Remove redundant const qualifiers

[stockfish] / src / nnue / layers / affine_transform.h

diff --git a/src/nnue/layers/affine_transform.h b/src/nnue/layers/affine_transform.h
index 22451915ba1eb2f547c823c52c95c703d5ebdba1..363b4916e37b40e80bbb719d9483f881929d75d2 100644 (file)
--- a/src/nnue/layers/affine_transform.h
+++ b/src/nnue/layers/affine_transform.h
@@ -1,6 +1,6 @@
 /*
   Stockfish, a UCI chess playing engine derived from Glaurung 2.1
-  Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+  Copyright (C) 2004-2023 The Stockfish developers (see AUTHORS file)
 
   Stockfish is free software: you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
@@ -25,7 +25,7 @@
 #include <algorithm>
 #include <type_traits>
 #include "../nnue_common.h"
-#include "../../simd.h"
+#include "simd.h"
 
 /*
   This file contains the definition for a fully connected layer (aka affine transform).
@@ -151,9 +151,15 @@ namespace Stockfish::Eval::NNUE::Layers {
   template <IndexType InDims, IndexType OutDims, typename Enabled = void>
   class AffineTransform;
 
+#if defined (USE_AVX512)
+  constexpr IndexType LargeInputSize = 2 * 64;
+#else
+  constexpr IndexType LargeInputSize = std::numeric_limits<IndexType>::max();
+#endif
+
   // A specialization for large inputs.
   template <IndexType InDims, IndexType OutDims>
-  class AffineTransform<InDims, OutDims, std::enable_if_t<(ceil_to_multiple<IndexType>(InDims, MaxSimdWidth) >= 2*64)>> {
+  class AffineTransform<InDims, OutDims, std::enable_if_t<(ceil_to_multiple<IndexType>(InDims, MaxSimdWidth) >= LargeInputSize)>> {
    public:
     // Input/output type
     using InputType = std::uint8_t;
@@ -170,36 +176,36 @@ namespace Stockfish::Eval::NNUE::Layers {
 
     using OutputBuffer = OutputType[PaddedOutputDimensions];
 
-    static_assert(PaddedInputDimensions >= 128, "Something went wrong. This specialization should not have been chosen.");
+    static_assert(PaddedInputDimensions >= LargeInputSize, "Something went wrong. This specialization should not have been chosen.");
 
 #if defined (USE_AVX512)
-    static constexpr const IndexType InputSimdWidth = 64;
-    static constexpr const IndexType MaxNumOutputRegs = 16;
+    static constexpr IndexType InputSimdWidth = 64;
+    static constexpr IndexType MaxNumOutputRegs = 16;
 #elif defined (USE_AVX2)
-    static constexpr const IndexType InputSimdWidth = 32;
-    static constexpr const IndexType MaxNumOutputRegs = 8;
+    static constexpr IndexType InputSimdWidth = 32;
+    static constexpr IndexType MaxNumOutputRegs = 8;
 #elif defined (USE_SSSE3)
-    static constexpr const IndexType InputSimdWidth = 16;
-    static constexpr const IndexType MaxNumOutputRegs = 8;
+    static constexpr IndexType InputSimdWidth = 16;
+    static constexpr IndexType MaxNumOutputRegs = 8;
 #elif defined (USE_NEON)
-    static constexpr const IndexType InputSimdWidth = 8;
-    static constexpr const IndexType MaxNumOutputRegs = 8;
+    static constexpr IndexType InputSimdWidth = 8;
+    static constexpr IndexType MaxNumOutputRegs = 8;
 #else
     // The fallback implementation will not have permuted weights.
     // We define these to avoid a lot of ifdefs later.
-    static constexpr const IndexType InputSimdWidth = 1;
-    static constexpr const IndexType MaxNumOutputRegs = 1;
+    static constexpr IndexType InputSimdWidth = 1;
+    static constexpr IndexType MaxNumOutputRegs = 1;
 #endif
 
     // A big block is a region in the weight matrix of the size [PaddedInputDimensions, NumOutputRegs].
     // A small block is a region of size [InputSimdWidth, 1]
 
