You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
@@ -107,15 +111,18 @@ Items are added as a python dictionary when invoking the MIGraphX execution prov
107
111
| migraphx_int8_enable | 1 or 0 | Enable int8 static quantization mode of the input model via the MIGraphX API. Requires calibration table path vars to be set (migraphx_int8_calibration_table_name=valid path).|
108
112
| migraphx_int8_calibration_table_name | <absolute path to calibration table> | Path to a set of input calibration data for int8 static model quantization. |
109
113
| migraphx_int8_use_native_calibration_table | 1 or 0 | Use a calibration table from Nvidia native int8 format or json dumped format. |
110
-
| migraphx_save_compiled_model | 1 or 0 | Enable saving a model as an MIGraphX (.mxr) format after compile when set to 1 |
111
-
| migraphx_save_compiled_path | path string | Path to write .mxr file (default is ./compiled_model.mxr). Path where the MIGraphX compiled model is stored. Requires migraphx_save_compiled_path to be set for this session |
112
-
| migraphx_load_compiled_model | 1 or 0 | Enable loading a model as an MIGraphX (.mxr) format after compile when set to 1 |
113
-
| migraphx_load_compiled_path | path string | Path to read .mxr file (default is ./compiled_model.mxr). Path where the MIGraphX compiled model is stored. |
114
114
| migraphx_exhaustive_tune | 1 or 0 (default 0) | Enable exhaustive tuning of parameters as part of compilation via the MIGraphX API. Adds additional compile time for a potential perf boost.|
115
115
| migraphx_mem_limit | INT | Set the memory limit used for memory arena. Default uses ORTs default_memory_arena_cfg value. |
116
116
| migraphx_external_alloc | Address | Address of external memory allocator function used for this EP. Useful for reading in larger models weights. |
117
117
| migraphx_external_free | Address | Address of external memory deallocator function used for this EP. Useful for unloadng what was allocated with the migraphx_external_alloc input. |
118
118
| migraphx_external_empty_cache | Address | Address of external memory cache used for this model. Useful for caching results of externally allocated models. |
119
+
| | | |
120
+
| Depricated | Release Removed | Description |
121
+
| migraphx_save_compiled_model | ROCm 6.4 | Enable saving a model as an MIGraphX (.mxr) format after compile when set to 1 |
122
+
| migraphx_save_compiled_path | ROcm 6.4 | Path to write .mxr file (default is ./compiled_model.mxr). Path where the MIGraphX compiled model is stored. Requires migraphx_save_compiled_path to be set for this session |
123
+
| migraphx_load_compiled_model | ROCm 6.4 | Enable loading a model as an MIGraphX (.mxr) format after compile when set to 1 |
124
+
| migraphx_load_compiled_path | ROCm 6.4 | Path to read .mxr file (default is ./compiled_model.mxr). Path where the MIGraphX compiled model is stored. |
125
+
119
126
120
127
121
128
## Environment Variables
@@ -143,11 +150,15 @@ Users can invoke Environment and Session variables in the same run but Environme
143
150
| ORT_MIGRAPHX_INT8_ENABLE | 1 or 0 | Enable int8 static quantization mode of the input model via the MIGraphX API.\n Requires calibration table path vars to be set (migraphx_int8_calibration_table_name=<valid path>).|
144
151
| ORT_MIGRAPHX_INT8_CALIBRATION_TABLE_NAME | <absolute path to calibration table> | Path to a set of input calibration data for int8 static model quantization. |
145
152
