/*
Stockfish, a UCI chess playing engine derived from Glaurung 2.1
- Copyright (C) 2004-2022 The Stockfish developers (see AUTHORS file)
+ Copyright (C) 2004-2024 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
#ifndef NNUE_FEATURE_TRANSFORMER_H_INCLUDED
#define NNUE_FEATURE_TRANSFORMER_H_INCLUDED
-#include "nnue_common.h"
-#include "nnue_architecture.h"
+#include <algorithm>
+#include <cassert>
+#include <cstdint>
+#include <cstring>
+#include <iosfwd>
+#include <utility>
-#include <cstring> // std::memset()
+#include "../position.h"
+#include "../types.h"
+#include "nnue_accumulator.h"
+#include "nnue_architecture.h"
+#include "nnue_common.h"
namespace Stockfish::Eval::NNUE {
- using BiasType = std::int16_t;
- using WeightType = std::int16_t;
- using PSQTWeightType = std::int32_t;
-
- // If vector instructions are enabled, we update and refresh the
- // accumulator tile by tile such that each tile fits in the CPU's
- // vector registers.
- #define VECTOR
-
- static_assert(PSQTBuckets % 8 == 0,
- "Per feature PSQT values cannot be processed at granularity lower than 8 at a time.");
-
- #ifdef USE_AVX512
- typedef __m512i vec_t;
- typedef __m256i psqt_vec_t;
- #define vec_load(a) _mm512_load_si512(a)
- #define vec_store(a,b) _mm512_store_si512(a,b)
- #define vec_add_16(a,b) _mm512_add_epi16(a,b)
- #define vec_sub_16(a,b) _mm512_sub_epi16(a,b)
- #define vec_load_psqt(a) _mm256_load_si256(a)
- #define vec_store_psqt(a,b) _mm256_store_si256(a,b)
- #define vec_add_psqt_32(a,b) _mm256_add_epi32(a,b)
- #define vec_sub_psqt_32(a,b) _mm256_sub_epi32(a,b)
- #define vec_zero_psqt() _mm256_setzero_si256()
- #define NumRegistersSIMD 32
-
- #elif USE_AVX2
- typedef __m256i vec_t;
- typedef __m256i psqt_vec_t;
- #define vec_load(a) _mm256_load_si256(a)
- #define vec_store(a,b) _mm256_store_si256(a,b)
- #define vec_add_16(a,b) _mm256_add_epi16(a,b)
- #define vec_sub_16(a,b) _mm256_sub_epi16(a,b)
- #define vec_load_psqt(a) _mm256_load_si256(a)
- #define vec_store_psqt(a,b) _mm256_store_si256(a,b)
- #define vec_add_psqt_32(a,b) _mm256_add_epi32(a,b)
- #define vec_sub_psqt_32(a,b) _mm256_sub_epi32(a,b)
- #define vec_zero_psqt() _mm256_setzero_si256()
- #define NumRegistersSIMD 16
-
- #elif USE_SSE2
- typedef __m128i vec_t;
- typedef __m128i psqt_vec_t;
- #define vec_load(a) (*(a))
- #define vec_store(a,b) *(a)=(b)
- #define vec_add_16(a,b) _mm_add_epi16(a,b)
- #define vec_sub_16(a,b) _mm_sub_epi16(a,b)
- #define vec_load_psqt(a) (*(a))
- #define vec_store_psqt(a,b) *(a)=(b)
- #define vec_add_psqt_32(a,b) _mm_add_epi32(a,b)
- #define vec_sub_psqt_32(a,b) _mm_sub_epi32(a,b)
- #define vec_zero_psqt() _mm_setzero_si128()
- #define NumRegistersSIMD (Is64Bit ? 16 : 8)
-
- #elif USE_MMX
- typedef __m64 vec_t;
- typedef __m64 psqt_vec_t;
- #define vec_load(a) (*(a))
- #define vec_store(a,b) *(a)=(b)
- #define vec_add_16(a,b) _mm_add_pi16(a,b)
- #define vec_sub_16(a,b) _mm_sub_pi16(a,b)
- #define vec_load_psqt(a) (*(a))
- #define vec_store_psqt(a,b) *(a)=(b)
- #define vec_add_psqt_32(a,b) _mm_add_pi32(a,b)
- #define vec_sub_psqt_32(a,b) _mm_sub_pi32(a,b)
- #define vec_zero_psqt() _mm_setzero_si64()
- #define NumRegistersSIMD 8
-
- #elif USE_NEON
- typedef int16x8_t vec_t;
- typedef int32x4_t psqt_vec_t;
- #define vec_load(a) (*(a))
- #define vec_store(a,b) *(a)=(b)
- #define vec_add_16(a,b) vaddq_s16(a,b)
- #define vec_sub_16(a,b) vsubq_s16(a,b)
- #define vec_load_psqt(a) (*(a))
- #define vec_store_psqt(a,b) *(a)=(b)
- #define vec_add_psqt_32(a,b) vaddq_s32(a,b)
- #define vec_sub_psqt_32(a,b) vsubq_s32(a,b)
- #define vec_zero_psqt() psqt_vec_t{0}
- #define NumRegistersSIMD 16
-
- #else
- #undef VECTOR
-
- #endif
-
-
- #ifdef VECTOR
-
- // Compute optimal SIMD register count for feature transformer accumulation.
