arm_compute v19.05
diff --git a/documentation/quantization__layer_8cl.xhtml b/documentation/quantization__layer_8cl.xhtml
index 3441b3b..3aad982 100644
--- a/documentation/quantization__layer_8cl.xhtml
+++ b/documentation/quantization__layer_8cl.xhtml
@@ -40,7 +40,7 @@
<img alt="Compute Library" src="https://raw.githubusercontent.com/ARM-software/ComputeLibrary/gh-pages/ACL_logo.png" style="max-width: 100%;margin-top: 15px;margin-left: 10px"/>
<td style="padding-left: 0.5em;">
<div id="projectname">
-  <span id="projectnumber">19.02</span>
+  <span id="projectnumber">19.05</span>
</div>
</td>
</tr>
@@ -97,7 +97,7 @@
<div class="header">
<div class="summary">
-<a href="#func-members">Functions</a> </div>
+<a href="#define-members">Macros</a> </div>
<div class="headertitle">
<div class="title">quantization_layer.cl File Reference</div> </div>
</div><!--header-->
@@ -106,181 +106,115 @@
</div>
<p><a href="quantization__layer_8cl_source.xhtml">Go to the source code of this file.</a></p>
<table class="memberdecls">
-<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="func-members"></a>
-Functions</h2></td></tr>
-<tr class="memitem:a1f16fa20ae72c76df16155e6a3ad20c5"><td class="memItemLeft" align="right" valign="top">__kernel void </td><td class="memItemRight" valign="bottom"><a class="el" href="quantization__layer_8cl.xhtml#a1f16fa20ae72c76df16155e6a3ad20c5">quantization_layer</a> (__global uchar *input_ptr, uint input_stride_x, uint input_step_x, uint input_stride_y, uint input_step_y, uint input_stride_z, uint input_step_z, uint input_offset_first_element_in_bytes, __global uchar *output_ptr, uint output_stride_x, uint output_step_x, uint output_stride_y, uint output_step_y, uint output_stride_z, uint output_step_z, uint output_offset_first_element_in_bytes, __global uchar *min_max_ptr, uint min_max_stride_x, uint min_max_step_x, uint min_max_offset_first_element_in_bytes)</td></tr>
-<tr class="memdesc:a1f16fa20ae72c76df16155e6a3ad20c5"><td class="mdescLeft"> </td><td class="mdescRight">This performs the quantization of floating point inputs to 8-bit unsigned integers. <a href="#a1f16fa20ae72c76df16155e6a3ad20c5">More...</a><br /></td></tr>
-<tr class="separator:a1f16fa20ae72c76df16155e6a3ad20c5"><td class="memSeparator" colspan="2"> </td></tr>
+<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="define-members"></a>
+Macros</h2></td></tr>
+<tr class="memitem:ab448fef4390583fc4b3832f8c590a2b0"><td class="memItemLeft" align="right" valign="top">#define </td><td class="memItemRight" valign="bottom"><a class="el" href="quantization__layer_8cl.xhtml#ab448fef4390583fc4b3832f8c590a2b0">CONVERT_RTE</a>(x, type)   (convert_##type##_rte((x)))</td></tr>
+<tr class="separator:ab448fef4390583fc4b3832f8c590a2b0"><td class="memSeparator" colspan="2"> </td></tr>
+<tr class="memitem:a85ea2f9c49c7c00d41c25f6d5e0687e1"><td class="memItemLeft" align="right" valign="top">#define </td><td class="memItemRight" valign="bottom"><a class="el" href="quantization__layer_8cl.xhtml#a85ea2f9c49c7c00d41c25f6d5e0687e1">CONVERT_RTE_VEC_STR</a>(x, type, size)   (convert_##type##size##_rte((x)))</td></tr>
+<tr class="separator:a85ea2f9c49c7c00d41c25f6d5e0687e1"><td class="memSeparator" colspan="2"> </td></tr>
+<tr class="memitem:a69f26e0fbc99605b9464b5500d4c461a"><td class="memItemLeft" align="right" valign="top">#define </td><td class="memItemRight" valign="bottom"><a class="el" href="quantization__layer_8cl.xhtml#a69f26e0fbc99605b9464b5500d4c461a">CONVERT_RTE_VEC</a>(x, type, size)   <a class="el" href="quantization__layer_8cl.xhtml#a85ea2f9c49c7c00d41c25f6d5e0687e1">CONVERT_RTE_VEC_STR</a>(x, type, size)</td></tr>
