mirror of
https://github.com/ml-explore/mlx.git
synced 2025-09-18 18:28:12 +08:00
docs update
This commit is contained in:

committed by
CircleCI Docs

parent
6e9288a41c
commit
f5dcb1c2b9
@@ -99,114 +99,129 @@ $(function() { codefold.init(0); });
|
||||
<div class="line"><a id="l00012" name="l00012"></a><span class="lineno"> 12</span> </div>
|
||||
<div class="line"><a id="l00013" name="l00013"></a><span class="lineno"> 13</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> Op></div>
|
||||
<div class="foldopen" id="foldopen00014" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00014" name="l00014"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#a1bd5918559850f3f80e3adee2391fe6a"> 14</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#a1bd5918559850f3f80e3adee2391fe6a">ternary_g_nd1</a>(</div>
|
||||
<div class="line"><a id="l00014" name="l00014"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#a3e610f3b01966bdbf23fdfebe5d2c508"> 14</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#a3e610f3b01966bdbf23fdfebe5d2c508">ternary_v2</a>(</div>
|
||||
<div class="line"><a id="l00015" name="l00015"></a><span class="lineno"> 15</span> device <span class="keyword">const</span> <span class="keywordtype">bool</span>* a,</div>
|
||||
<div class="line"><a id="l00016" name="l00016"></a><span class="lineno"> 16</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00017" name="l00017"></a><span class="lineno"> 17</span> device <span class="keyword">const</span> T* c,</div>
|
||||
<div class="line"><a id="l00018" name="l00018"></a><span class="lineno"> 18</span> device T* d,</div>
|
||||
<div class="line"><a id="l00019" name="l00019"></a><span class="lineno"> 19</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>& a_strides,</div>
|
||||
<div class="line"><a id="l00020" name="l00020"></a><span class="lineno"> 20</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>& b_strides,</div>
|
||||
<div class="line"><a id="l00021" name="l00021"></a><span class="lineno"> 21</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>& c_strides,</div>
|
||||
<div class="line"><a id="l00022" name="l00022"></a><span class="lineno"> 22</span> uint index [[thread_position_in_grid]]) {</div>
|
||||
<div class="line"><a id="l00023" name="l00023"></a><span class="lineno"> 23</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_strides);</div>
|
||||
<div class="line"><a id="l00024" name="l00024"></a><span class="lineno"> 24</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_strides);</div>
|
||||
<div class="line"><a id="l00025" name="l00025"></a><span class="lineno"> 25</span> <span class="keyword">auto</span> c_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a196a07022b812b241d4c06192c0fa83d">elem_to_loc_1</a>(index, c_strides);</div>
|
||||
<div class="line"><a id="l00026" name="l00026"></a><span class="lineno"> 26</span> d[index] = Op()(a[a_idx], b[b_idx], c[c_idx]);</div>
|
||||
<div class="line"><a id="l00027" name="l00027"></a><span class="lineno"> 27</span>}</div>
|
||||
<div class="line"><a id="l00019" name="l00019"></a><span class="lineno"> 19</span> uint2 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00020" name="l00020"></a><span class="lineno"> 20</span> uint2 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00021" name="l00021"></a><span class="lineno"> 21</span> <span class="keywordtype">size_t</span> offset = index.x + grid_dim.x * size_t(index.y);</div>
|
||||
<div class="line"><a id="l00022" name="l00022"></a><span class="lineno"> 22</span> d[offset] = Op()(a[offset], b[offset], c[offset]);</div>
|
||||
<div class="line"><a id="l00023" name="l00023"></a><span class="lineno"> 23</span>}</div>
|
||||
</div>
|
||||
<div class="line"><a id="l00028" name="l00028"></a><span class="lineno"> 28</span> </div>
|
||||
<div class="line"><a id="l00029" name="l00029"></a><span class="lineno"> 29</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> Op></div>
|
||||
<div class="foldopen" id="foldopen00030" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00030" name="l00030"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#afdf0d9d0cb21fcb3f176500785076af8"> 30</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#afdf0d9d0cb21fcb3f176500785076af8">ternary_g_nd2</a>(</div>
|
||||
<div class="line"><a id="l00031" name="l00031"></a><span class="lineno"> 31</span> device <span class="keyword">const</span> <span class="keywordtype">bool</span>* a,</div>
|
||||