-
- // We use __m* types as template arguments, which causes GCC to emit warnings
- // about losing some attribute information. This is irrelevant to us as we
- // only take their size, so the following pragma are harmless.
- #if defined(__GNUC__)
- #pragma GCC diagnostic push
- #pragma GCC diagnostic ignored "-Wignored-attributes"
- #endif
-
- template <typename SIMDRegisterType,
- typename LaneType,
- int NumLanes,
- int MaxRegisters>
- static constexpr int BestRegisterCount()
- {
- #define RegisterSize sizeof(SIMDRegisterType)
- #define LaneSize sizeof(LaneType)
-
- static_assert(RegisterSize >= LaneSize);
- static_assert(MaxRegisters <= NumRegistersSIMD);
- static_assert(MaxRegisters > 0);
- static_assert(NumRegistersSIMD > 0);
- static_assert(RegisterSize % LaneSize == 0);
- static_assert((NumLanes * LaneSize) % RegisterSize == 0);
-
- const int ideal = (NumLanes * LaneSize) / RegisterSize;
- if (ideal <= MaxRegisters)
- return ideal;
-
- // Look for the largest divisor of the ideal register count that is smaller than MaxRegisters
- for (int divisor = MaxRegisters; divisor > 1; --divisor)
- if (ideal % divisor == 0)
- return divisor;
-
- return 1;
- }
-
- static constexpr int NumRegs = BestRegisterCount<vec_t, WeightType, TransformedFeatureDimensions, NumRegistersSIMD>();
- static constexpr int NumPsqtRegs = BestRegisterCount<psqt_vec_t, PSQTWeightType, PSQTBuckets, NumRegistersSIMD>();
- #if defined(__GNUC__)
- #pragma GCC diagnostic pop
- #endif
- #endif
-
-
-
- // Input feature converter
- class FeatureTransformer {
+using BiasType = std::int16_t;
+using WeightType = std::int16_t;
+using PSQTWeightType = std::int32_t;
+
+// If vector instructions are enabled, we update and refresh the
+// accumulator tile by tile such that each tile fits in the CPU's
+// vector registers.
+#define VECTOR
+
+static_assert(PSQTBuckets % 8 == 0,
+ "Per feature PSQT values cannot be processed at granularity lower than 8 at a time.");
+
+#ifdef USE_AVX512
+using vec_t = __m512i;
+using psqt_vec_t = __m256i;
+ #define vec_load(a) _mm512_load_si512(a)
+ #define vec_store(a, b) _mm512_store_si512(a, b)
+ #define vec_add_16(a, b) _mm512_add_epi16(a, b)
+ #define vec_sub_16(a, b) _mm512_sub_epi16(a, b)
+ #define vec_mulhi_16(a, b) _mm512_mulhi_epi16(a, b)
+ #define vec_zero() _mm512_setzero_epi32()
+ #define vec_set_16(a) _mm512_set1_epi16(a)
+ #define vec_max_16(a, b) _mm512_max_epi16(a, b)
+ #define vec_min_16(a, b) _mm512_min_epi16(a, b)
+ #define vec_slli_16(a, b) _mm512_slli_epi16(a, b)
+ // Inverse permuted at load time
+ #define vec_packus_16(a, b) _mm512_packus_epi16(a, b)
+ #define vec_load_psqt(a) _mm256_load_si256(a)
+ #define vec_store_psqt(a, b) _mm256_store_si256(a, b)
+ #define vec_add_psqt_32(a, b) _mm256_add_epi32(a, b)
+ #define vec_sub_psqt_32(a, b) _mm256_sub_epi32(a, b)
+ #define vec_zero_psqt() _mm256_setzero_si256()
+ #define NumRegistersSIMD 16
+ #define MaxChunkSize 64
+
+#elif USE_AVX2
+using vec_t = __m256i;
+using psqt_vec_t = __m256i;