+<tr class="separator:a69f26e0fbc99605b9464b5500d4c461a"><td class="memSeparator" colspan="2"> </td></tr>
</table>
-<h2 class="groupheader">Function Documentation</h2>
-<a id="a1f16fa20ae72c76df16155e6a3ad20c5"></a>
-<h2 class="memtitle"><span class="permalink"><a href="#a1f16fa20ae72c76df16155e6a3ad20c5">◆ </a></span>quantization_layer()</h2>
+<h2 class="groupheader">Macro Definition Documentation</h2>
+<a id="ab448fef4390583fc4b3832f8c590a2b0"></a>
+<h2 class="memtitle"><span class="permalink"><a href="#ab448fef4390583fc4b3832f8c590a2b0">◆ </a></span>CONVERT_RTE</h2>
<div class="memitem">
<div class="memproto">
<table class="memname">
<tr>
- <td class="memname">__kernel void quantization_layer </td>
+ <td class="memname">#define CONVERT_RTE</td>
<td>(</td>
- <td class="paramtype">__global uchar * </td>
- <td class="paramname"><em>input_ptr</em>, </td>
+ <td class="paramtype"> </td>
+ <td class="paramname">x, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>input_stride_x</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>input_step_x</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>input_stride_y</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>input_step_y</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>input_stride_z</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>input_step_z</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>input_offset_first_element_in_bytes</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">__global uchar * </td>
- <td class="paramname"><em>output_ptr</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>output_stride_x</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>output_step_x</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>output_stride_y</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>output_step_y</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>output_stride_z</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>output_step_z</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>output_offset_first_element_in_bytes</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">__global uchar * </td>
- <td class="paramname"><em>min_max_ptr</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>min_max_stride_x</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>min_max_step_x</em>, </td>
- </tr>
- <tr>
- <td class="paramkey"></td>
- <td></td>
- <td class="paramtype">uint </td>
- <td class="paramname"><em>min_max_offset_first_element_in_bytes</em> </td>
+ <td class="paramtype"> </td>
+ <td class="paramname">type </td>
</tr>
<tr>
<td></td>
<td>)</td>
- <td></td><td></td>
+ <td></td><td>   (convert_##type##_rte((x)))</td>
</tr>
</table>
</div><div class="memdoc">
-<p>This performs the quantization of floating point inputs to 8-bit unsigned integers. </p>
-<dl class="params"><dt>Parameters</dt><dd>
- <table class="params">
- <tr><td class="paramdir">[in]</td><td class="paramname">input_ptr</td><td>Pointer to the source image. Supported data types: F32 </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">input_stride_x</td><td>Stride of the source image in X dimension (in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">input_step_x</td><td>input_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">input_stride_y</td><td>Stride of the source image in Y dimension (in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">input_step_y</td><td>input_stride_y * number of elements along Y processed per workitem(in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">input_stride_z</td><td>Stride of the source tensor in Z dimension (in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">input_step_z</td><td>input_stride_z * number of elements along Z processed per workitem(in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">input_offset_first_element_in_bytes</td><td>The offset of the first element in the source image </td></tr>
