<ahref="igemm__multiply__add_8h.html">Go to the documentation of this file.</a><divclass="fragment"><divclass="line"><aname="l00001"></a><spanclass="lineno"> 1</span> <spanclass="comment">/***************************************************************************************************</span></div><divclass="line"><aname="l00002"></a><spanclass="lineno"> 2</span> <spanclass="comment"> * Copyright (c) 2017-2018, NVIDIA CORPORATION. All rights reserved.</span></div><divclass="line"><aname="l00003"></a><spanclass="lineno"> 3</span> <spanclass="comment"> *</span></div><divclass="line"><aname="l00004"></a><spanclass="lineno"> 4</span> <spanclass="comment"> * Redistribution and use in source and binary forms, with or without modification, are permitted</span></div><divclass="line"><aname="l00005"></a><spanclass="lineno"> 5</span> <spanclass="comment"> * provided that the following conditions are met:</span></div><divclass="line"><aname="l00006"></a><spanclass="lineno"> 6</span> <spanclass="comment"> * * Redistributions of source code must retain the above copyright notice, this list of</span></div><divclass="line"><aname="l00007"></a><spanclass="lineno"> 7</span> <spanclass="comment"> * conditions and the following disclaimer.</span></div><divclass="line"><aname="l00008"></a><spanclass="lineno"> 8</span> <spanclass="comment"> * * Redistributions in binary form must reproduce the above copyright notice, this list of</span></div><divclass="line"><aname="l00009"></a><spanclass="lineno"> 9</span> <spanclass="comment"> * conditions and the following disclaimer in the documentation and/or other materials</span></div><divclass="line"><aname="l00010"></a><spanclass="lineno"> 10</span> <spanclass="comment"> * provided with the distribution.</span></div><divclass="line"><aname="l00011"></a><spanclass="lineno"> 11</span> <spanclass="comment"> * * Neither the name of the NVIDIA CORPORATION nor the names of its contributors may be used</span></div><divclass="line"><aname="l00012"></a><spanclass="lineno"> 12</span> <spanclass="comment"> * to endorse or promote products derived from this software without specific prior written</span></div><divclass="line"><aname="l00013"></a><spanclass="lineno"> 13</span> <spanclass="comment"> * permission.</span></div><divclass="line"><aname="l00014"></a><spanclass="lineno"> 14</span> <spanclass="comment"> *</span></div><divclass="line"><aname="l00015"></a><spanclass="lineno"> 15</span> <spanclass="comment"> * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR</span></div><divclass="line"><aname="l00016"></a><spanclass="lineno"> 16</span> <spanclass="comment"> * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND</span></div><divclass="line"><aname="l00017"></a><spanclass="lineno"> 17</span> <spanclass="comment"> * FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NVIDIA CORPORATION BE LIABLE</span></div><divclass="line"><aname="l00018"></a><spanclass="lineno"> 18</span> <spanclass="comment"> * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,</span></div><divclass="line"><aname="l00019"></a><spanclass="lineno"> 19</span> <spanclass="comment"> * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;</span></div><divclass="line"><aname="l00020"></a><spanclass="lineno"> 20</span> <spanclass="comment"> * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,</span></div><divclass="line"><aname="l00021"></a><spanclass="lineno"> 21</span> <spanclass="comment"> * STRICT LIABILITY, OR TOR (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE</span></div><divclass="line"><aname=
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_a8d0734b8e797576adcf89f70c62160d4"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#a8d0734b8e797576adcf89f70c62160d4">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::FragmentA</a></div><divclass="ttdeci">Fragment< ScalarA, AccumulatorsPerThread::kW *4 > FragmentA</div><divclass="ttdoc">The fragment for A. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:56</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_aad8a642f46c88e407a1150ee1d42b8dd"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#aad8a642f46c88e407a1150ee1d42b8dd">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::ThreadMultiplyAdd</a></div><divclass="ttdeci">CUTLASS_DEVICE ThreadMultiplyAdd()</div><divclass="ttdoc">Ctor. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:67</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_ad9d8e47e8896d8d4eab538aa78b56e47"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#ad9d8e47e8896d8d4eab538aa78b56e47">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::multiply_add</a></div><divclass="ttdeci">CUTLASS_DEVICE void multiply_add(FragmentA const &a, FragmentB const &b, Accumulators const &c, Accumulators &d)</div><divclass="ttdoc">Multiply : d = a*b + c. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:70</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_a703b329ebf14d78f576e83c5e6fe23a7"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#a703b329ebf14d78f576e83c5e6fe23a7">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::ScalarC</a></div><divclass="ttdeci">int ScalarC</div><divclass="ttdoc">The type for C and D. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:62</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_aa84c3d4efc7947d6efb75536c88043bd"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#aa84c3d4efc7947d6efb75536c88043bd">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::InstructionShape</a></div><divclass="ttdeci">Shape< 4, 1, 1 > InstructionShape</div><divclass="ttdoc">The shape of the instruction. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:44</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_a6bb1afd96da05370e61b38f2a93e40df"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#a6bb1afd96da05370e61b38f2a93e40df">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::ThreadsPerWarp</a></div><divclass="ttdeci">ThreadsPerWarp_ ThreadsPerWarp</div><divclass="ttdoc">The number of threads per warp. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:50</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_ac5cde71eb825b0a4311bd0ce982f47aa"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#ac5cde71eb825b0a4311bd0ce982f47aa">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::ThreadGemmShape</a></div><divclass="ttdeci">ThreadGemmShape_ ThreadGemmShape</div><divclass="ttdoc">Shape of the thread-level GEMM (K-by-N-by-M) </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:46</div></div>
<divclass="ttc"id="structcutlass_1_1Shape_html"><divclass="ttname"><ahref="structcutlass_1_1Shape.html">cutlass::Shape</a></div><divclass="ttdoc">A Shape implementing Layout Concept describing the dimensions of a cube. </div><divclass="ttdef"><b>Definition:</b> shape.h:64</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_ae3152470cbbba2310d9c83b9d5d43027"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#ae3152470cbbba2310d9c83b9d5d43027">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::AccumulatorsPerWarp</a></div><divclass="ttdeci">ShapeMul< ThreadGemmShape, ThreadsPerWarp >::Shape AccumulatorsPerWarp</div><divclass="ttdoc">The number of accumulators per warp. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:52</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_html"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd.html">cutlass::gemm::ThreadMultiplyAdd</a></div><divclass="ttdoc">Template performing matrix multiply-add operation within a thread. </div><divclass="ttdef"><b>Definition:</b> thread_multiply_add.h:44</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_aa88edf2e89062be00181f5dc4f4a0947"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#aa88edf2e89062be00181f5dc4f4a0947">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::AccumulatorsPerThread</a></div><divclass="ttdeci">ThreadGemmShape AccumulatorsPerThread</div><divclass="ttdoc">Aliased for compatibility. Will be removed in CUTLASS v2.0. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:48</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_a6439d8fc71727cc6d50f87eae549157e"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#a6439d8fc71727cc6d50f87eae549157e">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::FragmentB</a></div><divclass="ttdeci">Fragment< ScalarB, AccumulatorsPerThread::kH *4 > FragmentB</div><divclass="ttdoc">The fragment for B. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:60</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_a11be198f90afb859be51ec5feb5dcd2b"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#a11be198f90afb859be51ec5feb5dcd2b">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::ScalarA</a></div><divclass="ttdeci">int8_t ScalarA</div><divclass="ttdoc">The type for A. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:54</div></div>
<divclass="ttc"id="fragment_8h_html"><divclass="ttname"><ahref="fragment_8h.html">fragment.h</a></div><divclass="ttdoc">Defines Fragment, a statically-sized array for storing parts of matrices within a thread&#39;s registers...</div></div>
<divclass="ttc"id="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4_html_a6a9c4f906a4930f4fc415009ead2e05d"><divclass="ttname"><ahref="structcutlass_1_1gemm_1_1ThreadMultiplyAdd_3_01ThreadGemmShape___00_01ThreadsPerWarp___00_01int8__t_00_01int8__t_00_01int_01_4.html#a6a9c4f906a4930f4fc415009ead2e05d">cutlass::gemm::ThreadMultiplyAdd< ThreadGemmShape_, ThreadsPerWarp_, int8_t, int8_t, int >::ScalarB</a></div><divclass="ttdeci">int8_t ScalarB</div><divclass="ttdoc">The type for B. </div><divclass="ttdef"><b>Definition:</b> igemm_multiply_add.h:58</div></div>