+ #define vec_load(a) _mm256_load_si256(a)
+ #define vec_store(a, b) _mm256_store_si256(a, b)
+ #define vec_add_16(a, b) _mm256_add_epi16(a, b)
+ #define vec_sub_16(a, b) _mm256_sub_epi16(a, b)
+ #define vec_mulhi_16(a, b) _mm256_mulhi_epi16(a, b)
+ #define vec_zero() _mm256_setzero_si256()
+ #define vec_set_16(a) _mm256_set1_epi16(a)
+ #define vec_max_16(a, b) _mm256_max_epi16(a, b)
+ #define vec_min_16(a, b) _mm256_min_epi16(a, b)
+ #define vec_slli_16(a, b) _mm256_slli_epi16(a, b)
+ // Inverse permuted at load time
+ #define vec_packus_16(a, b) _mm256_packus_epi16(a, b)
+ #define vec_load_psqt(a) _mm256_load_si256(a)
+ #define vec_store_psqt(a, b) _mm256_store_si256(a, b)
+ #define vec_add_psqt_32(a, b) _mm256_add_epi32(a, b)
+ #define vec_sub_psqt_32(a, b) _mm256_sub_epi32(a, b)
+ #define vec_zero_psqt() _mm256_setzero_si256()
+ #define NumRegistersSIMD 16
+ #define MaxChunkSize 32
+
+#elif USE_SSE2
+using vec_t = __m128i;
+using psqt_vec_t = __m128i;
+ #define vec_load(a) (*(a))
+ #define vec_store(a, b) *(a) = (b)
+ #define vec_add_16(a, b) _mm_add_epi16(a, b)
+ #define vec_sub_16(a, b) _mm_sub_epi16(a, b)
+ #define vec_mulhi_16(a, b) _mm_mulhi_epi16(a, b)
+ #define vec_zero() _mm_setzero_si128()
+ #define vec_set_16(a) _mm_set1_epi16(a)
+ #define vec_max_16(a, b) _mm_max_epi16(a, b)
+ #define vec_min_16(a, b) _mm_min_epi16(a, b)
+ #define vec_slli_16(a, b) _mm_slli_epi16(a, b)
+ #define vec_packus_16(a, b) _mm_packus_epi16(a, b)
+ #define vec_load_psqt(a) (*(a))
+ #define vec_store_psqt(a, b) *(a) = (b)
+ #define vec_add_psqt_32(a, b) _mm_add_epi32(a, b)
+ #define vec_sub_psqt_32(a, b) _mm_sub_epi32(a, b)
+ #define vec_zero_psqt() _mm_setzero_si128()
+ #define NumRegistersSIMD (Is64Bit ? 16 : 8)
+ #define MaxChunkSize 16
+
+#elif USE_NEON
+using vec_t = int16x8_t;
+using psqt_vec_t = int32x4_t;
+ #define vec_load(a) (*(a))
+ #define vec_store(a, b) *(a) = (b)
+ #define vec_add_16(a, b) vaddq_s16(a, b)
+ #define vec_sub_16(a, b) vsubq_s16(a, b)
+ #define vec_mulhi_16(a, b) vqdmulhq_s16(a, b)
+ #define vec_zero() \
+ vec_t { 0 }
+ #define vec_set_16(a) vdupq_n_s16(a)
+ #define vec_max_16(a, b) vmaxq_s16(a, b)
+ #define vec_min_16(a, b) vminq_s16(a, b)
+ #define vec_slli_16(a, b) vshlq_s16(a, vec_set_16(b))
+ #define vec_packus_16(a, b) reinterpret_cast<vec_t>(vcombine_u8(vqmovun_s16(a), vqmovun_s16(b)))
+ #define vec_load_psqt(a) (*(a))
+ #define vec_store_psqt(a, b) *(a) = (b)
+ #define vec_add_psqt_32(a, b) vaddq_s32(a, b)
+ #define vec_sub_psqt_32(a, b) vsubq_s32(a, b)
+ #define vec_zero_psqt() \
+ psqt_vec_t { 0 }
+ #define NumRegistersSIMD 16
+ #define MaxChunkSize 16
+
+#else
+ #undef VECTOR
+
+#endif
+
+
+#ifdef VECTOR
+
+ // Compute optimal SIMD register count for feature transformer accumulation.
+
+ // We use __m* types as template arguments, which causes GCC to emit warnings
+ // about losing some attribute information. This is irrelevant to us as we
+ // only take their size, so the following pragma are harmless.
