llama.cpp
GDN: use replay instead of rollback
#31669
Open

GDN: use replay instead of rollback #31669

am17an wants to merge 11 commits into master from aman/gdn-rs-replay
am17an
Green-Sky
am17an
ggerganov ggerganov assigned ggerganov ggerganov 10 days ago
am17an am17an assigned ServeurpersoCom ServeurpersoCom 10 days ago
am17an
github-actions github-actions added model
github-actions github-actions added testing
github-actions github-actions added Vulkan
github-actions github-actions added ggml
github-actions github-actions added SYCL
github-actions github-actions added Apple Metal
github-actions github-actions added OpenCL
github-actions github-actions added Hexagon
github-actions github-actions added CUDA
github-actions github-actions added OpenVINO
github-actions github-actions added WebGPU
ServeurpersoCom
am17an am17an force pushed from d18aa713 to f46cacff 10 days ago
ServeurpersoCom
ServeurpersoCom
am17an
am17an
ServeurpersoCom
am17an
ServeurpersoCom
am17an
ServeurpersoCom
ServeurpersoCom
pwilkin
ServeurpersoCom
0cc4m
0cc4m commented on 2026-11-10
ggerganov
ggerganov commented on 2026-11-11
max-krasnyansky
am17an GDN: use replay instead of rollback [WIP]
4df118ea
am17an memory : move rs_max with the sequence when find_slot swaps cells
e93f8d08
ServeurpersoCom delta_net: drop the graph replay fallback
5b8d8a7b
ServeurpersoCom metal: replay the pending gated delta net tokens in the kernel
e6fbf3be
ServeurpersoCom vulkan: replay the pending gated delta net tokens in the shader
b85e15ad
ServeurpersoCom webgpu: replay the pending gated delta net tokens in the shader
30062d86
ServeurpersoCom delta_net: share the replay conv window views across layers
e0987318
ServeurpersoCom vulkan: clamp the gated delta net replay count to n_pend
36751006
max-krasnyansky hex-gdn: add support for replay
ab2fb0da
max-krasnyansky gdn: fix rebase conflicts (missing cparams, etc)
207d2c83
max-krasnyansky max-krasnyansky force pushed from 7cc4d991 to 207d2c83 3 days ago
max-krasnyansky
am17an llama : drop the fused GDN probes brought back by the rebase
33a1dbf7
am17an
am17an am17an marked this pull request as ready for review 1 day ago
am17an am17an requested a review 1 day ago
am17an am17an requested a review from CISC CISC 1 day ago
am17an am17an requested a review 1 day ago
am17an am17an requested a review 1 day ago
am17an am17an requested a review 1 day ago
am17an am17an requested a review 1 day ago
am17an am17an requested a review from wine99 wine99 1 day ago
am17an am17an requested a review 1 day ago
am17an am17an requested a review 1 day ago
am17an am17an requested a review from marty1885 marty1885 1 day ago
am17an am17an requested a review 1 day ago
am17an am17an requested a review 1 day ago
am17an am17an requested a review from JohannesGaessler JohannesGaessler 1 day ago
am17an am17an changed the title GDN: use replay instead of rollback [WIP] GDN: use replay instead of rollback 8 hours ago

Login to write a write a comment.

Login via GitHub

Labels
Milestone