mirror of
https://github.com/ml-explore/mlx.git
synced 2025-11-07 04:38:13 +08:00
rebase
This commit is contained in:
@@ -207,17 +207,17 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00099" name="l00099"></a><span class="lineno"> 99</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>& a_stride,</div>
|
||||
<div class="line"><a id="l00100" name="l00100"></a><span class="lineno"> 100</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>& b_stride,</div>
|
||||
<div class="line"><a id="l00101" name="l00101"></a><span class="lineno"> 101</span> uint index [[thread_position_in_grid]]) {</div>
|
||||
<div class="line"><a id="l00102" name="l00102"></a><span class="lineno"> 102</span> <span class="keyword">auto</span> a_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a196a07022b812b241d4c06192c0fa83d">elem_to_loc_1</a>(index, a_stride);</div>
|
||||
<div class="line"><a id="l00103" name="l00103"></a><span class="lineno"> 103</span> <span class="keyword">auto</span> b_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a196a07022b812b241d4c06192c0fa83d">elem_to_loc_1</a>(index, b_stride);</div>
|
||||
<div class="line"><a id="l00102" name="l00102"></a><span class="lineno"> 102</span> <span class="keyword">auto</span> a_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#ac612d0ae30b8257198339debe04916a3">elem_to_loc_1<size_t, uint></a>(index, a_stride);</div>
|
||||
<div class="line"><a id="l00103" name="l00103"></a><span class="lineno"> 103</span> <span class="keyword">auto</span> b_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#ac612d0ae30b8257198339debe04916a3">elem_to_loc_1<size_t, uint></a>(index, b_stride);</div>
|
||||
<div class="line"><a id="l00104" name="l00104"></a><span class="lineno"> 104</span> <span class="keyword">auto</span> out = Op()(a[a_idx], b[b_idx]);</div>
|
||||
<div class="line"><a id="l00105" name="l00105"></a><span class="lineno"> 105</span> c[index] = out[0];</div>
|
||||
<div class="line"><a id="l00106" name="l00106"></a><span class="lineno"> 106</span> d[index] = out[1];</div>
|
||||
<div class="line"><a id="l00107" name="l00107"></a><span class="lineno"> 107</span>}</div>
|
||||
</div>
|
||||
<div class="line"><a id="l00108" name="l00108"></a><span class="lineno"> 108</span> </div>
|
||||
<div class="line"><a id="l00109" name="l00109"></a><span class="lineno"> 109</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> U, <span class="keyword">typename</span> Op></div>
|
||||
<div class="line"><a id="l00109" name="l00109"></a><span class="lineno"> 109</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> U, <span class="keyword">typename</span> Op, <span class="keyword">typename</span> IdxT = <span class="keywordtype">size_t</span>></div>
|
||||
<div class="foldopen" id="foldopen00110" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00110" name="l00110"></a><span class="lineno"><a class="line" href="metal_2kernels_2binary__two_8h.html#a03f7c15a1607576755abb65c542ae347"> 110</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2binary__two_8h.html#a03f7c15a1607576755abb65c542ae347">binary_g_nd2</a>(</div>
|
||||
<div class="line"><a id="l00110" name="l00110"></a><span class="lineno"><a class="line" href="metal_2kernels_2binary__two_8h.html#a97b5613aff654d32c49225209a19bb95"> 110</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2binary__two_8h.html#a97b5613aff654d32c49225209a19bb95">binary_g_nd2</a>(</div>
|
||||
<div class="line"><a id="l00111" name="l00111"></a><span class="lineno"> 111</span> device <span class="keyword">const</span> T* a,</div>
|
||||
<div class="line"><a id="l00112" name="l00112"></a><span class="lineno"> 112</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00113" name="l00113"></a><span class="lineno"> 113</span> device U* c,</div>
|
||||
@@ -226,18 +226,18 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00116" name="l00116"></a><span class="lineno"> 116</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> b_strides[2],</div>
|
||||
