Doxygen for mul_mat kernels

at0m741 · at0m741 · commit 073579a12509 · 2025-05-31T15:44:53.000+02:00
diff --git a/docs/MatrixKernel_8hpp.html b/docs/MatrixKernel_8hpp.html
@@ -121,6 +121,7 @@
 <tr class="heading"><td colspan="2"><h2 class="groupheader"><a id="nested-classes" name="nested-classes"></a>
 Classes</h2></td></tr>
 <tr class="memitem:"><td class="memItemLeft" align="right" valign="top">class &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="classtensorium_1_1MatrixKernel.html">tensorium::MatrixKernel&lt; K &gt;</a></td></tr>
+<tr class="memdesc:"><td class="mdescLeft">&#160;</td><td class="mdescRight"><a class="el" href="classtensorium_1_1MatrixKernel.html" title="MatrixKernel provides specialized SIMD-accelerated matrix multiplication routines for statically-size...">MatrixKernel</a> provides specialized SIMD-accelerated matrix multiplication routines for statically-sized square matrices.  <a href="classtensorium_1_1MatrixKernel.html#details">More...</a><br /></td></tr>
 <tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
 </table><table class="memberdecls">
 <tr class="heading"><td colspan="2"><h2 class="groupheader"><a id="namespaces" name="namespaces"></a>
diff --git a/docs/MatrixKernel_8hpp_source.html b/docs/MatrixKernel_8hpp_source.html
diff --git a/docs/annotated.html b/docs/annotated.html
@@ -129,7 +129,7 @@
 <tr id="row_1_7_" class="odd"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1GemmKernelBig.html" target="_self">GemmKernelBig</a></td><td class="desc"></td></tr>
 <tr id="row_1_8_" class="even"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1GemmKernelBigger.html" target="_self">GemmKernelBigger</a></td><td class="desc"></td></tr>
 <tr id="row_1_9_" class="odd"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1Matrix.html" target="_self">Matrix</a></td><td class="desc">High-performance aligned matrix class with SIMD support </td></tr>
-<tr id="row_1_10_" class="even"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1MatrixKernel.html" target="_self">MatrixKernel</a></td><td class="desc"></td></tr>
+<tr id="row_1_10_" class="even"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1MatrixKernel.html" target="_self">MatrixKernel</a></td><td class="desc"><a class="el" href="classtensorium_1_1MatrixKernel.html" title="MatrixKernel provides specialized SIMD-accelerated matrix multiplication routines for statically-size...">MatrixKernel</a> provides specialized SIMD-accelerated matrix multiplication routines for statically-sized square matrices </td></tr>
 <tr id="row_1_11_" class="odd"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1Parser.html" target="_self">Parser</a></td><td class="desc"></td></tr>
 <tr id="row_1_12_" class="even"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1SpectalChebyshev.html" target="_self">SpectalChebyshev</a></td><td class="desc">Placeholder Chebyshev spectral method class </td></tr>
 <tr id="row_1_13_" class="odd"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1SpectralFFT.html" target="_self">SpectralFFT</a></td><td class="desc">Fast Fourier Transform (FFT) implementation using Cooley–Tukey algorithm </td></tr>
diff --git a/docs/classtensorium_1_1MatrixKernel.html b/docs/classtensorium_1_1MatrixKernel.html
diff --git a/docs/classtensorium_1_1MatrixKernel__coll__graph.map b/docs/classtensorium_1_1MatrixKernel__coll__graph.map
@@ -1,5 +1,5 @@
 <map id="tensorium::MatrixKernel&lt; K &gt;" name="tensorium::MatrixKernel&lt; K &gt;">
-<area shape="rect" id="Node000001" title=" " alt="" coords="5,460,215,721"/>
+<area shape="rect" id="Node000001" title="MatrixKernel provides specialized SIMD&#45;accelerated matrix multiplication routines for statically&#45;size..." alt="" coords="5,460,215,721"/>
 <area shape="rect" id="Node000002" href="$classtensorium_1_1Matrix.html" title=" " alt="" coords="40,5,180,412"/>
 <area shape="poly" id="edge1_Node000001_Node000002" title=" " alt="" coords="113,427,113,460,107,460,107,427"/>
 </map>
diff --git a/docs/classtensorium_1_1MatrixKernel__coll__graph.md5 b/docs/classtensorium_1_1MatrixKernel__coll__graph.md5
@@ -1 +1 @@
-86b9ab31e179ca77aef479132fc3faab
+b127ba75a2eeed943e675c24eed60ddc
diff --git a/docs/classtensorium_1_1MatrixKernel__coll__graph.svg b/docs/classtensorium_1_1MatrixKernel__coll__graph.svg
@@ -11,7 +11,7 @@
 <!-- Node1 -->
 <g id="Node000001" class="node">
 <title>Node1</title>
-<g id="a_Node000001"><a xlink:title=" ">
+<g id="a_Node000001"><a xlink:title="MatrixKernel provides specialized SIMD&#45;accelerated matrix multiplication routines for statically&#45;size...">
 <polygon fill="#999999" stroke="none" points="157,-196 0,-196 0,0 157,0 157,-196"/>