+ #if defined(__GNUC__)
+ #pragma GCC diagnostic push
+ #pragma GCC diagnostic ignored "-Wignored-attributes"
+ #endif
+
+template<typename SIMDRegisterType, typename LaneType, int NumLanes, int MaxRegisters>
+static constexpr int BestRegisterCount() {
+ #define RegisterSize sizeof(SIMDRegisterType)
+ #define LaneSize sizeof(LaneType)
+
+ static_assert(RegisterSize >= LaneSize);
+ static_assert(MaxRegisters <= NumRegistersSIMD);
+ static_assert(MaxRegisters > 0);
+ static_assert(NumRegistersSIMD > 0);
+ static_assert(RegisterSize % LaneSize == 0);
+ static_assert((NumLanes * LaneSize) % RegisterSize == 0);
+
+ const int ideal = (NumLanes * LaneSize) / RegisterSize;
+ if (ideal <= MaxRegisters)
+ return ideal;
+
+ // Look for the largest divisor of the ideal register count that is smaller than MaxRegisters
+ for (int divisor = MaxRegisters; divisor > 1; --divisor)
+ if (ideal % divisor == 0)
+ return divisor;
+
+ return 1;
+}
+ #if defined(__GNUC__)
+ #pragma GCC diagnostic pop
+ #endif
+#endif
+
+
+// Input feature converter
+template<IndexType TransformedFeatureDimensions,
+ Accumulator<TransformedFeatureDimensions> StateInfo::*accPtr>
+class FeatureTransformer {
- private:
// Number of output dimensions for one side
static constexpr IndexType HalfDimensions = TransformedFeatureDimensions;
- #ifdef VECTOR
- static constexpr IndexType TileHeight = NumRegs * sizeof(vec_t) / 2;
+ private:
+#ifdef VECTOR
+ static constexpr int NumRegs =
+ BestRegisterCount<vec_t, WeightType, TransformedFeatureDimensions, NumRegistersSIMD>();
+ static constexpr int NumPsqtRegs =
+ BestRegisterCount<psqt_vec_t, PSQTWeightType, PSQTBuckets, NumRegistersSIMD>();
+
+ static constexpr IndexType TileHeight = NumRegs * sizeof(vec_t) / 2;
static constexpr IndexType PsqtTileHeight = NumPsqtRegs * sizeof(psqt_vec_t) / 4;
static_assert(HalfDimensions % TileHeight == 0, "TileHeight must divide HalfDimensions");
static_assert(PSQTBuckets % PsqtTileHeight == 0, "PsqtTileHeight must divide PSQTBuckets");
- #endif
+#endif
public:
// Output type
using OutputType = TransformedFeatureType;
// Number of input/output dimensions
- static constexpr IndexType InputDimensions = FeatureSet::Dimensions;
+ static constexpr IndexType InputDimensions = FeatureSet::Dimensions;
static constexpr IndexType OutputDimensions = HalfDimensions;
// Size of forward propagation buffer
- static constexpr std::size_t BufferSize =
- OutputDimensions * sizeof(OutputType);
+ static constexpr std::size_t BufferSize = OutputDimensions * sizeof(OutputType);
// Hash value embedded in the evaluation file
static constexpr std::uint32_t get_hash_value() {
- return FeatureSet::HashValue ^ (OutputDimensions * 2);
+ return FeatureSet::HashValue ^ (OutputDimensions * 2);
}
- // Read network parameters
- bool read_parameters(std::istream& stream) {
-
- read_little_endian<BiasType >(stream, biases , HalfDimensions );
- read_little_endian<WeightType >(stream, weights , HalfDimensions * InputDimensions);
- read_little_endian<PSQTWeightType>(stream, psqtWeights, PSQTBuckets * InputDimensions);
+ static constexpr void order_packs([[maybe_unused]] uint64_t* v) {
+#if defined(USE_AVX512) // _mm512_packs_epi16 ordering
+ uint64_t tmp0 = v[2], tmp1 = v[3];
+ v[2] = v[8], v[3] = v[9];
+ v[8] = v[4], v[9] = v[5];
+ v[4] = tmp0, v[5] = tmp1;
+ tmp0 = v[6], tmp1 = v[7];
+ v[6] = v[10], v[7] = v[11];
+ v[10] = v[12], v[11] = v[13];
+ v[12] = tmp0, v[13] = tmp1;
+#elif defined(USE_AVX2) // _mm256_packs_epi16 ordering
+ std::swap(v[2], v[4]);
+ std::swap(v[3], v[5]);
+#endif
+ }
- return !stream.fail();
+ static constexpr void inverse_order_packs([[maybe_unused]] uint64_t* v) {
+#if defined(USE_AVX512) // Inverse _mm512_packs_epi16 ordering
+ uint64_t tmp0 = v[2], tmp1 = v[3];
+ v[2] = v[4], v[3] = v[5];
+ v[4] = v[8], v[5] = v[9];
