mudler-agentandEttore Di Giacinto 849c177834 feat(llama-cpp): bump llama.cpp and add moe_cache_mib option (#12729) ...

Bump LLAMA_VERSION to b76b13379d3b (upstream 0.7.0 plus later
commits). This is the same one-line change as the bot PR #12725.

Add the moe_cache_mib model option. It sets upstream's
moe_cache_size, a GPU cache for MoE experts kept in host memory,
from a size in MiB. It is off by default. Invalid, negative or
non-numeric values are ignored with a log line. Forks that predate
the field ignore the option, using the same prepare.sh probe as the
other fork-skew switches.

Document the option, and cover the MiB to bytes conversion with a
standalone unit test.

Assisted-by: Claude:claude-sonnet-5-5 g++

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Co-authored-by: Ettore Di Giacinto <mudler@localai.io>

2026-10-16 00:55:55 +02:00

..

audio-cpp

chore: ⬆️ Update 0xShug0/audio.cpp to 22cad6af5043a2d7c7f1d9a0e6c9d7996a80dfc8 (#12723)

2026-10-16 00:55:04 +02:00

bonsai

chore: ⬆️ Update PrismML-Eng/llama.cpp to e3de715222b2667c54a267441159a5729f11155e (#12680)

2026-10-15 12:14:19 +02:00

ds4

chore: ⬆️ Update antirez/ds4 to fc80bd695da76ee14bcb86820bf900a5cdfda806 (#12580)

2026-10-09 12:20:17 +02:00

grpc

fix: speedup git submodule update with --single-branch (#2847)

2024-07-13 22:32:25 +02:00

ik-llama-cpp

chore: ⬆️ Update ikawrakow/ik_llama.cpp to 8313388ab528503bf503aee0c51e7b30e1cda378 (#12724)

2026-10-16 00:55:24 +02:00

llama-cpp

feat(llama-cpp): bump llama.cpp and add moe_cache_mib option (#12729)

2026-10-16 00:55:55 +02:00

privacy-filter

refactor(backends): extract package-system-libs.sh from 31 package.sh (#11095)

2026-07-30 16:30:49 +02:00

turboquant

chore: ⬆️ Update TheTom/llama-cpp-turboquant to ca28b4bc0e531635d83fe530c6e6e8d694c4330d (#12699)

2026-10-15 14:26:33 +02:00

run-unit-tests.sh

feat(audio-cpp): add the audio.cpp native backend (#11141)

2026-07-30 12:11:56 +02:00