mirror of
https://git.datalinker.icu/comfyanonymous/ComfyUI
synced 2025-12-10 06:24:26 +08:00
e42682b24e
* flux: Do the xq and xk ropes one at a time This was doing independendent interleaved tensor math on the q and k tensors, leading to the holding of more than the minimum intermediates in VRAM. On a bad day, it would VRAM OOM on xk intermediates. Do everything q and then everything k, so torch can garbage collect all of qs intermediates before k allocates its intermediates. This reduces peak VRAM usage for some WAN2.2 inferences (at least). * wan: Optimize qkv intermediates on attention As commented. The former logic computed independent pieces of QKV in parallel which help more inference intermediates in VRAM spiking VRAM usage. Fully roping Q and garbage collecting the intermediates before touching K reduces the peak inference VRAM usage.