+ v[8] = tmp0, v[9] = tmp1;
+ tmp0 = v[6], tmp1 = v[7];
+ v[6] = v[12], v[7] = v[13];
+ v[12] = v[10], v[13] = v[11];
+ v[10] = tmp0, v[11] = tmp1;
+#elif defined(USE_AVX2) // Inverse _mm256_packs_epi16 ordering
+ std::swap(v[2], v[4]);
+ std::swap(v[3], v[5]);
+#endif
}
- // Write network parameters
- bool write_parameters(std::ostream& stream) const {
+ void permute_weights([[maybe_unused]] void (*order_fn)(uint64_t*)) const {
+#if defined(USE_AVX2)
+ #if defined(USE_AVX512)
+ constexpr IndexType di = 16;
+ #else
+ constexpr IndexType di = 8;
+ #endif
+ uint64_t* b = reinterpret_cast<uint64_t*>(const_cast<BiasType*>(&biases[0]));
+ for (IndexType i = 0; i < HalfDimensions * sizeof(BiasType) / sizeof(uint64_t); i += di)
+ order_fn(&b[i]);
- write_little_endian<BiasType >(stream, biases , HalfDimensions );
- write_little_endian<WeightType >(stream, weights , HalfDimensions * InputDimensions);
- write_little_endian<PSQTWeightType>(stream, psqtWeights, PSQTBuckets * InputDimensions);
+ for (IndexType j = 0; j < InputDimensions; ++j)
+ {
+ uint64_t* w =
+ reinterpret_cast<uint64_t*>(const_cast<WeightType*>(&weights[j * HalfDimensions]));
+ for (IndexType i = 0; i < HalfDimensions * sizeof(WeightType) / sizeof(uint64_t);
+ i += di)
+ order_fn(&w[i]);
+ }
+#endif
+ }
+
+ inline void scale_weights(bool read) const {
+ for (IndexType j = 0; j < InputDimensions; ++j)
+ {
+ WeightType* w = const_cast<WeightType*>(&weights[j * HalfDimensions]);
+ for (IndexType i = 0; i < HalfDimensions; ++i)
+ w[i] = read ? w[i] * 2 : w[i] / 2;
+ }
- return !stream.fail();
+ BiasType* b = const_cast<BiasType*>(biases);
+ for (IndexType i = 0; i < HalfDimensions; ++i)
+ b[i] = read ? b[i] * 2 : b[i] / 2;
}
- // Convert input features
- std::int32_t transform(const Position& pos, OutputType* output, int bucket) const {
- update_accumulator(pos, WHITE);
- update_accumulator(pos, BLACK);
+ // Read network parameters
+ bool read_parameters(std::istream& stream) {
- const Color perspectives[2] = {pos.side_to_move(), ~pos.side_to_move()};
- const auto& accumulation = pos.state()->accumulator.accumulation;
- const auto& psqtAccumulation = pos.state()->accumulator.psqtAccumulation;
+ read_leb_128<BiasType>(stream, biases, HalfDimensions);
+ read_leb_128<WeightType>(stream, weights, HalfDimensions * InputDimensions);
+ read_leb_128<PSQTWeightType>(stream, psqtWeights, PSQTBuckets * InputDimensions);
- const auto psqt = (
- psqtAccumulation[perspectives[0]][bucket]
- - psqtAccumulation[perspectives[1]][bucket]
- ) / 2;
+ permute_weights(inverse_order_packs);
+ scale_weights(true);
+ return !stream.fail();
+ }
+ // Write network parameters
+ bool write_parameters(std::ostream& stream) const {
- for (IndexType p = 0; p < 2; ++p)
- {
- const IndexType offset = (HalfDimensions / 2) * p;
+ permute_weights(order_packs);
+ scale_weights(false);
-#if defined(USE_AVX512)
+ write_leb_128<BiasType>(stream, biases, HalfDimensions);
+ write_leb_128<WeightType>(stream, weights, HalfDimensions * InputDimensions);
+ write_leb_128<PSQTWeightType>(stream, psqtWeights, PSQTBuckets * InputDimensions);
- constexpr IndexType OutputChunkSize = 512 / 8;
- static_assert((HalfDimensions / 2) % OutputChunkSize == 0);
- constexpr IndexType NumOutputChunks = HalfDimensions / 2 / OutputChunkSize;
+ permute_weights(inverse_order_packs);
+ scale_weights(true);
+ return !stream.fail();
+ }
- const __m512i Zero = _mm512_setzero_si512();
- const __m512i One = _mm512_set1_epi16(127);
- const __m512i Control = _mm512_setr_epi64(0, 2, 4, 6, 1, 3, 5, 7);
+ // Convert input features
+ std::int32_t transform(const Position& pos,
+ AccumulatorCaches::Cache<HalfDimensions>* cache,
+ OutputType* output,
+ int bucket) const {
+ update_accumulator<WHITE>(pos, cache);
+ update_accumulator<BLACK>(pos, cache);