<div class="line"><a id="l00117" name="l00117"></a><span class="lineno"> 117</span> uint2 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00118" name="l00118"></a><span class="lineno"> 118</span> uint2 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00119" name="l00119"></a><span class="lineno"> 119</span> <span class="keyword">auto</span> a_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#ad6c45cacca97899cd362df49c06fea79">elem_to_loc_2</a>(index, a_strides);</div>
|
||||
<div class="line"><a id="l00120" name="l00120"></a><span class="lineno"> 120</span> <span class="keyword">auto</span> b_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#ad6c45cacca97899cd362df49c06fea79">elem_to_loc_2</a>(index, b_strides);</div>
|
||||
<div class="line"><a id="l00121" name="l00121"></a><span class="lineno"> 121</span> <span class="keywordtype">size_t</span> out_idx = index.x + size_t(grid_dim.x) * index.y;</div>
|
||||
<div class="line"><a id="l00119" name="l00119"></a><span class="lineno"> 119</span> <span class="keyword">auto</span> a_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a43f33efc000962d6de881a3aab7458de">elem_to_loc_2<size_t, IdxT></a>(index, a_strides);</div>
|
||||
<div class="line"><a id="l00120" name="l00120"></a><span class="lineno"> 120</span> <span class="keyword">auto</span> b_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a43f33efc000962d6de881a3aab7458de">elem_to_loc_2<size_t, IdxT></a>(index, b_strides);</div>
|
||||
<div class="line"><a id="l00121" name="l00121"></a><span class="lineno"> 121</span> IdxT out_idx = index.x + IdxT(grid_dim.x) * index.y;</div>
|
||||
<div class="line"><a id="l00122" name="l00122"></a><span class="lineno"> 122</span> <span class="keyword">auto</span> out = Op()(a[a_idx], b[b_idx]);</div>
|
||||
<div class="line"><a id="l00123" name="l00123"></a><span class="lineno"> 123</span> c[out_idx] = out[0];</div>
|
||||
<div class="line"><a id="l00124" name="l00124"></a><span class="lineno"> 124</span> d[out_idx] = out[1];</div>
|
||||
<div class="line"><a id="l00125" name="l00125"></a><span class="lineno"> 125</span>}</div>
|
||||
</div>
|
||||
<div class="line"><a id="l00126" name="l00126"></a><span class="lineno"> 126</span> </div>
|
||||
<div class="line"><a id="l00127" name="l00127"></a><span class="lineno"> 127</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> U, <span class="keyword">typename</span> Op></div>
|
||||
<div class="line"><a id="l00127" name="l00127"></a><span class="lineno"> 127</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> U, <span class="keyword">typename</span> Op, <span class="keyword">typename</span> IdxT = <span class="keywordtype">size_t</span>></div>
|
||||
<div class="foldopen" id="foldopen00128" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00128" name="l00128"></a><span class="lineno"><a class="line" href="metal_2kernels_2binary__two_8h.html#a43e5943460996c43060d1f3aa1309ba6"> 128</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2binary__two_8h.html#a43e5943460996c43060d1f3aa1309ba6">binary_g_nd3</a>(</div>
|
||||
<div class="line"><a id="l00128" name="l00128"></a><span class="lineno"><a class="line" href="metal_2kernels_2binary__two_8h.html#aae07014f8dffa3649a5c7f4671e1268e"> 128</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2binary__two_8h.html#aae07014f8dffa3649a5c7f4671e1268e">binary_g_nd3</a>(</div>
|
||||
<div class="line"><a id="l00129" name="l00129"></a><span class="lineno"> 129</span> device <span class="keyword">const</span> T* a,</div>
|
||||
<div class="line"><a id="l00130" name="l00130"></a><span class="lineno"> 130</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00131" name="l00131"></a><span class="lineno"> 131</span> device U* c,</div>
|
||||
@@ -246,56 +246,59 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00134" name="l00134"></a><span class="lineno"> 134</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> b_strides[3],</div>
|
||||
<div class="line"><a id="l00135" name="l00135"></a><span class="lineno"> 135</span> uint3 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00136" name="l00136"></a><span class="lineno"> 136</span> uint3 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00137" name="l00137"></a><span class="lineno"> 137</span> <span class="keyword">auto</span> a_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a2c34ed54714c69e6e1b44344f9e6e330">elem_to_loc_3</a>(index, a_strides);</div>
