Update mixed_precision.md

docs/source/mixed_precision.md

@@ -20,14 +20,99 @@ The recently launched 3rd Gen Intel® Xeon® Scalable processor (codenamed Coope
 
 ## Mixed Precision Support Matrix
 
-|Framework     |BF16         |FP16         |
-|--------------|:-----------:|:-----------:|
-|TensorFlow    |✔     |:x:     |
-|PyTorch       |✔     |:x:     |
-|ONNX Runtime  |✔     |✔     |
-|MXNet         |✔     |:x:     |
-
-> **During quantization, BF16 conversion is default enabled, FP16 can be executed if 'device' of config is 'gpu'. Please refer to this [document](./quantization_mixed_precision.md) for its workflow.**
+<table class="center">
+    <thead>
+        <tr>
+            <th>Framework</th>
+            <th>Backend</th>
+            <th>Backend Library</th>
+            <th>Backend Value</th>
+            <th>Support Device(cpu as default)</th> 
+            <th>Support BF16</th>
+            <th>Support FP16</th>
+        </tr>
+    </thead>
+    <tbody>
+        <tr>
+            <td rowspan="2" align="left">PyTorch</td>
+            <td align="left">FX</td>
+            <td align="left">FBGEMM</td>
+            <td align="left">"default"</td>
+            <td align="left">cpu</td>
+            <td align="left">&#10004;</td>
+            <td align="left">:x:</td>
+        </tr>
+        <tr>
+            <td align="left">IPEX</td>
+            <td align="left">OneDNN</td>
+            <td align="left">"ipex"</td>
+            <td align="left">cpu</td>
+            <td align="left">&#10004;</td>
+            <td align="left">:x:</td>
+        </tr>
+        <tr>
+            <td rowspan="4" align="left">ONNX Runtime</td>
+            <td align="left">CPUExecutionProvider</td>
+            <td align="left">MLAS</td>
+            <td align="left">"default"</td>
+            <td align="left">cpu</td>
+            <td align="left">:x:</td>
+            <td align="left">:x:</td>
+        </tr>
+        <tr>
+            <td align="left">TensorrtExecutionProvider</td>
+            <td align="left">TensorRT</td>
+            <td align="left">"onnxrt_trt_ep"</td>
+            <td align="left">gpu</td>
+            <td align="left">:x:</td>
+            <td align="left">:x:</td>
+        </tr>
+        <tr>
+            <td align="left">CUDAExecutionProvider</td>
+            <td align="left">CUDA</td>
+            <td align="left">"onnxrt_cuda_ep"</td>
+            <td align="left">gpu</td>
+            <td align="left">&#10004;</td>
+            <td align="left">&#10004;</td>
+        </tr>
+        <tr>
+            <td align="left">DnnlExecutionProvider</td>
+            <td align="left">OneDNN</td>
+            <td align="left">"onnxrt_dnnl_ep"</td>
+            <td align="left">cpu</td>
+            <td align="left">&#10004;</td>
+            <td align="left">:x:</td>
+        </tr>
+        <tr>
+            <td rowspan="2" align="left">Tensorflow</td>
+            <td align="left">Tensorflow</td>
+            <td align="left">OneDNN</td>
+            <td align="left">"default"</td>
+            <td align="left">cpu</td>
+            <td align="left">&#10004;</td>
+            <td align="left">:x:</td>
+        </tr>
+        <tr>
+            <td align="left">ITEX</td>
+            <td align="left">OneDNN</td>
+            <td align="left">"itex"</td>
+            <td align="left">cpu | gpu</td>
+            <td align="left">&#10004;</td>
+            <td align="left">:x:</td>
+        </tr>  
+        <tr>
+            <td align="left">MXNet</td>
+            <td align="left">OneDNN</td>
+            <td align="left">OneDNN</td>
+            <td align="left">"default"</td>
+            <td align="left">cpu</td>
+            <td align="left">&#10004;</td>
+            <td align="left">:x:</td>
+        </tr>
+    </tbody>
+</table>
+
+> **During quantization, BF16 conversion is default enabled, FP16 can be executed if 'device' of config is 'gpu' and 'backend' is 'onnxrt_cuda_ep'. Please refer to this [document](./quantization_mixed_precision.md) for its workflow.**
 
 ## Get Started with Mixed Precision API
 
@@ -82,12 +167,14 @@ There are some pre-requirements to run mixed precision examples for each framewo
 
     #### ONNX Runtime
 
-    1. Hardware: GPU, set 'device' of config to 'gpu' and 'backend' to 'onnxrt_cuda_ep'.
-    2. Software: onnxruntime-gpu.
+    1. Hardware: GPU or CPU supports `avx512_bf16` instruction set.
+    2. Software: onnxruntime-gpu (for GPU) or onnxruntime-dnnl (for CPU, refer to this [doc](https://onnxruntime.ai/docs/build/eps.html#onednn) to build onnxruntime-dnnl from source code).
 
 - **FP16:**
 
     #### ONNX Runtime
 
-    1. Hardware: GPU, set 'device' of config to 'gpu' and 'backend' to 'onnxrt_cuda_ep'.
+    1. Hardware: GPU.
     2. Software: onnxruntime-gpu.
+
+> **Note: Please set the backend and device values of config according to [Mixed Precision Support Matrix](#mixed-precision-support-matrix) for mixed precision.**