<div class="line"><a id="l00032" name="l00032"></a><span class="lineno"> 32</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00033" name="l00033"></a><span class="lineno"> 33</span> device <span class="keyword">const</span> T* c,</div>
|
||||
<div class="line"><a id="l00034" name="l00034"></a><span class="lineno"> 34</span> device T* d,</div>
|
||||
<div class="line"><a id="l00035" name="l00035"></a><span class="lineno"> 35</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> a_strides[2],</div>
|
||||
<div class="line"><a id="l00036" name="l00036"></a><span class="lineno"> 36</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> b_strides[2],</div>
|
||||
<div class="line"><a id="l00037" name="l00037"></a><span class="lineno"> 37</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> c_strides[2],</div>
|
||||
<div class="line"><a id="l00038" name="l00038"></a><span class="lineno"> 38</span> uint2 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00039" name="l00039"></a><span class="lineno"> 39</span> uint2 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00040" name="l00040"></a><span class="lineno"> 40</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="l00041" name="l00041"></a><span class="lineno"> 41</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="l00042" name="l00042"></a><span class="lineno"> 42</span> <span class="keyword">auto</span> c_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#ad6c45cacca97899cd362df49c06fea79">elem_to_loc_2</a>(index, c_strides);</div>
|
||||
<div class="line"><a id="l00043" name="l00043"></a><span class="lineno"> 43</span> <span class="keywordtype">size_t</span> out_idx = index.x + (size_t)grid_dim.x * index.y;</div>
|
||||
<div class="line"><a id="l00044" name="l00044"></a><span class="lineno"> 44</span> d[out_idx] = Op()(a[a_idx], b[b_idx], c[c_idx]);</div>
|
||||
<div class="line"><a id="l00045" name="l00045"></a><span class="lineno"> 45</span>}</div>
|
||||
<div class="line"><a id="l00024" name="l00024"></a><span class="lineno"> 24</span> </div>
|
||||
<div class="line"><a id="l00025" name="l00025"></a><span class="lineno"> 25</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> Op></div>
|
||||
<div class="foldopen" id="foldopen00026" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00026" name="l00026"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#a1bd5918559850f3f80e3adee2391fe6a"> 26</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#a1bd5918559850f3f80e3adee2391fe6a">ternary_g_nd1</a>(</div>
|
||||
<div class="line"><a id="l00027" name="l00027"></a><span class="lineno"> 27</span> device <span class="keyword">const</span> <span class="keywordtype">bool</span>* a,</div>
|
||||
<div class="line"><a id="l00028" name="l00028"></a><span class="lineno"> 28</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00029" name="l00029"></a><span class="lineno"> 29</span> device <span class="keyword">const</span> T* c,</div>
|
||||
<div class="line"><a id="l00030" name="l00030"></a><span class="lineno"> 30</span> device T* d,</div>
|
||||
<div class="line"><a id="l00031" name="l00031"></a><span class="lineno"> 31</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>& a_strides,</div>
|
||||
<div class="line"><a id="l00032" name="l00032"></a><span class="lineno"> 32</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>& b_strides,</div>
|
||||
<div class="line"><a id="l00033" name="l00033"></a><span class="lineno"> 33</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>& c_strides,</div>
|
||||
<div class="line"><a id="l00034" name="l00034"></a><span class="lineno"> 34</span> uint index [[thread_position_in_grid]]) {</div>
|
||||
<div class="line"><a id="l00035" name="l00035"></a><span class="lineno"> 35</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_strides);</div>
|
||||
<div class="line"><a id="l00036" name="l00036"></a><span class="lineno"> 36</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_strides);</div>
|
||||
<div class="line"><a id="l00037" name="l00037"></a><span class="lineno"> 37</span> <span class="keyword">auto</span> c_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a196a07022b812b241d4c06192c0fa83d">elem_to_loc_1</a>(index, c_strides);</div>
|
||||
<div class="line"><a id="l00038" name="l00038"></a><span class="lineno"> 38</span> d[index] = Op()(a[a_idx], b[b_idx], c[c_idx]);</div>
|
||||
<div class="line"><a id="l00039" name="l00039"></a><span class="lineno"> 39</span>}</div>
|
||||
</div>
|
||||
<div class="line"><a id="l00046" name="l00046"></a><span class="lineno"> 46</span> </div>
|
||||