|
||||
<div class="line"><a id="l00138" name="l00138"></a><span class="lineno"> 138</span> <span class="keyword">auto</span> b_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a2c34ed54714c69e6e1b44344f9e6e330">elem_to_loc_3</a>(index, b_strides);</div>
|
||||
<div class="line"><a id="l00139" name="l00139"></a><span class="lineno"> 139</span> <span class="keywordtype">size_t</span> out_idx =</div>
|
||||
<div class="line"><a id="l00140" name="l00140"></a><span class="lineno"> 140</span> index.x + grid_dim.x * (index.y + size_t(grid_dim.y) * index.z);</div>
|
||||
<div class="line"><a id="l00141" name="l00141"></a><span class="lineno"> 141</span> <span class="keyword">auto</span> out = Op()(a[a_idx], b[b_idx]);</div>
|
||||
<div class="line"><a id="l00142" name="l00142"></a><span class="lineno"> 142</span> c[out_idx] = out[0];</div>
|
||||
<div class="line"><a id="l00143" name="l00143"></a><span class="lineno"> 143</span> d[out_idx] = out[1];</div>
|
||||
<div class="line"><a id="l00144" name="l00144"></a><span class="lineno"> 144</span>}</div>
|
||||
<div class="line"><a id="l00137" name="l00137"></a><span class="lineno"> 137</span> <span class="keyword">auto</span> a_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a650f8ea8cf9f9519da9e301aad0308dc">elem_to_loc_3<size_t, IdxT></a>(index, a_strides);</div>
|
||||
<div class="line"><a id="l00138" name="l00138"></a><span class="lineno"> 138</span> <span class="keyword">auto</span> b_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a650f8ea8cf9f9519da9e301aad0308dc">elem_to_loc_3<size_t, IdxT></a>(index, b_strides);</div>
|
||||
<div class="line"><a id="l00139" name="l00139"></a><span class="lineno"> 139</span> IdxT out_idx = index.x + grid_dim.x * (index.y + IdxT(grid_dim.y) * index.z);</div>
|
||||
<div class="line"><a id="l00140" name="l00140"></a><span class="lineno"> 140</span> <span class="keyword">auto</span> out = Op()(a[a_idx], b[b_idx]);</div>
|
||||
<div class="line"><a id="l00141" name="l00141"></a><span class="lineno"> 141</span> c[out_idx] = out[0];</div>
|
||||
<div class="line"><a id="l00142" name="l00142"></a><span class="lineno"> 142</span> d[out_idx] = out[1];</div>
|
||||
<div class="line"><a id="l00143" name="l00143"></a><span class="lineno"> 143</span>}</div>
|
||||
</div>
|
||||
<div class="line"><a id="l00145" name="l00145"></a><span class="lineno"> 145</span> </div>
|
||||
<div class="line"><a id="l00146" name="l00146"></a><span class="lineno"> 146</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> U, <span class="keyword">typename</span> Op, <span class="keywordtype">int</span> N = 1></div>
|
||||
<div class="foldopen" id="foldopen00147" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00147" name="l00147"></a><span class="lineno"><a class="line" href="metal_2kernels_2binary__two_8h.html#a6297badf47dece518bb4e67f02cffea8"> 147</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2binary__two_8h.html#a6297badf47dece518bb4e67f02cffea8">binary_g</a>(</div>
|
||||
<div class="line"><a id="l00148" name="l00148"></a><span class="lineno"> 148</span> device <span class="keyword">const</span> T* a,</div>
|
||||
<div class="line"><a id="l00149" name="l00149"></a><span class="lineno"> 149</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00150" name="l00150"></a><span class="lineno"> 150</span> device U* c,</div>
|
||||
<div class="line"><a id="l00151" name="l00151"></a><span class="lineno"> 151</span> device U* d,</div>
|
||||
<div class="line"><a id="l00152" name="l00152"></a><span class="lineno"> 152</span> constant <span class="keyword">const</span> <span class="keywordtype">int</span>* shape,</div>
|
||||
<div class="line"><a id="l00153" name="l00153"></a><span class="lineno"> 153</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>* a_strides,</div>
|
||||
<div class="line"><a id="l00154" name="l00154"></a><span class="lineno"> 154</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>* b_strides,</div>