+
+ const Color perspectives[2] = {pos.side_to_move(), ~pos.side_to_move()};
+ const auto& psqtAccumulation = (pos.state()->*accPtr).psqtAccumulation;
+ const auto psqt =
+ (psqtAccumulation[perspectives[0]][bucket] - psqtAccumulation[perspectives[1]][bucket])
+ / 2;
+
+ const auto& accumulation = (pos.state()->*accPtr).accumulation;
+
+ for (IndexType p = 0; p < 2; ++p)
+ {
+ const IndexType offset = (HalfDimensions / 2) * p;
+
+#if defined(VECTOR)
+
+ constexpr IndexType OutputChunkSize = MaxChunkSize;
+ static_assert((HalfDimensions / 2) % OutputChunkSize == 0);
+ constexpr IndexType NumOutputChunks = HalfDimensions / 2 / OutputChunkSize;
+
+ const vec_t Zero = vec_zero();
+ const vec_t One = vec_set_16(127 * 2);
+
+ const vec_t* in0 = reinterpret_cast<const vec_t*>(&(accumulation[perspectives[p]][0]));
+ const vec_t* in1 =
+ reinterpret_cast<const vec_t*>(&(accumulation[perspectives[p]][HalfDimensions / 2]));
+ vec_t* out = reinterpret_cast<vec_t*>(output + offset);
+
+ // Per the NNUE architecture, here we want to multiply pairs of
+ // clipped elements and divide the product by 128. To do this,
+ // we can naively perform min/max operation to clip each of the
+ // four int16 vectors, mullo pairs together, then pack them into
+ // one int8 vector. However, there exists a faster way.
+
+ // The idea here is to use the implicit clipping from packus to
+ // save us two vec_max_16 instructions. This clipping works due
+ // to the fact that any int16 integer below zero will be zeroed
+ // on packus.
+
+ // Consider the case where the second element is negative.
+ // If we do standard clipping, that element will be zero, which
+ // means our pairwise product is zero. If we perform packus and
+ // remove the lower-side clip for the second element, then our
+ // product before packus will be negative, and is zeroed on pack.
+ // The two operation produce equivalent results, but the second
+ // one (using packus) saves one max operation per pair.
+
+ // But here we run into a problem: mullo does not preserve the
+ // sign of the multiplication. We can get around this by doing
+ // mulhi, which keeps the sign. But that requires an additional
+ // tweak.
+
+ // mulhi cuts off the last 16 bits of the resulting product,
+ // which is the same as performing a rightward shift of 16 bits.
+ // We can use this to our advantage. Recall that we want to
+ // divide the final product by 128, which is equivalent to a
+ // 7-bit right shift. Intuitively, if we shift the clipped
+ // value left by 9, and perform mulhi, which shifts the product
+ // right by 16 bits, then we will net a right shift of 7 bits.
+ // However, this won't work as intended. Since we clip the
+ // values to have a maximum value of 127, shifting it by 9 bits
+ // might occupy the signed bit, resulting in some positive
+ // values being interpreted as negative after the shift.
+
+ // There is a way, however, to get around this limitation. When
+ // loading the network, scale accumulator weights and biases by
+ // 2. To get the same pairwise multiplication result as before,
+ // we need to divide the product by 128 * 2 * 2 = 512, which
+ // amounts to a right shift of 9 bits. So now we only have to
+ // shift left by 7 bits, perform mulhi (shifts right by 16 bits)
+ // and net a 9 bit right shift. Since we scaled everything by
+ // two, the values are clipped at 127 * 2 = 254, which occupies
+ // 8 bits. Shifting it by 7 bits left will no longer occupy the
+ // signed bit, so we are safe.
+
+ // Note that on NEON processors, we shift left by 6 instead
+ // because the instruction "vqdmulhq_s16" also doubles the
+ // return value after the multiplication, adding an extra shift
+ // to the left by 1, so we compensate by shifting less before
+ // the multiplication.