<div class="line"><a id="l00047" name="l00047"></a><span class="lineno"> 47</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> Op></div>
|
||||
<div class="foldopen" id="foldopen00048" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00048" name="l00048"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#a113df0c8a841b0e986900d580644e047"> 48</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#a113df0c8a841b0e986900d580644e047">ternary_g_nd3</a>(</div>
|
||||
<div class="line"><a id="l00049" name="l00049"></a><span class="lineno"> 49</span> device <span class="keyword">const</span> <span class="keywordtype">bool</span>* a,</div>
|
||||
<div class="line"><a id="l00050" name="l00050"></a><span class="lineno"> 50</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00051" name="l00051"></a><span class="lineno"> 51</span> device <span class="keyword">const</span> T* c,</div>
|
||||
<div class="line"><a id="l00052" name="l00052"></a><span class="lineno"> 52</span> device T* d,</div>
|
||||
<div class="line"><a id="l00053" name="l00053"></a><span class="lineno"> 53</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> a_strides[3],</div>
|
||||
<div class="line"><a id="l00054" name="l00054"></a><span class="lineno"> 54</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> b_strides[3],</div>
|
||||
<div class="line"><a id="l00055" name="l00055"></a><span class="lineno"> 55</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> c_strides[3],</div>
|
||||
<div class="line"><a id="l00056" name="l00056"></a><span class="lineno"> 56</span> uint3 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00057" name="l00057"></a><span class="lineno"> 57</span> uint3 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00058" name="l00058"></a><span class="lineno"> 58</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="l00059" name="l00059"></a><span class="lineno"> 59</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="l00060" name="l00060"></a><span class="lineno"> 60</span> <span class="keyword">auto</span> c_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a2c34ed54714c69e6e1b44344f9e6e330">elem_to_loc_3</a>(index, c_strides);</div>
|
||||
<div class="line"><a id="l00061" name="l00061"></a><span class="lineno"> 61</span> <span class="keywordtype">size_t</span> out_idx =</div>
|
||||
<div class="line"><a id="l00062" name="l00062"></a><span class="lineno"> 62</span> index.x + (size_t)grid_dim.x * (index.y + (<span class="keywordtype">size_t</span>)grid_dim.y * index.z);</div>
|
||||
<div class="line"><a id="l00063" name="l00063"></a><span class="lineno"> 63</span> d[out_idx] = Op()(a[a_idx], b[b_idx], c[c_idx]);</div>
|
||||
<div class="line"><a id="l00064" name="l00064"></a><span class="lineno"> 64</span>}</div>
|
||||
<div class="line"><a id="l00040" name="l00040"></a><span class="lineno"> 40</span> </div>
|
||||
<div class="line"><a id="l00041" name="l00041"></a><span class="lineno"> 41</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> Op></div>
|
||||
<div class="foldopen" id="foldopen00042" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00042" name="l00042"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#afdf0d9d0cb21fcb3f176500785076af8"> 42</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#afdf0d9d0cb21fcb3f176500785076af8">ternary_g_nd2</a>(</div>
|
||||
<div class="line"><a id="l00043" name="l00043"></a><span class="lineno"> 43</span> device <span class="keyword">const</span> <span class="keywordtype">bool</span>* a,</div>
|
||||
<div class="line"><a id="l00044" name="l00044"></a><span class="lineno"> 44</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00045" name="l00045"></a><span class="lineno"> 45</span> device <span class="keyword">const</span> T* c,</div>
|
||||
<div class="line"><a id="l00046" name="l00046"></a><span class="lineno"> 46</span> device T* d,</div>
|
||||
<div class="line"><a id="l00047" name="l00047"></a><span class="lineno"> 47</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> a_strides[2],</div>
|
||||
<div class="line"><a id="l00048" name="l00048"></a><span class="lineno"> 48</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> b_strides[2],</div>
|
||||
<div class="line"><a id="l00049" name="l00049"></a><span class="lineno"> 49</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> c_strides[2],</div>
|
||||
