[Bug]: Increasing batch size slows it down.[Performance]

Checklist

[X] The issue exists after disabling all extensions
[X] The issue exists on a clean installation of webui
[ ] The issue is caused by an extension, but I believe it is caused by a bug in the webui
[ ] The issue exists in the current version of the webui
[X] The issue has not been reported before recently
[ ] The issue has been reported before but has not been fixed yet

What happened?

Increasing the batch count during generation will slow it down compared to sd.webui. The difference in generation time was over 1 hour at 512x768 count:100 size:8.

Steps to reproduce the problem

input any prompt
set batchsize to any
Click [Generate].

What should have happened?

Generate faster than sd.webui

What browsers do you use to access the UI ?

Microsoft Edge

Sysinfo

sysinfo-2024-03-11-21-16.json

Console logs

Python 3.10.6 (tags/v3.10.6:9c7b4bd, Aug  1 2022, 21:53:49) [MSC v.1932 64 bit (AMD64)]
Version: f0.0.17v1.8.0rc-latest-276-g29be1da7
Commit hash: 29be1da7cf2b5dccfc70fbdd33eb35c56a31ffb7
Launching Web UI with arguments:
Total VRAM 8192 MB, total RAM 32658 MB
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce GTX 1070 : native
VAE dtype: torch.float32
CUDA Stream Activated:  False
Using pytorch cross attention
ControlNet preprocessor location: G:\webui_forge_cu121_torch21\webui\models\ControlNetPreprocessor
Loading weights [15012c538f] from G:\webui_forge_cu121_torch21\webui\models\Stable-diffusion\realisticVisionV51_v51VAE.safetensors
2024-03-12 06:02:34,453 - ControlNet - INFO - ControlNet UI callback registered.
Running on local URL:  http://127.0.0.1:7860

To create a public link, set `share=True` in `launch()`.
Startup time: 71.0s (initial startup: 0.1s, prepare environment: 25.2s, import torch: 15.8s, import gradio: 7.2s, setup paths: 8.9s, initialize shared: 0.7s, other imports: 5.5s, setup gfpgan: 0.1s, load scripts: 3.8s, create ui: 3.2s, gradio launch: 1.6s).
model_type EPS
UNet ADM Dimension 0
Using pytorch attention in VAE
Working with z of shape (1, 4, 32, 32) = 4096 dimensions.
Using pytorch attention in VAE
extra {'cond_stage_model.clip_l.logit_scale', 'cond_stage_model.clip_l.text_projection'}
To load target model SD1ClipModel
Begin to load 1 model
[Memory Management] Current Free GPU Memory (MB) =  7216.9990234375
[Memory Management] Model Memory (MB) =  454.2076225280762
[Memory Management] Minimal Inference Memory (MB) =  1024.0
[Memory Management] Estimated Remaining GPU Memory (MB) =  5738.791400909424
Moving model(s) has taken 0.17 seconds
Model loaded in 124.9s (load weights from disk: 4.2s, forge instantiate config: 0.7s, forge load real models: 117.2s, load textual inversion embeddings: 0.1s, calculate empty prompt: 2.7s).
To load target model BaseModel
Begin to load 1 model
[Memory Management] Current Free GPU Memory (MB) =  7187.494140625
[Memory Management] Model Memory (MB) =  3278.8199005126953
[Memory Management] Minimal Inference Memory (MB) =  1024.0
[Memory Management] Estimated Remaining GPU Memory (MB) =  2884.6742401123047
Moving model(s) has taken 0.67 seconds
100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [01:04<00:00,  3.20s/it]
To load target model AutoencoderKL████████████████████                                 | 20/40 [00:58<01:01,  3.08s/it]
Begin to load 1 model
[Memory Management] Current Free GPU Memory (MB) =  3872.1396484375
[Memory Management] Model Memory (MB) =  319.11416244506836
[Memory Management] Minimal Inference Memory (MB) =  1024.0
[Memory Management] Estimated Remaining GPU Memory (MB) =  2529.0254859924316
Moving model(s) has taken 0.23 seconds
100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [01:01<00:00,  3.09s/it]
To load target model AutoencoderKL█████████████████████████████████████████████████████| 40/40 [02:04<00:00,  3.10s/it]
Begin to load 1 model
[Memory Management] Current Free GPU Memory (MB) =  3871.63134765625
[Memory Management] Model Memory (MB) =  319.11416244506836
[Memory Management] Minimal Inference Memory (MB) =  1024.0
[Memory Management] Estimated Remaining GPU Memory (MB) =  2528.5171852111816
Moving model(s) has taken 0.07 seconds
Total progress: 100%|██████████████████████████████████████████████████████████████████| 40/40 [02:11<00:00,  3.30s/it]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 40/40 [02:11<00:00,  3.10s/it]

Additional information

input prompt:test, cat, floor negative prompt: img res: 512x512

simple test:(Batch count/size:1)

sd.webui v1.8.0 100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [00:09<00:00, 2.04it/s] sd.webui forge vf0.0.17v1.8.0rc-latest-276-g29be1da7 100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [00:08<00:00, 2.37it/s]

sd.webui v1.8.0 Console full logs

Python 3.10.6 (tags/v3.10.6:9c7b4bd, Aug 1 2022, 21:53:49) [MSC v.1932 64 bit (AMD64)] Version: v1.8.0 Commit hash: bef51aed032c0aaa5cfd80445bc4cf0d85b408b5 Launching Web UI with arguments: no module 'xformers'. Processing without... no module 'xformers'. Processing without... No module 'xformers'. Proceeding without it. Loading weights [15012c538f] from G:\sd.webui\webui\models\Stable-diffusion\realisticVisionV51_v51VAE.safetensors Creating model from config: G:\sd.webui\webui\configs\v1-inference.yaml Running on local URL: http://127.0.0.1:7860

To create a public link, set share=True in launch(). Startup time: 9.6s (prepare environment: 2.1s, import torch: 3.4s, import gradio: 1.0s, setup paths: 0.9s, initialize shared: 0.2s, other imports: 0.5s, load scripts: 0.8s, create ui: 0.4s, gradio launch: 0.4s). Applying attention optimization: Doggettx... done. Model loaded in 2.5s (load weights from disk: 0.5s, create model: 0.7s, apply weights to model: 1.1s, calculate empty prompt: 0.1s). 100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [00:57<00:00, 2.87s/it] 100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [00:57<00:00, 2.86s/it] Total progress: 100%|██████████████████████████████████████████████████████████████████| 40/40 [02:00<00:00, 3.02s/it] Total progress: 100%|██████████████████████████████████████████████████████████████████| 40/40 [02:00<00:00, 2.86s/it]

lllyasviel / stable-diffusion-webui-forge