sync : ggml #3264
ggerganov
merged
26 commits into
master
from
sync-ggml-25-06-18
examples : include examples in msvc disable warn (ggml/1270)
d15be12a
ggml : remove unused ggml_context_container (ggml/1272)
44dbaa5e
ggml : disable warnings for tests when using MSVC (ggml/1273)
0ea0bb8d
rpc : nicer error messages for RPC server crash (llama/14076)
9b850cf9
Vulkan: Don't default to CPU device (like llvmpipe), even if no other…
19f70d6a
opencl: add `mul_mv_id_q4_0_f32_8x_flat` (llama/14003)
ea295ba2
vulkan: Track descriptor pools/sets per-context (llama/14109)
9de0db5b
vulkan: Better thread-safety for command pools/buffers (llama/14116)
ada4c552
Implement GGML_CPU_ALL_VARIANTS for ARM (llama/14080)
fc447c1a
cmake : handle whitepsaces in path during metal build (llama/14126)
7bc2b17b
sycl: Remove not needed copy f16->f32 for dnnl mul mat (llama/14125)
1e6693cb
SYCL: Bump oneMath commit (llama/14152)
c583825a
sycl: Adding additional cpy dbg print output (llama/14034)
28560f97
HIP: Replace usage of depricated preprocessor macro __AMDGCN_WAVEFRON…
39f31e3a
CUDA/HIP: fix ssm_scan on devices where warp size is not 32 (llama/14…
6c721d60
ggml-cpu : rework weak alias on apple targets (llama/14146)
4640699d
vulkan: mutex around vkQueueSubmit (llama/14127)
2c5f56b2
ggml: Add Android support for GGML_CPU_ALL_VARIANTS (llama/14206)
c4049194
HIP: disable rocwmma on gfx12 by default until rocm 7.0 (llama/14202)
acb45d4d
cmake: clean up external project logic for vulkan-shaders-gen (llama/…
91f8856b
llama : add thread safety test (llama/14035)
311fa204
musa: fix build warning (unused variable) (llama/14231)
2400bdb3
ggml-cpu : remove the weak alias trick (llama/14221)
9686aece
cmake: remove shader-gen step-targets from ggml-vulkan (llama/14226)
ad8b05b0
sync : ggml
a8d075df
talk-llama : sync llama.cpp
9cc9701e
danbev
approved these changes
on 2025-06-18
Login to write a write a comment.
Login via GitHub