<div class="line"><a id="l00050" name="l00050"></a><span class="lineno"> 50</span> uint2 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00051" name="l00051"></a><span class="lineno"> 51</span> uint2 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00052" name="l00052"></a><span class="lineno"> 52</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="l00053" name="l00053"></a><span class="lineno"> 53</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="l00054" name="l00054"></a><span class="lineno"> 54</span> <span class="keyword">auto</span> c_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#ad6c45cacca97899cd362df49c06fea79">elem_to_loc_2</a>(index, c_strides);</div>
|
||||
<div class="line"><a id="l00055" name="l00055"></a><span class="lineno"> 55</span> <span class="keywordtype">size_t</span> out_idx = index.x + (size_t)grid_dim.x * index.y;</div>
|
||||
<div class="line"><a id="l00056" name="l00056"></a><span class="lineno"> 56</span> d[out_idx] = Op()(a[a_idx], b[b_idx], c[c_idx]);</div>
|
||||
<div class="line"><a id="l00057" name="l00057"></a><span class="lineno"> 57</span>}</div>
|
||||
</div>
|
||||
<div class="line"><a id="l00065" name="l00065"></a><span class="lineno"> 65</span> </div>
|
||||
<div class="line"><a id="l00066" name="l00066"></a><span class="lineno"> 66</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> Op, <span class="keywordtype">int</span> DIM></div>
|
||||
<div class="foldopen" id="foldopen00067" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00067" name="l00067"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#a94f4695f471837525e5ca995285bc6aa"> 67</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#a94f4695f471837525e5ca995285bc6aa">ternary_g_nd</a>(</div>
|
||||
<div class="line"><a id="l00068" name="l00068"></a><span class="lineno"> 68</span> device <span class="keyword">const</span> <span class="keywordtype">bool</span>* a,</div>
|
||||
<div class="line"><a id="l00069" name="l00069"></a><span class="lineno"> 69</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00070" name="l00070"></a><span class="lineno"> 70</span> device <span class="keyword">const</span> T* c,</div>
|
||||
<div class="line"><a id="l00071" name="l00071"></a><span class="lineno"> 71</span> device T* d,</div>
|
||||
<div class="line"><a id="l00072" name="l00072"></a><span class="lineno"> 72</span> constant <span class="keyword">const</span> <span class="keywordtype">int</span> shape[DIM],</div>
|
||||
<div class="line"><a id="l00073" name="l00073"></a><span class="lineno"> 73</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> a_strides[DIM],</div>
|
||||
<div class="line"><a id="l00074" name="l00074"></a><span class="lineno"> 74</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> b_strides[DIM],</div>
|
||||
<div class="line"><a id="l00075" name="l00075"></a><span class="lineno"> 75</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> c_strides[DIM],</div>
|
||||
<div class="line"><a id="l00076" name="l00076"></a><span class="lineno"> 76</span> uint3 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00077" name="l00077"></a><span class="lineno"> 77</span> uint3 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00078" name="l00078"></a><span class="lineno"> 78</span> <span class="keyword">auto</span> idx =</div>
|
||||
<div class="line"><a id="l00079" name="l00079"></a><span class="lineno"> 79</span> elem_to_loc_3_nd<DIM>(index, shape, a_strides, b_strides, c_strides);</div>
|
||||
<div class="line"><a id="l00080" name="l00080"></a><span class="lineno"> 80</span> <span class="keywordtype">size_t</span> out_idx =</div>
|
||||
<div class="line"><a id="l00081" name="l00081"></a><span class="lineno"> 81</span> index.x + (size_t)grid_dim.x * (index.y + (<span class="keywordtype">size_t</span>)grid_dim.y * index.z);</div>
|
||||
<div class="line"><a id="l00082" name="l00082"></a><span class="lineno"> 82</span> d[out_idx] = Op()(a[idx.x], b[idx.y], c[idx.z]);</div>
|
||||
<div class="line"><a id="l00083" name="l00083"></a><span class="lineno"> 83</span>}</div>
|
||||
<div class="line"><a id="l00058" name="l00058"></a><span class="lineno"> 58</span> </div>
|
||||
<div class="line"><a id="l00059" name="l00059"></a><span class="lineno"> 59</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> Op></div>
|
||||
<div class="foldopen" id="foldopen00060" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00060" name="l00060"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#a113df0c8a841b0e986900d580644e047"> 60</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#a113df0c8a841b0e986900d580644e047">ternary_g_nd3</a>(</div>
|
||||
<div class="line"><a id="l00061" name="l00061"></a><span class="lineno"> 61</span> device <span class="keyword">const</span> <span class="keywordtype">bool</span>* a,</div>
