llama.cpp
ggml-zdnn: fix #15414, activate FP16 and BF16 acceleration and incorrect zTensor free
#15839
Merged
Merged
ggml-zdnn: fix #15414, activate FP16 and BF16 acceleration and incorrect zTensor free #15839
taronaeo
merged
20 commits into
ggml-org:master
from
taronaeo:fix/15414-llama-set-rows
ggml-zdnn: first rev to fix #15414
47509d42
ggml-zdnn: trying to fix set_tensors without needing additional if guard
6e780a41
Revert "ggml-zdnn: trying to fix set_tensors without needing addition…
bf285e0a
ggml-zdnn: clean up set_tensor
f4ec752d
ggml-zdnn: remove old code
e0bae5dd
ggml-zdnn: attempt at init bias in init_tensor
7de719a4
ggml-zdnn: add assert
81e2025a
ggml-zdnn: attempt at fixing sigsegv
5c31f9b2
ggml-zdnn: attempt at fixing double free
1a6d62b7
ggml-zdnn: fix incorrect ztensor free
8279a1c7
ggml-zdnn: add comments to look back
0a08b1df
ggml-zdnn: activate fp16 and bf16
9ed5947a
ggml-zdnn: clean up matmul codepath
99311bc9
ggml-zdnn: fix compiler warnings
d5b32ff6
ggml-zdnn: fix more compiler warnings
53b2ad94
ggml-zdnn: rm origtensor from .get_tensor
4f6be460
ggml-cpu: clean up s390x simd
0da4b6aa
Revert "ggml-cpu: clean up s390x simd"
5542933d
docs: update ops and build for zDNN
f231a9d8
slaren
commented
on 2025-09-09
ggml-zdnn: rm ggml_backend_zdnn_init from header files
326ce817
slaren
approved these changes
on 2025-09-12
Reviewers
Assignees
No one assigned
Labels
documentation
ggml
IBM zDNN
Milestone
No milestone
Login to write a write a comment.
Login via GitHub