[Bug] --sampler-rng cuda breaks SDXL on vulkan

[daniandtheweb]

Git commit

f532972

Operating System & Version

"Arch"

GGML backends

Vulkan

Command-line arguments used

./sd -M img_gen -p "a cat" --sampling-method euler_a --steps 20 --scheduler gits -W 1024 -H 1024 -b 1 --cfg-scale 5 -s -1 --clip-skip -1 --embd-dir /home/daniandtheweb/Workspace/sd.cpp-webui/models/embeddings/ --lora-model-dir /home/daniandtheweb/Workspace/sd.cpp-webui/models/loras/ -t 0 --rng cuda --sampler-rng cuda --lora-apply-mode auto -o /home/daniandtheweb/Workspace/sd.cpp-webui/outputs/txt2img/1763304506.png --model /home/daniandtheweb/Workspace/sd.cpp-webui/models/checkpoints/plantMilkModelSuite_hempII.safetensors --vae /home/daniandtheweb/Workspace/sd.cpp-webui/models/vae/sdxl_vae_fp16_fix.safetensors --preview proj --preview-path /home/daniandtheweb/Workspace/sd.cpp-webui/outputs/txt2img/1763304506_preview.png --preview-interval 1 --diffusion-fa --vae-conv-direct --color

Steps to reproduce

Running the inference with --sampler-rng cuda doesn't correctly denoise the image.

What you expected to happen

A clear output image.

What actually happened

Complete noise.

Logs / error messages / stack trace

Additional context / environment details

I've tested this issue on SDXL, on SD 1.5 this I can't seem to trigger the same issue. If instead of cuda I use std_default the denoising happens normally.

[leejet]

When rng and sampler-rng use the same type but different RNG instances, this issue is expected. #982 prevents that situation.

[daniandtheweb]

Thanks for the quick fix.