|
||||
<div class="line"><a id="l00062" name="l00062"></a><span class="lineno"> 62</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00063" name="l00063"></a><span class="lineno"> 63</span> device <span class="keyword">const</span> T* c,</div>
|
||||
<div class="line"><a id="l00064" name="l00064"></a><span class="lineno"> 64</span> device T* d,</div>
|
||||
<div class="line"><a id="l00065" name="l00065"></a><span class="lineno"> 65</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> a_strides[3],</div>
|
||||
<div class="line"><a id="l00066" name="l00066"></a><span class="lineno"> 66</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> b_strides[3],</div>
|
||||
<div class="line"><a id="l00067" name="l00067"></a><span class="lineno"> 67</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> c_strides[3],</div>
|
||||
<div class="line"><a id="l00068" name="l00068"></a><span class="lineno"> 68</span> uint3 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00069" name="l00069"></a><span class="lineno"> 69</span> uint3 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00070" name="l00070"></a><span class="lineno"> 70</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="l00071" name="l00071"></a><span class="lineno"> 71</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="l00072" name="l00072"></a><span class="lineno"> 72</span> <span class="keyword">auto</span> c_idx = <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a2c34ed54714c69e6e1b44344f9e6e330">elem_to_loc_3</a>(index, c_strides);</div>
|
||||
<div class="line"><a id="l00073" name="l00073"></a><span class="lineno"> 73</span> <span class="keywordtype">size_t</span> out_idx =</div>
|
||||
<div class="line"><a id="l00074" name="l00074"></a><span class="lineno"> 74</span> index.x + (size_t)grid_dim.x * (index.y + (<span class="keywordtype">size_t</span>)grid_dim.y * index.z);</div>
|
||||
<div class="line"><a id="l00075" name="l00075"></a><span class="lineno"> 75</span> d[out_idx] = Op()(a[a_idx], b[b_idx], c[c_idx]);</div>
|
||||
<div class="line"><a id="l00076" name="l00076"></a><span class="lineno"> 76</span>}</div>
|
||||
</div>
|
||||
<div class="line"><a id="l00084" name="l00084"></a><span class="lineno"> 84</span> </div>
|
||||
<div class="line"><a id="l00085" name="l00085"></a><span class="lineno"> 85</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> Op></div>
|
||||
<div class="foldopen" id="foldopen00086" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00086" name="l00086"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#ab55f2a7fbbdae6798217f8253e4bacb3"> 86</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#ab55f2a7fbbdae6798217f8253e4bacb3">ternary_g</a>(</div>
|
||||
<div class="line"><a id="l00087" name="l00087"></a><span class="lineno"> 87</span> device <span class="keyword">const</span> <span class="keywordtype">bool</span>* a,</div>
|
||||
<div class="line"><a id="l00088" name="l00088"></a><span class="lineno"> 88</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00089" name="l00089"></a><span class="lineno"> 89</span> device <span class="keyword">const</span> T* c,</div>
|
||||
<div class="line"><a id="l00090" name="l00090"></a><span class="lineno"> 90</span> device T* d,</div>
|
||||
<div class="line"><a id="l00091" name="l00091"></a><span class="lineno"> 91</span> constant <span class="keyword">const</span> <span class="keywordtype">int</span>* shape,</div>
|
||||
<div class="line"><a id="l00092" name="l00092"></a><span class="lineno"> 92</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>* a_strides,</div>
|
||||
<div class="line"><a id="l00093" name="l00093"></a><span class="lineno"> 93</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>* b_strides,</div>
|
||||
<div class="line"><a id="l00094" name="l00094"></a><span class="lineno"> 94</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>* c_strides,</div>
|
||||
<div class="line"><a id="l00095" name="l00095"></a><span class="lineno"> 95</span> constant <span class="keyword">const</span> <span class="keywordtype">int</span>& ndim,</div>
|
||||
<div class="line"><a id="l00096" name="l00096"></a><span class="lineno"> 96</span> uint3 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00097" name="l00097"></a><span class="lineno"> 97</span> uint3 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00098" name="l00098"></a><span class="lineno"> 98</span> <span class="keyword">auto</span> idx =</div>
|
||||