|
||||
<div class="line"><a id="l00155" name="l00155"></a><span class="lineno"> 155</span> constant <span class="keyword">const</span> <span class="keywordtype">int</span>& ndim,</div>
|
||||
<div class="line"><a id="l00156" name="l00156"></a><span class="lineno"> 156</span> uint3 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00157" name="l00157"></a><span class="lineno"> 157</span> uint3 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00158" name="l00158"></a><span class="lineno"> 158</span> <span class="keyword">auto</span> idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a01c9309978a6c12f79b6e4108728a953">elem_to_loc_2_nd</a>(</div>
|
||||
<div class="line"><a id="l00159" name="l00159"></a><span class="lineno"> 159</span> {N * index.x, index.y, index.z}, shape, a_strides, b_strides, ndim);</div>
|
||||
<div class="line"><a id="l00160" name="l00160"></a><span class="lineno"> 160</span> <span class="keyword">auto</span> xshape = shape[ndim - 1];</div>
|
||||
<div class="line"><a id="l00161" name="l00161"></a><span class="lineno"> 161</span> <span class="keywordtype">size_t</span> out_idx =</div>
|
||||
<div class="line"><a id="l00162" name="l00162"></a><span class="lineno"> 162</span> N * index.x + xshape * (index.y + size_t(grid_dim.y) * index.z);</div>
|
||||
<div class="line"><a id="l00163" name="l00163"></a><span class="lineno"> 163</span> <span class="keyword">auto</span> a_xstride = a_strides[ndim - 1];</div>
|
||||
<div class="line"><a id="l00164" name="l00164"></a><span class="lineno"> 164</span> <span class="keyword">auto</span> b_xstride = b_strides[ndim - 1];</div>
|
||||
<div class="line"><a id="l00165" name="l00165"></a><span class="lineno"> 165</span> <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0; i < N && (int(N * index.x) + i) < xshape; ++i) {</div>
|
||||
<div class="line"><a id="l00166" name="l00166"></a><span class="lineno"> 166</span> <span class="keyword">auto</span> out = Op()(a[idx.x], b[idx.y]);</div>
|
||||
<div class="line"><a id="l00167" name="l00167"></a><span class="lineno"> 167</span> c[out_idx] = out[0];</div>
|
||||
<div class="line"><a id="l00168" name="l00168"></a><span class="lineno"> 168</span> d[out_idx++] = out[1];</div>
|
||||
<div class="line"><a id="l00169" name="l00169"></a><span class="lineno"> 169</span> idx.x += a_xstride;</div>
|
||||
<div class="line"><a id="l00170" name="l00170"></a><span class="lineno"> 170</span> idx.y += b_xstride;</div>
|
||||
<div class="line"><a id="l00171" name="l00171"></a><span class="lineno"> 171</span> }</div>
|
||||
<div class="line"><a id="l00172" name="l00172"></a><span class="lineno"> 172</span>}</div>
|
||||
<div class="line"><a id="l00144" name="l00144"></a><span class="lineno"> 144</span> </div>
|
||||
<div class="line"><a id="l00145" name="l00145"></a><span class="lineno"> 145</span><span class="keyword">template</span> <</div>
|
||||
<div class="line"><a id="l00146" name="l00146"></a><span class="lineno"> 146</span> <span class="keyword">typename</span> T,</div>
|
||||
<div class="line"><a id="l00147" name="l00147"></a><span class="lineno"> 147</span> <span class="keyword">typename</span> U,</div>
|
||||
<div class="line"><a id="l00148" name="l00148"></a><span class="lineno"> 148</span> <span class="keyword">typename</span> Op,</div>
|
||||
<div class="line"><a id="l00149" name="l00149"></a><span class="lineno"> 149</span> <span class="keywordtype">int</span> N = 1,</div>
|
||||
<div class="line"><a id="l00150" name="l00150"></a><span class="lineno"> 150</span> <span class="keyword">typename</span> IdxT = <span class="keywordtype">size_t</span>></div>
|
||||
<div class="foldopen" id="foldopen00151" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00151" name="l00151"></a><span class="lineno"><a class="line" href="metal_2kernels_2binary__two_8h.html#aaf6edb734cea627bca4f6540dc338fbd"> 151</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2binary__two_8h.html#aaf6edb734cea627bca4f6540dc338fbd">binary_g</a>(</div>
|
||||