 <text text-anchor="start" x="5" y="-182.5" font-family="Helvetica,sans-Serif" font-size="10.00">tensorium::MatrixKernel&lt; K &gt;</text>
 <text text-anchor="start" x="77" y="-166.5" font-family="Helvetica,sans-Serif" font-size="10.00"> </text>
diff --git a/docs/classtensorium_1_1MatrixKernel__inherit__graph.map b/docs/classtensorium_1_1MatrixKernel__inherit__graph.map
@@ -1,5 +1,5 @@
 <map id="tensorium::MatrixKernel&lt; K &gt;" name="tensorium::MatrixKernel&lt; K &gt;">
-<area shape="rect" id="Node000001" title=" " alt="" coords="5,460,215,721"/>
+<area shape="rect" id="Node000001" title="MatrixKernel provides specialized SIMD&#45;accelerated matrix multiplication routines for statically&#45;size..." alt="" coords="5,460,215,721"/>
 <area shape="rect" id="Node000002" href="$classtensorium_1_1Matrix.html" title=" " alt="" coords="40,5,180,412"/>
 <area shape="poly" id="edge1_Node000001_Node000002" title=" " alt="" coords="113,427,113,460,107,460,107,427"/>
 </map>
diff --git a/docs/classtensorium_1_1MatrixKernel__inherit__graph.md5 b/docs/classtensorium_1_1MatrixKernel__inherit__graph.md5
@@ -1 +1 @@
-86b9ab31e179ca77aef479132fc3faab
+b127ba75a2eeed943e675c24eed60ddc
diff --git a/docs/classtensorium_1_1MatrixKernel__inherit__graph.svg b/docs/classtensorium_1_1MatrixKernel__inherit__graph.svg
@@ -11,7 +11,7 @@
 <!-- Node1 -->
 <g id="Node000001" class="node">
 <title>Node1</title>
-<g id="a_Node000001"><a xlink:title=" ">
+<g id="a_Node000001"><a xlink:title="MatrixKernel provides specialized SIMD&#45;accelerated matrix multiplication routines for statically&#45;size...">
 <polygon fill="#999999" stroke="none" points="157,-196 0,-196 0,0 157,0 157,-196"/>
 <text text-anchor="start" x="5" y="-182.5" font-family="Helvetica,sans-Serif" font-size="10.00">tensorium::MatrixKernel&lt; K &gt;</text>
 <text text-anchor="start" x="77" y="-166.5" font-family="Helvetica,sans-Serif" font-size="10.00"> </text>
diff --git a/docs/hierarchy.html b/docs/hierarchy.html
@@ -129,7 +129,7 @@
 <tr id="row_19_" class="odd"><td class="entry"><span style="width:16px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classMathsUtils.html" target="_self">MathsUtils</a></td><td class="desc"></td></tr>
 <tr id="row_20_" class="even"><td class="entry"><span style="width:16px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1Matrix.html" target="_self">tensorium::Matrix&lt; K, RowMajor &gt;</a></td><td class="desc">High-performance aligned matrix class with SIMD support </td></tr>
 <tr id="row_21_" class="odd"><td class="entry"><span style="width:0px;display:inline-block;">&#160;</span><span id="arr_21_" class="arrow" onclick="dynsection.toggleFolder('21_')">&#9660;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1Matrix.html" target="_self">tensorium::Matrix&lt; K, true &gt;</a></td><td class="desc"></td></tr>
-<tr id="row_21_0_" class="even"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1MatrixKernel.html" target="_self">tensorium::MatrixKernel&lt; K &gt;</a></td><td class="desc"></td></tr>
+<tr id="row_21_0_" class="even"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1MatrixKernel.html" target="_self">tensorium::MatrixKernel&lt; K &gt;</a></td><td class="desc"><a class="el" href="classtensorium_1_1MatrixKernel.html" title="MatrixKernel provides specialized SIMD-accelerated matrix multiplication routines for statically-size...">MatrixKernel</a> provides specialized SIMD-accelerated matrix multiplication routines for statically-sized square matrices </td></tr>
 <tr id="row_22_" class="odd"><td class="entry"><span style="width:16px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium__RG_1_1Metric.html" target="_self">tensorium_RG::Metric&lt; T &gt;</a></td><td class="desc">A callable 4D metric class for general relativity (Minkowski, Schwarzschild, Kerr, etc.) </td></tr>
 <tr id="row_23_" class="even"><td class="entry"><span style="width:16px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1Parser.html" target="_self">tensorium::Parser</a></td><td class="desc"></td></tr>
 <tr id="row_24_" class="odd"><td class="entry"><span style="width:16px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="structAlignedAllocator_1_1rebind.html" target="_self">AlignedAllocator&lt; T, Alignment &gt;::rebind&lt; U &gt;</a></td><td class="desc">Rebinding structure for allocator traits </td></tr>
diff --git a/docs/inherit_graph_30.map b/docs/inherit_graph_30.map
@@ -1,5 +1,5 @@
 <map id="Graphical Class Hierarchy" name="Graphical Class Hierarchy">
 <area shape="rect" id="Node000000" href="$classtensorium_1_1Matrix.html" title=" " alt="" coords="5,5,154,50"/>
-<area shape="rect" id="Node000001" href="$classtensorium_1_1MatrixKernel.html" title=" " alt="" coords="202,14,419,42"/>