<div class="line"><a id="l00099" name="l00099"></a><span class="lineno"> 99</span> <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a59d3221f4fbcc7e340af0a743fae054b">elem_to_loc_3_nd</a>(index, shape, a_strides, b_strides, c_strides, ndim);</div>
|
||||
<div class="line"><a id="l00100" name="l00100"></a><span class="lineno"> 100</span> <span class="keywordtype">size_t</span> out_idx = index.x + grid_dim.x * (index.y + grid_dim.y * index.z);</div>
|
||||
<div class="line"><a id="l00101" name="l00101"></a><span class="lineno"> 101</span> d[out_idx] = Op()(a[idx.x], b[idx.y], c[idx.z]);</div>
|
||||
<div class="line"><a id="l00102" name="l00102"></a><span class="lineno"> 102</span>}</div>
|
||||
<div class="line"><a id="l00077" name="l00077"></a><span class="lineno"> 77</span> </div>
|
||||
<div class="line"><a id="l00078" name="l00078"></a><span class="lineno"> 78</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> Op, <span class="keywordtype">int</span> DIM></div>
|
||||
<div class="foldopen" id="foldopen00079" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00079" name="l00079"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#a94f4695f471837525e5ca995285bc6aa"> 79</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#a94f4695f471837525e5ca995285bc6aa">ternary_g_nd</a>(</div>
|
||||
<div class="line"><a id="l00080" name="l00080"></a><span class="lineno"> 80</span> device <span class="keyword">const</span> <span class="keywordtype">bool</span>* a,</div>
|
||||
<div class="line"><a id="l00081" name="l00081"></a><span class="lineno"> 81</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00082" name="l00082"></a><span class="lineno"> 82</span> device <span class="keyword">const</span> T* c,</div>
|
||||
<div class="line"><a id="l00083" name="l00083"></a><span class="lineno"> 83</span> device T* d,</div>
|
||||
<div class="line"><a id="l00084" name="l00084"></a><span class="lineno"> 84</span> constant <span class="keyword">const</span> <span class="keywordtype">int</span> shape[DIM],</div>
|
||||
<div class="line"><a id="l00085" name="l00085"></a><span class="lineno"> 85</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> a_strides[DIM],</div>
|
||||
<div class="line"><a id="l00086" name="l00086"></a><span class="lineno"> 86</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> b_strides[DIM],</div>
|
||||
<div class="line"><a id="l00087" name="l00087"></a><span class="lineno"> 87</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span> c_strides[DIM],</div>
|
||||
<div class="line"><a id="l00088" name="l00088"></a><span class="lineno"> 88</span> uint3 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00089" name="l00089"></a><span class="lineno"> 89</span> uint3 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00090" name="l00090"></a><span class="lineno"> 90</span> <span class="keyword">auto</span> idx =</div>
|
||||
<div class="line"><a id="l00091" name="l00091"></a><span class="lineno"> 91</span> elem_to_loc_3_nd<DIM>(index, shape, a_strides, b_strides, c_strides);</div>
|
||||
<div class="line"><a id="l00092" name="l00092"></a><span class="lineno"> 92</span> <span class="keywordtype">size_t</span> out_idx =</div>
|
||||
<div class="line"><a id="l00093" name="l00093"></a><span class="lineno"> 93</span> index.x + (size_t)grid_dim.x * (index.y + (<span class="keywordtype">size_t</span>)grid_dim.y * index.z);</div>
|
||||
<div class="line"><a id="l00094" name="l00094"></a><span class="lineno"> 94</span> d[out_idx] = Op()(a[idx.x], b[idx.y], c[idx.z]);</div>
|
||||
<div class="line"><a id="l00095" name="l00095"></a><span class="lineno"> 95</span>}</div>
|
||||
</div>
|
||||
<div class="line"><a id="l00096" name="l00096"></a><span class="lineno"> 96</span> </div>
|
||||
<div class="line"><a id="l00097" name="l00097"></a><span class="lineno"> 97</span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> Op></div>
|
||||
<div class="foldopen" id="foldopen00098" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00098" name="l00098"></a><span class="lineno"><a class="line" href="metal_2kernels_2ternary_8h.html#ab55f2a7fbbdae6798217f8253e4bacb3"> 98</a></span>[[kernel]] <span class="keywordtype">void</span> <a class="code hl_function" href="metal_2kernels_2ternary_8h.html#ab55f2a7fbbdae6798217f8253e4bacb3">ternary_g</a>(</div>
|
||||
<div class="line"><a id="l00099" name="l00099"></a><span class="lineno"> 99</span> device <span class="keyword">const</span> <span class="keywordtype">bool</span>* a,</div>
|
||||
<div class="line"><a id="l00100" name="l00100"></a><span class="lineno"> 100</span> device <span class="keyword">const</span> T* b,</div>
