PR #15832 metal : make the backend async

metal : make the backend async #15832

ggerganov wants to merge 9 commits into master from gg/metal-async

github-actions added ggml

github-actions added Apple Metal

ggerganov force pushed from 2b8074d9 to 84ae8368 2 days ago

ggerganov force pushed from 84ae8368 to c4fe8b0f 2 days ago

ggerganov marked this pull request as ready for review 2 days ago

ggerganov commented on 2025-09-08

metal : make the backend async

97b96c1a

cont : add comments, extend op offload, clean up

c5637cf3

metal : fix batch size for MUL_MAT_ID

bdff7729

metal : remove deprecated ggml_backend_metal_buffer_from_ptr

d91ba85d

ggerganov force pushed from 1526218e to d91ba85d 1 day ago

ggerganov force pushed from f369bdb7 to 8d4835d3 1 day ago

ggerganov commented on 2025-09-09

metal : create only metal buffers, no wrapping of host memory

85aaf52b

ggerganov force pushed from 8d4835d3 to 85aaf52b 1 day ago

ggerganov commented on 2025-09-09

metal : restore .alloc_buffer for buffer_from_ptr_type

7fc2b3d5

metal : remove broken implementation of GGML_OP_SET

f288225d

metal : clean-up loose ends, ready for tests

0926cb49

metal : back to a single queue per device

3f62ee8b

slaren commented on 2025-09-09

ggerganov closed this 7 hours ago

Reviewers

slaren

Assignees

No one assigned

Labels

ggml Apple Metal

Milestone

No milestone