-    static constexpr const IndexType NumOutputRegs = std::min(MaxNumOutputRegs, OutputDimensions);
-    static constexpr const IndexType SmallBlockSize = InputSimdWidth;
-    static constexpr const IndexType BigBlockSize = NumOutputRegs * PaddedInputDimensions;
-    static constexpr const IndexType NumSmallBlocksInBigBlock = BigBlockSize / SmallBlockSize;
-    static constexpr const IndexType NumSmallBlocksPerOutput = PaddedInputDimensions / SmallBlockSize;
-    static constexpr const IndexType NumBigBlocks = OutputDimensions / NumOutputRegs;
+    static constexpr IndexType NumOutputRegs = std::min(MaxNumOutputRegs, OutputDimensions);
+    static constexpr IndexType SmallBlockSize = InputSimdWidth;
+    static constexpr IndexType BigBlockSize = NumOutputRegs * PaddedInputDimensions;
+    static constexpr IndexType NumSmallBlocksInBigBlock = BigBlockSize / SmallBlockSize;
+    static constexpr IndexType NumSmallBlocksPerOutput = PaddedInputDimensions / SmallBlockSize;
+    static constexpr IndexType NumBigBlocks = OutputDimensions / NumOutputRegs;
 
     static_assert(OutputDimensions % NumOutputRegs == 0);
 
@@ -235,10 +241,10 @@ namespace Stockfish::Eval::NNUE::Layers {
 
     // Read network parameters
     bool read_parameters(std::istream& stream) {
-      for (std::size_t i = 0; i < OutputDimensions; ++i)
+      for (IndexType i = 0; i < OutputDimensions; ++i)
         biases[i] = read_little_endian<BiasType>(stream);
 
-      for (std::size_t i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
+      for (IndexType i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
         weights[get_weight_index(i)] = read_little_endian<WeightType>(stream);
 
       return !stream.fail();
@@ -246,10 +252,10 @@ namespace Stockfish::Eval::NNUE::Layers {
 
     // Write network parameters
     bool write_parameters(std::ostream& stream) const {
-      for (std::size_t i = 0; i < OutputDimensions; ++i)
+      for (IndexType i = 0; i < OutputDimensions; ++i)
           write_little_endian<BiasType>(stream, biases[i]);
 
-      for (std::size_t i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
+      for (IndexType i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
         write_little_endian<WeightType>(stream, weights[get_weight_index(i)]);
 
       return !stream.fail();
@@ -369,7 +375,7 @@ namespace Stockfish::Eval::NNUE::Layers {
   };
 
   template <IndexType InDims, IndexType OutDims>
-  class AffineTransform<InDims, OutDims, std::enable_if_t<(ceil_to_multiple<IndexType>(InDims, MaxSimdWidth) < 2*64)>> {
+  class AffineTransform<InDims, OutDims, std::enable_if_t<(ceil_to_multiple<IndexType>(InDims, MaxSimdWidth) < LargeInputSize)>> {
    public:
     // Input/output type
     // Input/output type
@@ -387,11 +393,11 @@ namespace Stockfish::Eval::NNUE::Layers {
 
     using OutputBuffer = OutputType[PaddedOutputDimensions];
 
-    static_assert(PaddedInputDimensions < 128, "Something went wrong. This specialization should not have been chosen.");
+    static_assert(PaddedInputDimensions < LargeInputSize, "Something went wrong. This specialization should not have been chosen.");
 
 #if defined (USE_SSSE3)
-    static constexpr const IndexType OutputSimdWidth = SimdWidth / 4;
-    static constexpr const IndexType InputSimdWidth = SimdWidth;
+    static constexpr IndexType OutputSimdWidth = SimdWidth / 4;
+    static constexpr IndexType InputSimdWidth = SimdWidth;
 #endif
 
     // Hash value embedded in the evaluation file
@@ -422,9 +428,9 @@ namespace Stockfish::Eval::NNUE::Layers {
 
     // Read network parameters
     bool read_parameters(std::istream& stream) {
-      for (std::size_t i = 0; i < OutputDimensions; ++i)
+      for (IndexType i = 0; i < OutputDimensions; ++i)
         biases[i] = read_little_endian<BiasType>(stream);
-      for (std::size_t i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
+      for (IndexType i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
         weights[get_weight_index(i)] = read_little_endian<WeightType>(stream);
 
       return !stream.fail();
@@ -432,10 +438,10 @@ namespace Stockfish::Eval::NNUE::Layers {
 
     // Write network parameters
     bool write_parameters(std::ostream& stream) const {
-      for (std::size_t i = 0; i < OutputDimensions; ++i)
+      for (IndexType i = 0; i < OutputDimensions; ++i)
         write_little_endian<BiasType>(stream, biases[i]);
 
-      for (std::size_t i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
+      for (IndexType i = 0; i < OutputDimensions * PaddedInputDimensions; ++i)
         write_little_endian<WeightType>(stream, weights[get_weight_index(i)]);
 
       return !stream.fail();