|
||||
<div class="line"><a id="l00101" name="l00101"></a><span class="lineno"> 101</span> device <span class="keyword">const</span> T* c,</div>
|
||||
<div class="line"><a id="l00102" name="l00102"></a><span class="lineno"> 102</span> device T* d,</div>
|
||||
<div class="line"><a id="l00103" name="l00103"></a><span class="lineno"> 103</span> constant <span class="keyword">const</span> <span class="keywordtype">int</span>* shape,</div>
|
||||
<div class="line"><a id="l00104" name="l00104"></a><span class="lineno"> 104</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>* a_strides,</div>
|
||||
<div class="line"><a id="l00105" name="l00105"></a><span class="lineno"> 105</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>* b_strides,</div>
|
||||
<div class="line"><a id="l00106" name="l00106"></a><span class="lineno"> 106</span> constant <span class="keyword">const</span> <span class="keywordtype">size_t</span>* c_strides,</div>
|
||||
<div class="line"><a id="l00107" name="l00107"></a><span class="lineno"> 107</span> constant <span class="keyword">const</span> <span class="keywordtype">int</span>& ndim,</div>
|
||||
<div class="line"><a id="l00108" name="l00108"></a><span class="lineno"> 108</span> uint3 index [[thread_position_in_grid]],</div>
|
||||
<div class="line"><a id="l00109" name="l00109"></a><span class="lineno"> 109</span> uint3 grid_dim [[threads_per_grid]]) {</div>
|
||||
<div class="line"><a id="l00110" name="l00110"></a><span class="lineno"> 110</span> <span class="keyword">auto</span> idx =</div>
|
||||
<div class="line"><a id="l00111" name="l00111"></a><span class="lineno"> 111</span> <a class="code hl_function" href="backend_2metal_2kernels_2utils_8h.html#a59d3221f4fbcc7e340af0a743fae054b">elem_to_loc_3_nd</a>(index, shape, a_strides, b_strides, c_strides, ndim);</div>
|
||||
<div class="line"><a id="l00112" name="l00112"></a><span class="lineno"> 112</span> <span class="keywordtype">size_t</span> out_idx = index.x + grid_dim.x * (index.y + grid_dim.y * index.z);</div>
|
||||
<div class="line"><a id="l00113" name="l00113"></a><span class="lineno"> 113</span> d[out_idx] = Op()(a[idx.x], b[idx.y], c[idx.z]);</div>
|
||||
<div class="line"><a id="l00114" name="l00114"></a><span class="lineno"> 114</span>}</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:123</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:135</div></div>
|
||||
<div class="ttc" id="abackend_2metal_2kernels_2utils_8h_html_a59d3221f4fbcc7e340af0a743fae054b"><div class="ttname"><a href="backend_2metal_2kernels_2utils_8h.html#a59d3221f4fbcc7e340af0a743fae054b">elem_to_loc_3_nd</a></div><div class="ttdeci">METAL_FUNC uint3 elem_to_loc_3_nd(uint3 elem, constant const int *shape, constant const size_t *a_strides, constant const size_t *b_strides, constant const size_t *c_strides, int ndim)</div><div class="ttdef"><b>Definition</b> utils.h:220</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:129</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_a113df0c8a841b0e986900d580644e047"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#a113df0c8a841b0e986900d580644e047">ternary_g_nd3</a></div><div class="ttdeci">void ternary_g_nd3(device const bool *a, device const T *b, device const T *c, device T *d, constant const size_t a_strides[3], constant const size_t b_strides[3], constant const size_t c_strides[3], uint3 index, uint3 grid_dim)</div><div class="ttdef"><b>Definition</b> ternary.h:48</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_a1bd5918559850f3f80e3adee2391fe6a"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#a1bd5918559850f3f80e3adee2391fe6a">ternary_g_nd1</a></div><div class="ttdeci">void ternary_g_nd1(device const bool *a, device const T *b, device const T *c, device T *d, constant const size_t &a_strides, constant const size_t &b_strides, constant const size_t &c_strides, uint index)</div><div class="ttdef"><b>Definition</b> ternary.h:14</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_a113df0c8a841b0e986900d580644e047"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#a113df0c8a841b0e986900d580644e047">ternary_g_nd3</a></div><div class="ttdeci">void ternary_g_nd3(device const bool *a, device const T *b, device const T *c, device T *d, constant const size_t a_strides[3], constant const size_t b_strides[3], constant const size_t c_strides[3], uint3 index, uint3 grid_dim)</div><div class="ttdef"><b>Definition</b> ternary.h:60</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_a1bd5918559850f3f80e3adee2391fe6a"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#a1bd5918559850f3f80e3adee2391fe6a">ternary_g_nd1</a></div><div