Sage Attention: out of resource

I'm on ComfyUI and I'm trying out the basic z-image workflow. Everything works, except that this error keeps popping up during image generation:

Error running sage attention: out of resource: shared memory, Required: 67584, Hardware limit: 65536. Reducing block sizes or num_stages may help., using pytorch attention instead

I know my graphics card is small (RTX 2060 Laptop, 6 GB VRAM, 64 GB RAM), but other models, like FLUX, run perfectly with Sage Attention even at 4K resolutions; my workflow is shown in the photo. How could I speed up image generation with z-image? Even using quantized (and therefore smaller) models or changing resolutions doesn't change the situation