- <tr><td class="paramdir">[out]</td><td class="paramname">output_ptr</td><td>Pointer to the destination image. Supported data types: U8 </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">output_stride_x</td><td>Stride of the destination image in X dimension (in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">output_step_x</td><td>output_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">output_stride_y</td><td>Stride of the destination image in Y dimension (in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">output_step_y</td><td>output_stride_y * number of elements along Y processed per workitem(in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">output_stride_z</td><td>Stride of the source tensor in Z dimension (in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">output_step_z</td><td>output_stride_z * number of elements along Z processed per workitem(in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">output_offset_first_element_in_bytes</td><td>The offset of the first element in the destination image </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">min_max_ptr</td><td>Pointer to the min/max vector. Minimum value in position 0, maximum value in position 1. Supported data types: F32. </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">min_max_stride_x</td><td>Stride of the min/max vector in X dimension (in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">min_max_step_x</td><td>min_max_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
- <tr><td class="paramdir">[in]</td><td class="paramname">min_max_offset_first_element_in_bytes</td><td>The offset of the first element in the min/max vector </td></tr>
- </table>
- </dd>
-</dl>
+<p class="definition">Definition at line <a class="el" href="quantization__layer_8cl_source.xhtml#l00026">26</a> of file <a class="el" href="quantization__layer_8cl_source.xhtml">quantization_layer.cl</a>.</p>
-<p class="definition">Definition at line <a class="el" href="quantization__layer_8cl_source.xhtml#l00049">49</a> of file <a class="el" href="quantization__layer_8cl_source.xhtml">quantization_layer.cl</a>.</p>
-<div class="fragment"><div class="line"><a name="l00053"></a><span class="lineno"> 53</span> {</div><div class="line"><a name="l00054"></a><span class="lineno"> 54</span>  <span class="comment">// Get pixels pointer</span></div><div class="line"><a name="l00055"></a><span class="lineno"> 55</span>  <a class="code" href="struct_tensor3_d.xhtml">Tensor3D</a> input = <a class="code" href="src_2core_2_c_l_2cl__kernels_2_helpers_8h.xhtml#a31c8c760f08fb1a331b16b7c204321dc">CONVERT_TO_TENSOR3D_STRUCT</a>(input);</div><div class="line"><a name="l00056"></a><span class="lineno"> 56</span>  <a class="code" href="struct_tensor3_d.xhtml">Tensor3D</a> output = <a class="code" href="src_2core_2_c_l_2cl__kernels_2_helpers_8h.xhtml#a31c8c760f08fb1a331b16b7c204321dc">CONVERT_TO_TENSOR3D_STRUCT</a>(output);</div><div class="line"><a name="l00057"></a><span class="lineno"> 57</span> </div><div class="line"><a name="l00058"></a><span class="lineno"> 58</span>  <span class="comment">// min_max_value.s0 = min, min_max_value.s1 = max</span></div><div class="line"><a name="l00059"></a><span class="lineno"> 59</span>  <span class="keyword">const</span> float2 min_max_value = vload2(0, (__global <span