Model WanVAE prepared for dynamic VRAM loading. 483MB Staged. 0 patches attached. Force pre-loaded 60 weights: 89 KB.
[INFO] Requested to load QwenImageTEModel_
[INFO] Model QwenImageTEModel_ prepared for dynamic VRAM loading. 8193MB Staged. 0 patches attached. Force pre-loaded 122 weights: 561 KB.
[INFO] [OmniXPU] norm first use: RMSNorm shape=torch.Size([768, 1280])
[INFO] Found quantization metadata version 1
[INFO] Detected mixed precision quantization
[INFO] Using mixed precision operations
[INFO] Native ops: int8_tensorwise , emulated ops: mxfp8, w6a8_int8, float8_e5m2, float8_e4m3fn, convrot_w4a4, asym_w4a8_int8, nvfp4
[INFO] model weight dtype torch.bfloat16, manual cast: torch.bfloat16
[INFO] model_type FLUX
[INFO] Model storage policy: fast_disk=False paths=['<本地路径>/qwen_image_edit_2511_int8_convrot.safetensors']
[INFO] [OmniXPU] LoRA memory budget cached: source=810.7MiB/2160 tensors model_added=810.0MiB model_budget=810.0MiB clip_added=0.0MiB clip_budget=0.0MiB
[INFO] [OmniXPU] LoRA memory budget cached: source=225.0MiB/1440 tensors model_added=225.0MiB model_budget=1035.0MiB clip_added=0.0MiB clip_budget=0.0MiB
[INFO] [OmniXPU] LoRA model budget applied: cached=1035.0MiB/2880 tensors requested=2347.1->3382.1MiB minimum=1173.5->2208.5MiB xpu_memory=available xpu_allocated=0.6MiB xpu_reserved=2.0MiB xpu_free=21914.5MiB xpu_total=31977.4MiB
[INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=2208.5MiB
[INFO] Requested to load QwenImage
[INFO] Model QwenImage prepared for dynamic VRAM loading. 19582MB Staged. 840 patches attached. Force pre-loaded 241 weights: 70 KB.
[INFO] [OmniXPU] LoRA model load complete: xpu_memory=available xpu_allocated=0.8MiB xpu_reserved=2.0MiB xpu_free=21914.3MiB xpu_total=31977.4MiB
  0%|                                                                | 0/6 [00:00<?, ?it/s,   Model Initializing ...  ][INFO] [OmniXPU] attention CUTE #1: heads=24 seq=8434 dtype=torch.bfloat16
[INFO] [OmniXPU] attention CUTE #2: heads=24 seq=8434 dtype=torch.bfloat16
[INFO] [OmniXPU] attention CUTE #3: heads=24 seq=8434 dtype=torch.bfloat16
100%|████████████████████████████████████████████████████████████████████████████████████| 6/6 [00:08<00:00,  1.47s/it]
[INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=550.0MiB
[INFO] Model WanVAE prepared for dynamic VRAM loading. 483MB Staged. 0 patches attached. Force pre-loaded 60 weights: 89 KB.
[INFO] Prompt executed in 179.64 seconds
[INFO] got prompt
[INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=5921.6MiB
[INFO] Model WanVAE prepared for dynamic VRAM loading. 483MB Staged. 0 patches attached. Force pre-loaded 60 weights: 89 KB.
[INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=6016.1MiB
[INFO] Model WanVAE prepared for dynamic VRAM loading. 483MB Staged. 0 patches attached. Force pre-loaded 60 weights: 89 KB.
[INFO] Model QwenImageTEModel_ prepared for dynamic VRAM loading. 8193MB Staged. 0 patches attached. Force pre-loaded 122 weights: 561 KB.
[INFO] [OmniXPU] LoRA model budget applied: cached=1035.0MiB/2880 tensors requested=2347.1->3382.1MiB minimum=1173.5->2208.5MiB xpu_memory=available xpu_allocated=0.8MiB xpu_reserved=2.0MiB xpu_free=8034.0MiB xpu_total=31977.4MiB
[INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=2208.5MiB
[INFO] Model QwenImage prepared for dynamic VRAM loading. 19582MB Staged. 840 patches attached. Force pre-loaded 241 weights: 70 KB.
[INFO] [OmniXPU] LoRA model load complete: xpu_memory=available xpu_allocated=0.8MiB xpu_reserved=2.0MiB xpu_free=8034.2MiB xpu_total=31977.4MiB
100%|████████████████████████████████████████████████████████████████████████████████████| 6/6 [00:08<00:00,  1.47s/it]
[INFO] [OmniXPU] AIMDO inference budget: devices=0 minimum=550.0MiB
[INFO] Model WanVAE prepared for dynamic VRAM loading. 483MB Staged. 0 patches attached. Force pre-loaded 60 weights: 89 KB.
[INFO] Prompt executed in 19.51 seconds