| ORT_MIGRAPHX_INT8_USE_NATIVE_CALIBRATION_TABLE | 1 or 0 | Use a calibration table from Nvidia native int8 format or json dumped format. |
146
-
| ORT_MIGRAPHX_SAVE_COMPILED_MODEL | 1 or 0 | Enable saving a model as an MIGraphX (.mxr) format after compile. |
147
-
| ORT_MIGRAPHX_SAVE_COMPILED_PATH | <string> - Path to write .mxr file (default is ./compiled_model.mxr) | Path where the MIGraphX compiled model is stored. Requires ORT_MIGRAPHX_SAVE_COMPILED_MODEL to be set to 1. |
148
-
| ORT_MIGRAPHX_LOAD_COMPILED_MODEL | 1 or 0 | Enable loading a model as an MIGraphX (.mxr) format after compile. |
149
-
| ORT_MIGRAPHX_LOAD_COMPILED_PATH | <string> - Path to read .mxr file (default is ./compiled_model.mxr) | Path where the MIGraphX compiled model is stored. Requires ORT_MIGRAPHX_LOAD_COMPILED_MODEL to be set to 1. |
150
153
| ORT_MIGRAPHX_EXHAUSTIVE_TUNE | 1 or 0 (default 0) | Enable exhaustive tuning of parameters as part of compilation via the MIGraphX API. Adds additional compile time for a potential perf boost. |
154
+
| ORT_MIGRAPHX_MODEL_CACHE_PATH | <string> | Path to read and write model specific data such as weights or other model specific data |
155
+
| ORT_MIGRAPHX_MODEL_PATH | <string> | Path to read and write .mxr path occurs after MIGraphX model compile complete |
156
+
| | | |
157
+
| Depricated | ROCm Version removed | Description |
158
+
| ORT_MIGRAPHX_SAVE_COMPILED_MODEL | ROCm 6.4 | Enable saving a model as an MIGraphX (.mxr) format after compile. ( 0 or 1) |
159
+
| ORT_MIGRAPHX_SAVE_COMPILED_PATH | ROCm 6.4 | Path where the MIGraphX compiled model is stored. Requires ORT_MIGRAPHX_SAVE_COMPILED_MODEL to be set to 1. |
160
+
| ORT_MIGRAPHX_LOAD_COMPILED_MODEL | ROCm 6.4 | Enable loading a model as an MIGraphX (.mxr) format after compile. (0 or 1) |
161
+
| ORT_MIGRAPHX_LOAD_COMPILED_PATH | ROCm 6.4 | Path where the MIGraphX compiled model is stored. Requires ORT_MIGRAPHX_LOAD_COMPILED_MODEL to be set to 1. |
The ROCm Execution Provider enables hardware accelerated computation on AMD ROCm-enabled GPUs.
13
13
14
+
** NOTE ** As of ROCm 7.1 There will be no more ROCm Execution Provider support provider by Microsoft
15
+
16
+
Please Migrate your applications to use the [MIGraphX Execution Provider](https://onnxruntime.ai/docs/execution-providers/MIGraphX-ExecutionProvider.html#migraphx-execution-provider)
17
+
18
+
ROCm 7.0 is the last offiicaly AMD supported distribution of this provider and all builds going forward (ROCm 7.1+) Will have ROCm EP removed.
19
+
20
+
Please refer to this [Pull Request](https://github.com/microsoft/onnxruntime/pull/25181) for background.
21
+
14
22
## Contents
15
23
{: .no_toc }
16
24
@@ -36,9 +44,18 @@ Below is the matrix of supported ROCm versions corresponding to Ubuntu builds.
36
44
As of ROCm 6.0.2 Links for prebuild Python Wheels (.whl) are linked below corresponding to python versions for the host OS based on Ubuntu support.
37
45
All links can be found on AMD's [repo.radeon manylinux page](https://repo.radeon.com/rocm/manylinux) for each corresponding to the ROCm release.
38
46
47
+
ROCm 7.0 Will be the last officially supported AMD Release which includes ROCm Execution Provider
48
+
Please use [MIGraphX Execution Provider](https://onnxruntime.ai/docs/execution-providers/MIGraphX-ExecutionProvider.html#migraphx-execution-provider) For your applications instead.
49
+
39
50
| ONNX Runtime Version | ROCm Version | Python 3.8 | Python 3.9 | Python 3.10 | Python 3.12 |
0 commit comments