+<area shape="rect" id="Node000001" href="$classtensorium_1_1MatrixKernel.html" title="MatrixKernel provides specialized SIMD&#45;accelerated matrix multiplication routines for statically&#45;size..." alt="" coords="202,14,419,42"/>
 <area shape="poly" id="edge3_Node000000_Node000001" title=" " alt="" coords="169,25,201,25,201,30,169,30"/>
 </map>
diff --git a/docs/inherit_graph_30.md5 b/docs/inherit_graph_30.md5
@@ -1 +1 @@
-d4d94c662a8fc22a693f82e4b966e234
+ca866c6450c3ffcc8227091e19bf6478
diff --git a/docs/inherit_graph_30.svg b/docs/inherit_graph_30.svg
@@ -21,7 +21,7 @@
 <!-- Node1 -->
 <g id="Node000001" class="node">
 <title>Node1</title>
-<g id="a_Node000001"><a xlink:href="classtensorium_1_1MatrixKernel.html" target="_top" xlink:title=" ">
+<g id="a_Node000001"><a xlink:href="classtensorium_1_1MatrixKernel.html" target="_top" xlink:title="MatrixKernel provides specialized SIMD&#45;accelerated matrix multiplication routines for statically&#45;size...">
 <polygon fill="white" stroke="#666666" points="310.25,-27.12 147.25,-27.12 147.25,-6.38 310.25,-6.38 310.25,-27.12"/>
 <text text-anchor="middle" x="228.75" y="-13.62" font-family="Helvetica,sans-Serif" font-size="10.00">tensorium::MatrixKernel&lt; K &gt;</text>
 </a>
diff --git a/docs/namespaces.html b/docs/namespaces.html
@@ -131,7 +131,7 @@
 <tr id="row_2_8_" class="odd"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1GemmKernelBig.html" target="_self">GemmKernelBig</a></td><td class="desc"></td></tr>
 <tr id="row_2_9_" class="even"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1GemmKernelBigger.html" target="_self">GemmKernelBigger</a></td><td class="desc"></td></tr>
 <tr id="row_2_10_" class="odd"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1Matrix.html" target="_self">Matrix</a></td><td class="desc">High-performance aligned matrix class with SIMD support </td></tr>
-<tr id="row_2_11_" class="even"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1MatrixKernel.html" target="_self">MatrixKernel</a></td><td class="desc"></td></tr>
+<tr id="row_2_11_" class="even"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1MatrixKernel.html" target="_self">MatrixKernel</a></td><td class="desc"><a class="el" href="classtensorium_1_1MatrixKernel.html" title="MatrixKernel provides specialized SIMD-accelerated matrix multiplication routines for statically-size...">MatrixKernel</a> provides specialized SIMD-accelerated matrix multiplication routines for statically-sized square matrices </td></tr>
 <tr id="row_2_12_" class="odd"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1Parser.html" target="_self">Parser</a></td><td class="desc"></td></tr>
 <tr id="row_2_13_" class="even"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1SpectalChebyshev.html" target="_self">SpectalChebyshev</a></td><td class="desc">Placeholder Chebyshev spectral method class </td></tr>
 <tr id="row_2_14_" class="odd"><td class="entry"><span style="width:32px;display:inline-block;">&#160;</span><span class="icona"><span class="icon">C</span></span><a class="el" href="classtensorium_1_1SpectralFFT.html" target="_self">SpectralFFT</a></td><td class="desc">Fast Fourier Transform (FFT) implementation using Cooley–Tukey algorithm </td></tr>
diff --git a/docs/namespacetensorium.html b/docs/namespacetensorium.html
@@ -139,6 +139,7 @@
 <tr class="memdesc:"><td class="mdescLeft">&#160;</td><td class="mdescRight">High-performance aligned matrix class with SIMD support.  <a href="classtensorium_1_1Matrix.html#details">More...</a><br /></td></tr>
 <tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
 <tr class="memitem:"><td class="memItemLeft" align="right" valign="top">class &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="classtensorium_1_1MatrixKernel.html">MatrixKernel</a></td></tr>
+<tr class="memdesc:"><td class="mdescLeft">&#160;</td><td class="mdescRight"><a class="el" href="classtensorium_1_1MatrixKernel.html" title="MatrixKernel provides specialized SIMD-accelerated matrix multiplication routines for statically-size...">MatrixKernel</a> provides specialized SIMD-accelerated matrix multiplication routines for statically-sized square matrices.  <a href="classtensorium_1_1MatrixKernel.html#details">More...</a><br /></td></tr>
 <tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
 <tr class="memitem:"><td class="memItemLeft" align="right" valign="top">class &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="classtensorium_1_1Parser.html">Parser</a></td></tr>
 <tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
diff --git a/includes/Tensorium/Core/MatrixKernels/MatrixKernel.hpp b/includes/Tensorium/Core/MatrixKernels/MatrixKernel.hpp
@@ -2,7 +2,18 @@
 
