Skip to content

ggml-zdnn: fix #15414, activate FP16 and BF16 acceleration and incorrect zTensor free #15839

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 20 commits into from
Sep 12, 2025
+7,760 −3,539
Merged

ggml-zdnn: fix #15414, activate FP16 and BF16 acceleration and incorrect zTensor free #15839

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
20 commits
Select commit Hold shift + click to select a range
47509d4
ggml-zdnn: first rev to fix #15414
taronaeo Sep 6, 2025
6e780a4
ggml-zdnn: trying to fix set_tensors without needing additional if guard
taronaeo Sep 6, 2025
bf285e0
Revert "ggml-zdnn: trying to fix set_tensors without needing addition…
taronaeo Sep 6, 2025
f4ec752
ggml-zdnn: clean up set_tensor
taronaeo Sep 6, 2025
e0bae5d
ggml-zdnn: remove old code
taronaeo Sep 6, 2025
7de719a
ggml-zdnn: attempt at init bias in init_tensor
taronaeo Sep 6, 2025
81e2025
ggml-zdnn: add assert
taronaeo Sep 6, 2025
5c31f9b
ggml-zdnn: attempt at fixing sigsegv
taronaeo Sep 6, 2025
1a6d62b
ggml-zdnn: attempt at fixing double free
taronaeo Sep 6, 2025
8279a1c
ggml-zdnn: fix incorrect ztensor free
taronaeo Sep 6, 2025
0a08b1d
ggml-zdnn: add comments to look back
taronaeo Sep 6, 2025
9ed5947
ggml-zdnn: activate fp16 and bf16
taronaeo Sep 6, 2025
99311bc
ggml-zdnn: clean up matmul codepath
taronaeo Sep 6, 2025
d5b32ff
ggml-zdnn: fix compiler warnings
taronaeo Sep 6, 2025
53b2ad9
ggml-zdnn: fix more compiler warnings
taronaeo Sep 6, 2025
4f6be46
ggml-zdnn: rm origtensor from .get_tensor
taronaeo Sep 6, 2025
0da4b6a
ggml-cpu: clean up s390x simd
taronaeo Sep 7, 2025
5542933
Revert "ggml-cpu: clean up s390x simd"
taronaeo Sep 7, 2025
f231a9d
docs: update ops and build for zDNN
taronaeo Sep 7, 2025
326ce81
ggml-zdnn: rm ggml_backend_zdnn_init from header files
taronaeo Sep 10, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions docs/build-s390x.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -241,8 +241,8 @@ IBM VXE/VXE2 SIMD acceleration depends on the BLAS implementation. It is strongl
| | VX/VXE/VXE2 | zDNN | Spyre |
|------------|-------------|------|-------|
| FP32 | ✅ | ✅ | ❓ |
| FP16 | ✅ | ❓ | ❓ |
| BF16 | 🚫 | ❓ | ❓ |
| FP16 | ✅ | ✅ | ❓ |
| BF16 | 🚫 | ✅ | ❓ |
| Q4_0 | ✅ | ❓ | ❓ |
| Q4_1 | ✅ | ❓ | ❓ |
| MXFP4 | 🚫 | ❓ | ❓ |
Expand Down Expand Up @@ -272,4 +272,4 @@ IBM VXE/VXE2 SIMD acceleration depends on the BLAS implementation. It is strongl
- 🚫 - acceleration unavailable, will still run using scalar implementation
- ❓ - acceleration unknown, please contribute if you can test it yourself

Last Updated by **Aaron Teo ([email protected])** on Sep 6, 2025.
Last Updated by **Aaron Teo ([email protected])** on Sep 7, 2025.
7 changes: 7 additions & 0 deletions docs/ops.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,7 @@ Legend:
| ACC | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ |
| ADD | ❌ | ✅ | ✅ | ✅ | 🟡 | 🟡 | ✅ | ✅ | ❌ |
| ADD1 | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | ✅ | ❌ | ❌ |
| ADD_ID | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
| ARANGE | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ | ❌ |
| ARGMAX | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ |
| ARGSORT | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ |
Expand All @@ -26,6 +27,7 @@ Legend:
| CONT | ❌ | 🟡 | ✅ | ✅ | ✅ | 🟡 | 🟡 | 🟡 | ❌ |
| CONV_2D | ❌ | ❌ | ✅ | ❌ | ❌ | ✅ | ❌ | ✅ | ❌ |
| CONV_2D_DW | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ | ✅ | ❌ |
| CONV_3D | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
| CONV_TRANSPOSE_1D | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ |
| CONV_TRANSPOSE_2D | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ |
| COS | ❌ | ✅ | ✅ | ✅ | 🟡 | ❌ | ✅ | 🟡 | ❌ |
Expand All @@ -49,9 +51,11 @@ Legend:
| GET_ROWS | ❌ | 🟡 | ✅ | 🟡 | ✅ | 🟡 | 🟡 | 🟡 | ❌ |
| GET_ROWS_BACK | ❌ | ❌ | 🟡 | 🟡 | ❌ | ❌ | ❌ | ❌ | ❌ |
| GROUP_NORM | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ |
| GROUP_NORM_MUL_ADD | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
| HARDSIGMOID | ❌ | ✅ | ✅ | 🟡 | 🟡 | ❌ | 🟡 | ❌ | ❌ |
| HARDSWISH | ❌ | ✅ | ✅ | 🟡 | 🟡 | ❌ | 🟡 | ❌ | ❌ |
| IM2COL | ❌ | ✅ | ✅ | ✅ | 🟡 | ✅ | ✅ | ✅ | ❌ |
| IM2COL_3D | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
| L2_NORM | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ |
| LEAKY_RELU | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ |
| LOG | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | ✅ | ❌ | ❌ |
Expand All @@ -61,7 +65,9 @@ Legend:
| MUL_MAT_ID | ❌ | 🟡 | ✅ | ✅ | ✅ | 🟡 | 🟡 | ✅ | ❌ |
| NEG | ❌ | ✅ | ✅ | 🟡 | 🟡 | ❌ | 🟡 | ❌ | ❌ |
| NORM | ❌ | ✅ | ✅ | ✅ | 🟡 | ✅ | ✅ | 🟡 | ❌ |
| NORM_MUL_ADD | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
| OPT_STEP_ADAMW | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ | ✅ | ❌ |
| OPT_STEP_SGD | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
| OUT_PROD | 🟡 | ❌ | 🟡 | 🟡 | ❌ | ❌ | 🟡 | ❌ | ❌ |
| PAD | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ |
| PAD_REFLECT_1D | ❌ | ✅ | ✅ | ❌ | ✅ | ❌ | ❌ | ❌ | ❌ |
Expand Down Expand Up @@ -98,6 +104,7 @@ Legend:
| SUM | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | ✅ | ✅ | ❌ |
| SUM_ROWS | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ |
| SWIGLU | ❌ | ✅ | ✅ | ✅ | 🟡 | ✅ | ✅ | 🟡 | ❌ |
| SWIGLU_OAI | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
| TANH | ❌ | ✅ | ✅ | 🟡 | 🟡 | ✅ | 🟡 | 🟡 | ❌ |
| TIMESTEP_EMBEDDING | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ |
| UPSCALE | ❌ | 🟡 | ✅ | ✅ | 🟡 | ✅ | 🟡 | ✅ | ❌ |
Loading
Loading