<div class="line"><a id="l00152" name="l00152"></a><span class="lineno"> 152</span> device <span class="keyword">const</span> T* a,</div>
|
||||
<div class="line"><a id="l00153" name="l00153"></a><span class="lineno"> 153</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00154" name="l00154"></a><span class="lineno"> 154</span> device U* c,</div>
|
||||
<div class="line"><a id="l00155" name="l00155"></a><span class="lineno"> 155</span> device U* d,</div>
|
||||
<div class="line"><a id="l00156" name="l00156"></a><span class="lineno"> 156</span> constant <span class="keyword">const</span> <span class="keywordtype">int</span>* shape,</div>
|
||||
<div class="line"><a id="l00157" name="l00157"></a><span class="lineno"> 157</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>* a_strides,</div>
|
||||
<div class="line"><a id="l00158" name="l00158"></a><span class="lineno"> 158</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>* b_strides,</div>
|
||||
<div class="line"><a id="l00159" name="l00159"></a><span class="lineno"> 159</span> constant <span class="keyword">const</span> <span class="keywordtype">int</span>& ndim,</div>
|
||||
<div class="line"><a id="l00160" name="l00160"></a><span class="lineno"> 160</span> uint3 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00161" name="l00161"></a><span class="lineno"> 161</span> uint3 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00162" name="l00162"></a><span class="lineno"> 162</span> <span class="keyword">auto</span> idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a66a2d7eec0262b12db16cd6c781ccf9a">elem_to_loc_2_nd<size_t, IdxT></a>(</div>
|
||||
<div class="line"><a id="l00163" name="l00163"></a><span class="lineno"> 163</span> {N * index.x, index.y, index.z}, shape, a_strides, b_strides, ndim);</div>
|
||||
<div class="line"><a id="l00164" name="l00164"></a><span class="lineno"> 164</span> <span class="keyword">auto</span> xshape = shape[ndim - 1];</div>
|
||||
<div class="line"><a id="l00165" name="l00165"></a><span class="lineno"> 165</span> IdxT out_idx = N * index.x + xshape * (index.y + IdxT(grid_dim.y) * index.z);</div>
|
||||
<div class="line"><a id="l00166" name="l00166"></a><span class="lineno"> 166</span> IdxT a_xstride = a_strides[ndim - 1];</div>
|
||||
<div class="line"><a id="l00167" name="l00167"></a><span class="lineno"> 167</span> IdxT b_xstride = b_strides[ndim - 1];</div>
|
||||
<div class="line"><a id="l00168" name="l00168"></a><span class="lineno"> 168</span> <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0; i < N && (int(N * index.x) + i) < xshape; ++i) {</div>
|
||||
<div class="line"><a id="l00169" name="l00169"></a><span class="lineno"> 169</span> <span class="keyword">auto</span> out = Op()(a[idx.x], b[idx.y]);</div>
|
||||
<div class="line"><a id="l00170" name="l00170"></a><span class="lineno"> 170</span> c[out_idx] = out[0];</div>
|
||||
<div class="line"><a id="l00171" name="l00171"></a><span class="lineno"> 171</span> d[out_idx++] = out[1];</div>
|
||||
<div class="line"><a id="l00172" name="l00172"></a><span class="lineno"> 172</span> idx.x += a_xstride;</div>
|
||||
<div class="line"><a id="l00173" name="l00173"></a><span class="lineno"> 173</span> idx.y += b_xstride;</div>
|
||||
<div class="line"><a id="l00174" name="l00174"></a><span class="lineno"> 174</span> }</div>
|
||||
<div class="line"><a id="l00175" name="l00175"></a><span class="lineno"> 175</span>}</div>
|
||||
</div>
|
||||
<div class="ttc" id="abackend_2metal_2kernels_2utils_8h_html_a01c9309978a6c12f79b6e4108728a953"><div class="ttname"><a href="backend_2metal_2kernels_2utils_8h.html#a01c9309978a6c12f79b6e4108728a953">elem_to_loc_2_nd</a></div><div class="ttdeci">METAL_FUNC ulong2 elem_to_loc_2_nd(uint3 elem, constant const int *shape, constant const stride_t *a_strides, constant const stride_t *b_strides, int ndim)</div><div class="ttdef"><b>Definition</b> utils.h:153</div></div>
|
||||
<div class="ttc" id="abackend_2metal_2kernels_2utils_8h_html_a196a07022b812b241d4c06192c0fa83d"><div class="ttname"><a href="backend_2metal_2kernels_2utils_8h.html#a196a07022b812b241d4c06192c0fa83d">elem_to_loc_1</a></div><div class="ttdeci">METAL_FUNC stride_t elem_to_loc_1(uint elem, constant const stride_t &stride)</div><div class="ttdef"><b>Definition</b> utils.h:133</div></div>
