root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-28 12:24:35 +00:00
Code Issues Actions 3 Packages Projects Releases Wiki Activity
1,913 Commits 359 Branches 2,870 Tags 394 MiB
1453215165
Commit Graph

14 Commits

Author SHA1 Message Date
Georgi Gerganov
1453215165
kompute : fix ggml_add kernel 2024-01-19 00:09:16 +02:00
Jared Van Bortel
610394fff8 fix supported ops for kompute backend 2024-01-18 16:36:18 -05:00
Jared Van Bortel
7addf2b878 never try to evaluate an empty command buffer 
This fixes the immediate crashes with test-backend-ops - when
evaluatating individual no-ops like OP_VIEW, it tries to submit an empty
command buffer, which crashes RADV and hangs AMDVLK.
 2024-01-18 16:33:28 -05:00
Jared Van Bortel
0f1a958a51 actually fix this assertion 2024-01-18 11:49:49 -05:00
Jared Van Bortel
a97935e098 clean up old backend code 2024-01-18 11:49:49 -05:00
Georgi Gerganov
696faa8660
kompute : fix rope_f32 and scale ops (#5008) 2024-01-18 11:49:39 -05:00
Jared Van Bortel
02b9bafe29 kompute : ignore exceptions in ggml_vk_available_devices (#12) 
Signed-off-by: Jared Van Bortel <jared@nomic.ai>
 2024-01-17 14:02:41 -05:00
Jared Van Bortel
de9b0bbbe4 add sanity check and fix kompute teardown order 2024-01-17 10:09:27 -05:00
Jared Van Bortel
50579f27e9 attempt to get test-backend-ops working 2024-01-10 16:37:01 -05:00
Jared Van Bortel
8a99f69895 fix assertion failure 2024-01-10 13:44:34 -05:00
Jared Van Bortel
d5670d6e46 kompute : initial attempt at ggml-backend v2 support 2024-01-10 12:26:31 -05:00
Jared Van Bortel
8072706210 kompute : always destroy Manager via the destructor 2023-12-15 16:23:24 -05:00
Jared Van Bortel
b906e126ca kompute : fix compile warnings 2023-12-13 17:49:45 -05:00
Jared Van Bortel
3e09e127eb rename ggml-vulkan -> ggml-kompute 2023-12-13 17:49:45 -05:00