class="ttdeci">void ternary_g_nd1(device const bool *a, device const T *b, device const T *c, device T *d, constant const size_t &a_strides, constant const size_t &b_strides, constant const size_t &c_strides, uint index)</div><div class="ttdef"><b>Definition</b> ternary.h:26</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_a3e610f3b01966bdbf23fdfebe5d2c508"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#a3e610f3b01966bdbf23fdfebe5d2c508">ternary_v2</a></div><div class="ttdeci">void ternary_v2(device const bool *a, device const T *b, device const T *c, device T *d, uint2 index, uint2 grid_dim)</div><div class="ttdef"><b>Definition</b> ternary.h:14</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_a83f93644d21ee774e06e8190d0725ccb"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#a83f93644d21ee774e06e8190d0725ccb">ternary_v</a></div><div class="ttdeci">void ternary_v(device const bool *a, device const T *b, device const T *c, device T *d, uint index)</div><div class="ttdef"><b>Definition</b> ternary.h:4</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_a94f4695f471837525e5ca995285bc6aa"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#a94f4695f471837525e5ca995285bc6aa">ternary_g_nd</a></div><div class="ttdeci">void ternary_g_nd(device const bool *a, device const T *b, device const T *c, device T *d, constant const int shape[DIM], constant const size_t a_strides[DIM], constant const size_t b_strides[DIM], constant const size_t c_strides[DIM], uint3 index, uint3 grid_dim)</div><div class="ttdef"><b>Definition</b> ternary.h:67</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_ab55f2a7fbbdae6798217f8253e4bacb3"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#ab55f2a7fbbdae6798217f8253e4bacb3">ternary_g</a></div><div class="ttdeci">void ternary_g(device const bool *a, device const T *b, device const T *c, device T *d, constant const int *shape, constant const size_t *a_strides, constant const size_t *b_strides, constant const size_t *c_strides, constant const int &ndim, uint3 index, uint3 grid_dim)</div><div class="ttdef"><b>Definition</b> ternary.h:86</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_afdf0d9d0cb21fcb3f176500785076af8"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#afdf0d9d0cb21fcb3f176500785076af8">ternary_g_nd2</a></div><div class="ttdeci">void ternary_g_nd2(device const bool *a, device const T *b, device const T *c, device T *d, constant const size_t a_strides[2], constant const size_t b_strides[2], constant const size_t c_strides[2], uint2 index, uint2 grid_dim)</div><div class="ttdef"><b>Definition</b> ternary.h:30</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_a94f4695f471837525e5ca995285bc6aa"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#a94f4695f471837525e5ca995285bc6aa">ternary_g_nd</a></div><div class="ttdeci">void ternary_g_nd(device const bool *a, device const T *b, device const T *c, device T *d, constant const int shape[DIM], constant const size_t a_strides[DIM], constant const size_t b_strides[DIM], constant const size_t c_strides[DIM], uint3 index, uint3 grid_dim)</div><div class="ttdef"><b>Definition</b> ternary.h:79</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_ab55f2a7fbbdae6798217f8253e4bacb3"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#ab55f2a7fbbdae6798217f8253e4bacb3">ternary_g</a></div><div class="ttdeci">void ternary_g(device const bool *a, device const T *b, device const T *c, device T *d, constant const int *shape, constant const size_t *a_strides, constant const size_t *b_strides, constant const size_t *c_strides, constant const int &ndim, uint3 index, uint3 grid_dim)</div><div class="ttdef"><b>Definition</b> ternary.h:98</div></div>
|
||||
<div class="ttc" id="ametal_2kernels_2ternary_8h_html_afdf0d9d0cb21fcb3f176500785076af8"><div class="ttname"><a href="metal_2kernels_2ternary_8h.html#afdf0d9d0cb21fcb3f176500785076af8">ternary_g_nd2</a></div><div class="ttdeci">void ternary_g_nd2(device const bool *a, device const T *b, device const T *c, device T *d, constant const size_t a_strides[2], constant const size_t b_strides[2], constant const size_t c_strides[2], uint2 index, uint2 grid_dim)</div><div class="ttdef"><b>Definition</b> ternary.h:42</div></div>
|
||||
</div><!-- fragment --></div><!-- contents -->
|
||||
<!-- start footer part -->
|
||||
<hr class="footer"/><address class="footer"><small>
|
||||
|
Reference in New Issue
Block a user