|
||||
<div class="ttc" id="abackend_2metal_2kernels_2utils_8h_html_a2c34ed54714c69e6e1b44344f9e6e330"><div class="ttname"><a href="backend_2metal_2kernels_2utils_8h.html#a2c34ed54714c69e6e1b44344f9e6e330">elem_to_loc_3</a></div><div class="ttdeci">METAL_FUNC stride_t elem_to_loc_3(uint3 elem, constant const stride_t strides[3])</div><div class="ttdef"><b>Definition</b> utils.h:145</div></div>
|
||||
<div class="ttc" id="abackend_2metal_2kernels_2utils_8h_html_ad6c45cacca97899cd362df49c06fea79"><div class="ttname"><a href="backend_2metal_2kernels_2utils_8h.html#ad6c45cacca97899cd362df49c06fea79">elem_to_loc_2</a></div><div class="ttdeci">METAL_FUNC stride_t elem_to_loc_2(uint2 elem, constant const stride_t strides[2])</div><div class="ttdef"><b>Definition</b> utils.h:139</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_a03f7c15a1607576755abb65c542ae347"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#a03f7c15a1607576755abb65c542ae347">binary_g_nd2</a></div><div class="ttdeci">void binary_g_nd2(device const T *a, device const T *b, device U *c, device U *d, constant const size_t a_strides[2], constant const size_t b_strides[2], uint2 index, uint2 grid_dim)</div><div class="ttdef"><b>Definition</b> binary_two.h:110</div></div>
|
||||
<div class="ttc" id="abackend_2metal_2kernels_2utils_8h_html_a43f33efc000962d6de881a3aab7458de"><div class="ttname"><a href="backend_2metal_2kernels_2utils_8h.html#a43f33efc000962d6de881a3aab7458de">elem_to_loc_2</a></div><div class="ttdeci">METAL_FUNC IdxT elem_to_loc_2(uint2 elem, constant const StrideT strides[2])</div><div class="ttdef"><b>Definition</b> utils.h:145</div></div>
|
||||
<div class="ttc" id="abackend_2metal_2kernels_2utils_8h_html_a650f8ea8cf9f9519da9e301aad0308dc"><div class="ttname"><a href="backend_2metal_2kernels_2utils_8h.html#a650f8ea8cf9f9519da9e301aad0308dc">elem_to_loc_3</a></div><div class="ttdeci">METAL_FUNC IdxT elem_to_loc_3(uint3 elem, constant const StrideT strides[3])</div><div class="ttdef"><b>Definition</b> utils.h:150</div></div>
|
||||
<div class="ttc" id="abackend_2metal_2kernels_2utils_8h_html_a66a2d7eec0262b12db16cd6c781ccf9a"><div class="ttname"><a href="backend_2metal_2kernels_2utils_8h.html#a66a2d7eec0262b12db16cd6c781ccf9a">elem_to_loc_2_nd</a></div><div class="ttdeci">METAL_FUNC vec< IdxT, 2 > elem_to_loc_2_nd(uint3 elem, constant const int *shape, constant const StrideT *a_strides, constant const StrideT *b_strides, int ndim)</div><div class="ttdef"><b>Definition</b> utils.h:159</div></div>
|
||||
<div class="ttc" id="abackend_2metal_2kernels_2utils_8h_html_ac612d0ae30b8257198339debe04916a3"><div class="ttname"><a href="backend_2metal_2kernels_2utils_8h.html#ac612d0ae30b8257198339debe04916a3">elem_to_loc_1</a></div><div class="ttdeci">METAL_FUNC IdxT elem_to_loc_1(uint elem, constant const StrideT &stride)</div><div class="ttdef"><b>Definition</b> utils.h:140</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_a08822ff98ea6f61a98b49a9e9a38b891"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#a08822ff98ea6f61a98b49a9e9a38b891">binary_sv2</a></div><div class="ttdeci">void binary_sv2(device const T *a, device const T *b, device U *c, device U *d, uint2 index, uint2 grid_dim)</div><div class="ttdef"><b>Definition</b> binary_two.h:52</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_a12dbda74fa460812177ccb9aeee6e1ca"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#a12dbda74fa460812177ccb9aeee6e1ca">binary_vs</a></div><div class="ttdeci">void binary_vs(device const T *a, device const T *b, device U *c, device U *d, uint index)</div><div class="ttdef"><b>Definition</b> binary_two.h:28</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_a12e80730e43dfaa4c79ce8d5f99edc50"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#a12e80730e43dfaa4c79ce8d5f99edc50">binary_vv2</a></div><div class="ttdeci">void binary_vv2(device const T *a, device const T *b, device U *c, device U *d, uint2 