+
+ constexpr int shift =
+ #if defined(USE_SSE2)
+ 7;
+ #else
+ 6;
+ #endif
- const __m512i* in0 = reinterpret_cast<const __m512i*>(&(accumulation[perspectives[p]][0]));
- const __m512i* in1 = reinterpret_cast<const __m512i*>(&(accumulation[perspectives[p]][HalfDimensions / 2]));
- __m512i* out = reinterpret_cast< __m512i*>(output + offset);
+ for (IndexType j = 0; j < NumOutputChunks; ++j)
+ {
+ const vec_t sum0a =
+ vec_slli_16(vec_max_16(vec_min_16(in0[j * 2 + 0], One), Zero), shift);
+ const vec_t sum0b =
+ vec_slli_16(vec_max_16(vec_min_16(in0[j * 2 + 1], One), Zero), shift);
+ const vec_t sum1a = vec_min_16(in1[j * 2 + 0], One);
+ const vec_t sum1b = vec_min_16(in1[j * 2 + 1], One);
- for (IndexType j = 0; j < NumOutputChunks; j += 1)
- {
- const __m512i sum0a = _mm512_max_epi16(_mm512_min_epi16(in0[j * 2 + 0], One), Zero);
- const __m512i sum0b = _mm512_max_epi16(_mm512_min_epi16(in0[j * 2 + 1], One), Zero);
- const __m512i sum1a = _mm512_max_epi16(_mm512_min_epi16(in1[j * 2 + 0], One), Zero);
- const __m512i sum1b = _mm512_max_epi16(_mm512_min_epi16(in1[j * 2 + 1], One), Zero);
+ const vec_t pa = vec_mulhi_16(sum0a, sum1a);
+ const vec_t pb = vec_mulhi_16(sum0b, sum1b);
- const __m512i pa = _mm512_srli_epi16(_mm512_mullo_epi16(sum0a, sum1a), 7);
- const __m512i pb = _mm512_srli_epi16(_mm512_mullo_epi16(sum0b, sum1b), 7);
+ out[j] = vec_packus_16(pa, pb);
+ }
- out[j] = _mm512_permutexvar_epi64(Control, _mm512_packs_epi16(pa, pb));
- }
+#else
-#elif defined(USE_AVX2)
+ for (IndexType j = 0; j < HalfDimensions / 2; ++j)
+ {
+ BiasType sum0 = accumulation[static_cast<int>(perspectives[p])][j + 0];
+ BiasType sum1 =
+ accumulation[static_cast<int>(perspectives[p])][j + HalfDimensions / 2];
+ sum0 = std::clamp<BiasType>(sum0, 0, 127 * 2);
+ sum1 = std::clamp<BiasType>(sum1, 0, 127 * 2);
+ output[offset + j] = static_cast<OutputType>(unsigned(sum0 * sum1) / 512);
+ }
- constexpr IndexType OutputChunkSize = 256 / 8;
- static_assert((HalfDimensions / 2) % OutputChunkSize == 0);
- constexpr IndexType NumOutputChunks = HalfDimensions / 2 / OutputChunkSize;
+#endif
+ }
- const __m256i Zero = _mm256_setzero_si256();
- const __m256i One = _mm256_set1_epi16(127);
- constexpr int Control = 0b11011000;
+ return psqt;
+ } // end of function transform()
- const __m256i* in0 = reinterpret_cast<const __m256i*>(&(accumulation[perspectives[p]][0]));
- const __m256i* in1 = reinterpret_cast<const __m256i*>(&(accumulation[perspectives[p]][HalfDimensions / 2]));
- __m256i* out = reinterpret_cast< __m256i*>(output + offset);
+ void hint_common_access(const Position& pos,
+ AccumulatorCaches::Cache<HalfDimensions>* cache) const {
+ hint_common_access_for_perspective<WHITE>(pos, cache);
+ hint_common_access_for_perspective<BLACK>(pos, cache);
+ }
- for (IndexType j = 0; j < NumOutputChunks; j += 1)
- {
- const __m256i sum0a = _mm256_max_epi16(_mm256_min_epi16(in0[j * 2 + 0], One), Zero);
- const __m256i sum0b = _mm256_max_epi16(_mm256_min_epi16(in0[j * 2 + 1], One), Zero);
- const __m256i sum1a = _mm256_max_epi16(_mm256_min_epi16(in1[j * 2 + 0], One), Zero);
- const __m256i sum1b = _mm256_max_epi16(_mm256_min_epi16(in1[j * 2 + 1], One), Zero);
+ private:
+ template<Color Perspective>
+ StateInfo* try_find_computed_accumulator(const Position& pos) const {
+ // Look for a usable accumulator of an earlier position. We keep track
+ // of the estimated gain in terms of features to be added/subtracted.