class="keywordtype">float</span> *)(min_max_ptr + min_max_offset_first_element_in_bytes));</div><div class="line"><a name="l00060"></a><span class="lineno"> 60</span> </div><div class="line"><a name="l00061"></a><span class="lineno"> 61</span>  <span class="keyword">const</span> float4 <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ab026a1aeec48074092f548c130c323f4">vmin</a> = (float4)min_max_value.s0;</div><div class="line"><a name="l00062"></a><span class="lineno"> 62</span>  <span class="keyword">const</span> float4 vrange = (float4)(min_max_value.s1 - min_max_value.s0);</div><div class="line"><a name="l00063"></a><span class="lineno"> 63</span> </div><div class="line"><a name="l00064"></a><span class="lineno"> 64</span>  <span class="comment">// Load data</span></div><div class="line"><a name="l00065"></a><span class="lineno"> 65</span>  float4 data = vload4(0, (__global <span class="keywordtype">float</span> *)input.<a class="code" href="struct_tensor3_d.xhtml#acf52c23cbd7424606c10a606524e3e32">ptr</a>);</div><div class="line"><a name="l00066"></a><span class="lineno"> 66</span> </div><div class="line"><a name="l00067"></a><span class="lineno"> 67</span>  <span class="comment">// Map float values to range [0.0, 1.0]</span></div><div class="line"><a name="l00068"></a><span class="lineno"> 68</span>  data = (data - <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ab026a1aeec48074092f548c130c323f4">vmin</a>) / vrange;</div><div class="line"><a name="l00069"></a><span class="lineno"> 69</span> </div><div class="line"><a name="l00070"></a><span class="lineno"> 70</span>  <span class="comment">// Quantize and saturate</span></div><div class="line"><a name="l00071"></a><span class="lineno"> 71</span>  uchar4 res = convert_uchar4_sat(data * 256.0f);</div><div class="line"><a name="l00072"></a><span class="lineno"> 72</span> </div><div class="line"><a name="l00073"></a><span class="lineno"> 73</span>  <span class="comment">// Store result</span></div><div class="line"><a name="l00074"></a><span class="lineno"> 74</span>  vstore4(res, 0, (__global uchar *)output.<a class="code" href="struct_tensor3_d.xhtml#acf52c23cbd7424606c10a606524e3e32">ptr</a>);</div><div class="line"><a name="l00075"></a><span class="lineno"> 75</span> }</div><div class="ttc" id="struct_tensor3_d_xhtml"><div class="ttname"><a href="struct_tensor3_d.xhtml">Tensor3D</a></div><div class="ttdoc">Structure to hold 3D tensor information.</div><div class="ttdef"><b>Definition:</b> <a href="src_2core_2_c_l_2cl__kernels_2_helpers_8h_source.xhtml#l00160">helpers.h:160</a></div></div>
-<div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_ab026a1aeec48074092f548c130c323f4"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#ab026a1aeec48074092f548c130c323f4">arm_compute::wrapper::vmin</a></div><div class="ttdeci">uint8x8_t vmin(const uint8x8_t &a, const uint8x8_t &b)</div><div class="ttdef"><b>Definition:</b> <a href="min_8h_source.xhtml#l00039">min.h:39</a></div></div>
-<div class="ttc" id="src_2core_2_c_l_2cl__kernels_2_helpers_8h_xhtml_a31c8c760f08fb1a331b16b7c204321dc"><div class="ttname"><a href="src_2core_2_c_l_2cl__kernels_2_helpers_8h.xhtml#a31c8c760f08fb1a331b16b7c204321dc">CONVERT_TO_TENSOR3D_STRUCT</a></div><div class="ttdeci">#define CONVERT_TO_TENSOR3D_STRUCT(name)</div><div class="ttdef"><b>Definition:</b> <a href="src_2core_2_c_l_2cl__kernels_2_helpers_8h_source.xhtml#l00128">helpers.h:128</a></div></div>
