P104-100显卡性能测试

前言

我有两块p104-100，一个是8针电源的msi，另一个是6针电源猜测是耕升吧。我之前看到有评论说，8针的更耗电，性能高一些，有些好奇，在同样的平台中测试了对比

功耗

其实两个卡，在计算绘图时，整机功耗大致都在240-260w，不过msi的似乎更倾向于靠近260w，数值相对稳定

性能

耕升：sdxl，首次217秒，后续126秒

MSI：sdxl，首次200秒，后续126秒

耕升数据

Total VRAM 8109 MB, total RAM 15446 MB
pytorch version: 2.4.1+cu121
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA P104-100 : cudaMallocAsync

Requested to load SDXLClipModel
Loading 1 new model
loaded completely 0.0 1560.802734375 True
Requested to load SDXL
Loading 1 new model
loaded completely 0.0 4897.0483474731445 True
100%|███████████████████████████████████████████| 20/20 [01:19<00:00,  3.98s/it]
Requested to load SDXLRefiner
Loading 1 new model
loaded completely 0.0 4309.7050857543945 True
100%|█████████████████████████████████████████████| 5/5 [00:16<00:00,  3.25s/it]
Requested to load AutoencoderKL
Loading 1 new model
loaded completely 0.0 319.11416244506836 True
Prompt executed in 217.91 seconds
Requested to load SDXL
Loading 1 new model
loaded completely 0.0 4897.0483474731445 True
100%|███████████████████████████████████████████| 20/20 [01:18<00:00,  3.94s/it]
Requested to load SDXLRefiner
Loading 1 new model
loaded completely 0.0 4309.7050857543945 True
100%|█████████████████████████████████████████████| 5/5 [00:16<00:00,  3.25s/it]
Requested to load AutoencoderKL
Loading 1 new model
loaded completely 0.0 319.11416244506836 True
Prompt executed in 126.11 seconds

msi数据

Total VRAM 8109 MB, total RAM 15446 MB
pytorch version: 2.4.1+cu121
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA P104-100 : cudaMallocAsync

Requested to load SDXLClipModel
Loading 1 new model
loaded completely 0.0 1560.802734375 True
Requested to load SDXL
Loading 1 new model
loaded completely 0.0 4897.0483474731445 True
100%|███████████████████████████████████████████| 20/20 [01:19<00:00,  3.98s/it]
Requested to load SDXLRefiner
Loading 1 new model
loaded completely 0.0 4309.7050857543945 True
100%|█████████████████████████████████████████████| 5/5 [00:16<00:00,  3.26s/it]
Requested to load AutoencoderKL
Loading 1 new model
loaded completely 0.0 319.11416244506836 True
Prompt executed in 199.99 seconds
Requested to load SDXL
Loading 1 new model
loaded completely 0.0 4897.0483474731445 True
100%|███████████████████████████████████████████| 20/20 [01:19<00:00,  3.95s/it]
Requested to load SDXLRefiner
Loading 1 new model
loaded completely 0.0 4309.7050857543945 True
100%|█████████████████████████████████████████████| 5/5 [00:16<00:00,  3.28s/it]
Requested to load AutoencoderKL
Loading 1 new model
loaded completely 0.0 319.11416244506836 True
Prompt executed in 125.97 seconds

喜欢 (0)赏

[赞赏使我快乐]

分享 (0)