 #include "../Matrix.hpp"
 
-namespace tensorium { 
+namespace tensorium {
+
+	/**
+	 * @brief MatrixKernel provides specialized SIMD-accelerated matrix multiplication routines 
+	 *        for statically-sized square matrices.
+	 * 
+	 * This class inherits from a column-major `Matrix<K, true>` and offers optimized
+	 * kernels for specific square sizes (2x2, 3x3, ..., 64x64), using AVX/SIMD intrinsics 
+	 * for high performance. The kernels exploit register-level blocking and FMADD chaining.
+	 * 
+	 * @tparam K Scalar type (float, double, etc.)
+	 */
 	template <typename K>
 		class MatrixKernel : public Matrix<K, true> {
 			public :
@@ -13,17 +24,30 @@ namespace tensorium {
 
 				using Simd = simd::SimdTraits<K, DefaultISA>;
 				using reg  = typename Simd::reg;
-
+				/** 
+				 * @brief Construct a MatrixKernel from a column-major matrix.
+				 * @param m Source matrix.
+				 */
 				MatrixKernel(const Matrix<K, true>& m) : Matrix<K, true>(m) {}
-
+				/** 
+				 * @brief Construct a MatrixKernel from a row-major matrix by copying elements.
+				 * @param m Source matrix.
+				 */
 				MatrixKernel(const Matrix<K, false>& m) : Matrix<K, true>(m.rows, m.cols) {
 					for (size_t i = 0; i < m.rows; ++i)
 						for (size_t j = 0; j < m.cols; ++j)
 							(*this)(i, j) = m(i, j);
 				}
 