-<div class="ttc" id="struct_tensor3_d_xhtml_acf52c23cbd7424606c10a606524e3e32"><div class="ttname"><a href="struct_tensor3_d.xhtml#acf52c23cbd7424606c10a606524e3e32">Tensor3D::ptr</a></div><div class="ttdeci">__global uchar * ptr</div><div class="ttdoc">Pointer to the starting postion of the buffer.</div><div class="ttdef"><b>Definition:</b> <a href="src_2core_2_c_l_2cl__kernels_2_helpers_8h_source.xhtml#l00162">helpers.h:162</a></div></div>
-</div><!-- fragment -->
-<p class="reference">References <a class="el" href="src_2core_2_c_l_2cl__kernels_2_helpers_8h_source.xhtml#l00128">CONVERT_TO_TENSOR3D_STRUCT</a>, <a class="el" href="src_2core_2_c_l_2cl__kernels_2_helpers_8h_source.xhtml#l00162">Tensor3D::ptr</a>, and <a class="el" href="min_8h_source.xhtml#l00039">arm_compute::wrapper::vmin()</a>.</p>
+</div>
+</div>
+<a id="a69f26e0fbc99605b9464b5500d4c461a"></a>
+<h2 class="memtitle"><span class="permalink"><a href="#a69f26e0fbc99605b9464b5500d4c461a">◆ </a></span>CONVERT_RTE_VEC</h2>
+
+<div class="memitem">
+<div class="memproto">
+ <table class="memname">
+ <tr>
+ <td class="memname">#define CONVERT_RTE_VEC</td>
+ <td>(</td>
+ <td class="paramtype"> </td>
+ <td class="paramname">x, </td>
+ </tr>
+ <tr>
+ <td class="paramkey"></td>
+ <td></td>
+ <td class="paramtype"> </td>
+ <td class="paramname">type, </td>
+ </tr>
+ <tr>
+ <td class="paramkey"></td>
+ <td></td>
+ <td class="paramtype"> </td>
+ <td class="paramname">size </td>
+ </tr>
+ <tr>
+ <td></td>
+ <td>)</td>
+ <td></td><td>   <a class="el" href="quantization__layer_8cl.xhtml#a85ea2f9c49c7c00d41c25f6d5e0687e1">CONVERT_RTE_VEC_STR</a>(x, type, size)</td>
+ </tr>
+ </table>
+</div><div class="memdoc">
+
+<p class="definition">Definition at line <a class="el" href="quantization__layer_8cl_source.xhtml#l00028">28</a> of file <a class="el" href="quantization__layer_8cl_source.xhtml">quantization_layer.cl</a>.</p>
+
+</div>
+</div>
+<a id="a85ea2f9c49c7c00d41c25f6d5e0687e1"></a>
+<h2 class="memtitle"><span class="permalink"><a href="#a85ea2f9c49c7c00d41c25f6d5e0687e1">◆ </a></span>CONVERT_RTE_VEC_STR</h2>
+
+<div class="memitem">
+<div class="memproto">
+ <table class="memname">
+ <tr>
+ <td class="memname">#define CONVERT_RTE_VEC_STR</td>
+ <td>(</td>
+ <td class="paramtype"> </td>
+ <td class="paramname">x, </td>
+ </tr>
+ <tr>
+ <td class="paramkey"></td>
+ <td></td>
+ <td class="paramtype"> </td>
+ <td class="paramname">type, </td>
+ </tr>
+ <tr>
+ <td class="paramkey"></td>
+ <td></td>
+ <td class="paramtype"> </td>
+ <td class="paramname">size </td>
+ </tr>
+ <tr>
+ <td></td>
+ <td>)</td>
+ <td></td><td>   (convert_##type##size##_rte((x)))</td>
+ </tr>
+ </table>
+</div><div class="memdoc">
+
+<p class="definition">Definition at line <a class="el" href="quantization__layer_8cl_source.xhtml#l00027">27</a> of file <a class="el" href="quantization__layer_8cl_source.xhtml">quantization_layer.cl</a>.</p>
</div>
</div>
@@ -290,7 +224,7 @@
<div id="nav-path" class="navpath"><!-- id is needed for treeview function! -->
<ul>
<li class="navelem"><a class="el" href="dir_68267d1309a1af8e8297ef4c3efbcdba.xhtml">src</a></li><li class="navelem"><a class="el" href="dir_aebb8dcc11953d78e620bbef0b9e2183.xhtml">core</a></li><li class="navelem"><a class="el" href="dir_8c278f79c760e5c5fbd911f9870614c1.xhtml">CL</a></li><li class="navelem"><a class="el" href="dir_25885286e9dad4fa105b7b25a8031bbf.xhtml">cl_kernels</a></li><li class="navelem"><a class="el" href="quantization__layer_8cl.xhtml">quantization_layer.cl</a></li>
- <li class="footer">Generated on Thu Feb 28 2019 12:24:56 for Compute Library by
+ <li class="footer">Generated on Thu May 23 2019 17:11:26 for Compute Library by
<a href="http://www.doxygen.org/index.html">
<img class="footer" src="doxygen.png" alt="doxygen"/></a> 1.8.15 </li>
</ul>