index, uint2 grid_dim)</div><div class="ttdef"><b>Definition</b> binary_two.h:80</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_a273d2f31691f2c64623c2a97eab344be"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#a273d2f31691f2c64623c2a97eab344be">binary_vs2</a></div><div class="ttdeci">void binary_vs2(device const T *a, device const T *b, device U *c, device U *d, uint2 index, uint2 grid_dim)</div><div class="ttdef"><b>Definition</b> binary_two.h:66</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_a43e5943460996c43060d1f3aa1309ba6"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#a43e5943460996c43060d1f3aa1309ba6">binary_g_nd3</a></div><div class="ttdeci">void binary_g_nd3(device const T *a, device const T *b, device U *c, device U *d, constant const size_t a_strides[3], constant const size_t b_strides[3], uint3 index, uint3 grid_dim)</div><div class="ttdef"><b>Definition</b> binary_two.h:128</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_a6297badf47dece518bb4e67f02cffea8"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#a6297badf47dece518bb4e67f02cffea8">binary_g</a></div><div class="ttdeci">void binary_g(device const T *a, device const T *b, device U *c, device U *d, constant const int *shape, constant const size_t *a_strides, constant const size_t *b_strides, constant const int &ndim, uint3 index, uint3 grid_dim)</div><div class="ttdef"><b>Definition</b> binary_two.h:147</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_a97b5613aff654d32c49225209a19bb95"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#a97b5613aff654d32c49225209a19bb95">binary_g_nd2</a></div><div class="ttdeci">void binary_g_nd2(device const T *a, device const T *b, device U *c, device U *d, constant const size_t a_strides[2], constant const size_t b_strides[2], uint2 index, uint2 grid_dim)</div><div class="ttdef"><b>Definition</b> binary_two.h:110</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_aae07014f8dffa3649a5c7f4671e1268e"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#aae07014f8dffa3649a5c7f4671e1268e">binary_g_nd3</a></div><div class="ttdeci">void binary_g_nd3(device const T *a, device const T *b, device U *c, device U *d, constant const size_t a_strides[3], constant const size_t b_strides[3], uint3 index, uint3 grid_dim)</div><div class="ttdef"><b>Definition</b> binary_two.h:128</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_aaf6edb734cea627bca4f6540dc338fbd"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#aaf6edb734cea627bca4f6540dc338fbd">binary_g</a></div><div class="ttdeci">void binary_g(device const T *a, device const T *b, device U *c, device U *d, constant const int *shape, constant const size_t *a_strides, constant const size_t *b_strides, constant const int &ndim, uint3 index, uint3 grid_dim)</div><div class="ttdef"><b>Definition</b> binary_two.h:151</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_ab18c6ecf5065275c93701efd095c916c"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#ab18c6ecf5065275c93701efd095c916c">binary_sv</a></div><div class="ttdeci">void binary_sv(device const T *a, device const T *b, device U *c, device U *d, uint index)</div><div class="ttdef"><b>Definition</b> binary_two.h:16</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_ab4324f594c007a6895540b77ad5d89d9"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#ab4324f594c007a6895540b77ad5d89d9">binary_vv</a></div><div class="ttdeci">void binary_vv(device const T *a, device const T *b, device U *c, device U *d, uint index)</div><div class="ttdef"><b>Definition</b> binary_two.h:40</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2binary__two_8h_html_ad1fad37c168192b212a4294f4cf78133"><div class="ttname"><a href="metal_2kernels_2binary__two_8h.html#ad1fad37c168192b212a4294f4cf78133">binary_g_nd1</a></div><div class="ttdeci">void binary_g_nd1(device const T *a, device const T *b, device U *c, device U *d, constant const size_t &a_stride, constant const size_t &b_stride, uint index)</div><div class="ttdef"><b>Definition</b> binary_two.h:94</div></div>
|
||||
|
||||
Reference in New Issue
Block a user