+				/**
+				 * @brief Construct an empty column-major matrix kernel of size (r × c).
+				 */
 				MatrixKernel(size_t r, size_t c) : Matrix<K, true>(r, c) {}
-
+				/** 
+				 * @brief Multiply two 2×2 matrices using SIMD.
+				 * @param B Right-hand matrix.
+				 * @return Result of multiplication.
+				 */
 				inline Matrix<K> mul_mat2x2(const MatrixKernel<K>& B) const {
 					using Simd = simd::SimdTraits<K, DefaultISA>;
 					using reg = typename Simd::reg;
@@ -45,9 +69,11 @@ namespace tensorium {
 
 					return C;
 				}
-
-
-		
+				/** 
+				 * @brief Multiply two 3×3 matrices using SIMD.
+				 * @param B Right-hand matrix.
+				 * @return Result of multiplication.
+				 */
 				inline Matrix<K> mul_mat3x3(const MatrixKernel<K>& B) const {
 					using Simd = simd::SimdTraits<K, DefaultISA>;
 					using reg  = typename Simd::reg;
@@ -80,6 +106,11 @@ namespace tensorium {
 				}
 
 
+				/** 
+				 * @brief Multiply two 4×4 matrices using SIMD.
+				 * @param B Right-hand matrix.
+				 * @return Result of multiplication.
+				 */
 				inline Matrix<K> mul_mat4x4(const MatrixKernel<K>& B) const {
 					using Simd = simd::SimdTraits<K, DefaultISA>;
 					using reg  = typename Simd::reg;
@@ -154,6 +185,11 @@ namespace tensorium {
 					return result;
 				}
 
+				/** 
+				 * @brief Multiply two 8×8 matrices using SIMD.
+				 * @param B Right-hand matrix.
+				 * @return Result of multiplication.
+				 */
 				inline Matrix<K> mul_mat8x8(const MatrixKernel<K>& B) const {
 					using Simd = simd::SimdTraits<K, DefaultISA>;
 					using reg  = typename Simd::reg;
@@ -190,6 +226,12 @@ namespace tensorium {
 				}
 
 
+				/** 
+				 * @brief Multiply two 16×16 matrices using SIMD with FMADD accumulation.
+				 *        This function splits each row into two registers (low/high).
+				 * @param B Right-hand matrix.
+				 * @return Result of multiplication.
+				 */
 				inline Matrix<K> mul_mat16x16(const MatrixKernel<K>& B) const {
 					using Simd = simd::SimdTraits<K, DefaultISA>;
 					using reg  = typename Simd::reg;
@@ -220,20 +262,25 @@ namespace tensorium {
 
 					return result;
 				}
-
+				/** 
+				 * @brief Multiply two 32×32 matrices using SIMD.
+				 *        Each row is split into two registers (16 elements each).
+				 * @param B Right-hand matrix.
+				 * @return Result of multiplication.
+				 */
 				inline Matrix<K> mul_mat32x32(const MatrixKernel<K>& B) const {
 					using Simd = simd::SimdTraits<K, DefaultISA>;
 					using reg  = typename Simd::reg;
 
 					Matrix<K> result(32, 32);
 
 					reg brow[32][2];
-					#pragma unroll(2)
+#pragma unroll(2)
 					for (int k = 0; k < 32; ++k) {
 						brow[k][0] = Simd::loadu(&B.data[k * 32 + 0]);
 						brow[k][1] = Simd::loadu(&B.data[k * 32 + 16]);
 					}
-					#pragma unroll(2)
+#pragma unroll(2)
 					for (int i = 0; i < 32; ++i) {
 						const K* a = &data[i * 32];
 						reg acc0 = Simd::mul(Simd::set1(a[0]), brow[0][0]);
@@ -252,7 +299,13 @@ namespace tensorium {
 					return result;
 				}
 
-
+				/** 
+				 * @brief Multiply two 64×64 matrices using SIMD.
+				 *        Each row is split into 4 SIMD registers (4×16 elements).
+				 *        Vectorized FMADD chaining is used for performance.
+				 * @param B Right-hand matrix.
+				 * @return Result of multiplication.
+				 */
 				inline Matrix<K> mul_mat64x64(const MatrixKernel<K>& B) const {
 					using Simd = simd::SimdTraits<K, DefaultISA>;
 					using reg  = typename Simd::reg;

Original file line number	Diff line number	Diff line change
`@@ -1 +1 @@`
`1`		`-86b9ab31e179ca77aef479132fc3faab`
	`1`	`+b127ba75a2eeed943e675c24eed60ddc`