+ StateInfo* st = pos.state();
+ int gain = FeatureSet::refresh_cost(pos);
+ while (st->previous && !(st->*accPtr).computed[Perspective])
+ {
+ // This governs when a full feature refresh is needed and how many
+ // updates are better than just one full refresh.
+ if (FeatureSet::requires_refresh(st, Perspective)
+ || (gain -= FeatureSet::update_cost(st) + 1) < 0)
+ break;
+ st = st->previous;
+ }
+ return st;
+ }
- const __m256i pa = _mm256_srli_epi16(_mm256_mullo_epi16(sum0a, sum1a), 7);
- const __m256i pb = _mm256_srli_epi16(_mm256_mullo_epi16(sum0b, sum1b), 7);
+ // It computes the accumulator of the next position, or updates the
+ // current position's accumulator if CurrentOnly is true.
+ template<Color Perspective, bool CurrentOnly>
+ void update_accumulator_incremental(const Position& pos, StateInfo* computed) const {
+ assert((computed->*accPtr).computed[Perspective]);
+ assert(computed->next != nullptr);
+
+#ifdef VECTOR
+ // Gcc-10.2 unnecessarily spills AVX2 registers if this array
+ // is defined in the VECTOR code below, once in each branch.
+ vec_t acc[NumRegs];
+ psqt_vec_t psqt[NumPsqtRegs];
+#endif
- out[j] = _mm256_permute4x64_epi64(_mm256_packs_epi16(pa, pb), Control);
- }
+ const Square ksq = pos.square<KING>(Perspective);
-#elif defined(USE_SSE2)
+ // The size must be enough to contain the largest possible update.
+ // That might depend on the feature set and generally relies on the
+ // feature set's update cost calculation to be correct and never allow
+ // updates with more added/removed features than MaxActiveDimensions.
+ FeatureSet::IndexList removed, added;
- constexpr IndexType OutputChunkSize = 128 / 8;
- static_assert((HalfDimensions / 2) % OutputChunkSize == 0);
- constexpr IndexType NumOutputChunks = HalfDimensions / 2 / OutputChunkSize;
+ if constexpr (CurrentOnly)
+ for (StateInfo* st = pos.state(); st != computed; st = st->previous)
+ FeatureSet::append_changed_indices<Perspective>(ksq, st->dirtyPiece, removed,
+ added);
+ else
+ FeatureSet::append_changed_indices<Perspective>(ksq, computed->next->dirtyPiece,
+ removed, added);
- const __m128i Zero = _mm_setzero_si128();
- const __m128i One = _mm_set1_epi16(127);
+ StateInfo* next = CurrentOnly ? pos.state() : computed->next;
+ assert(!(next->*accPtr).computed[Perspective]);
- const __m128i* in0 = reinterpret_cast<const __m128i*>(&(accumulation[perspectives[p]][0]));
- const __m128i* in1 = reinterpret_cast<const __m128i*>(&(accumulation[perspectives[p]][HalfDimensions / 2]));
- __m128i* out = reinterpret_cast< __m128i*>(output + offset);
+#ifdef VECTOR
+ if ((removed.size() == 1 || removed.size() == 2) && added.size() == 1)
+ {
+ auto accIn =
+ reinterpret_cast<const vec_t*>(&(computed->*accPtr).accumulation[Perspective][0]);
+ auto accOut = reinterpret_cast<vec_t*>(&(next->*accPtr).accumulation[Perspective][0]);
- for (IndexType j = 0; j < NumOutputChunks; j += 1)
- {
- const __m128i sum0a = _mm_max_epi16(_mm_min_epi16(in0[j * 2 + 0], One), Zero);
- const __m128i sum0b = _mm_max_epi16(_mm_min_epi16(in0[j * 2 + 1], One), Zero);
- const __m128i sum1a = _mm_max_epi16(_mm_min_epi16(in1[j * 2 + 0], One), Zero);
- const __m128i sum1b = _mm_max_epi16(_mm_min_epi16(in1[j * 2 + 1], One), Zero);
+ const IndexType offsetR0 = HalfDimensions * removed[0];
+ auto columnR0 = reinterpret_cast<const vec_t*>(&weights[offsetR0]);
+ const IndexType offsetA = HalfDimensions * added[0];
+ auto columnA = reinterpret_cast<const vec_t*>(&weights[offsetA]);
- const __m128i pa = _mm_srli_epi16(_mm_mullo_epi16(sum0a, sum1a), 7);
- const __m128i pb = _mm_srli_epi16(_mm_mullo_epi16(sum0b, sum1b), 7);
+ if (removed.size() == 1)
+ {
+ for (IndexType i = 0; i < HalfDimensions * sizeof(WeightType) / sizeof(vec_t); ++i)
+ accOut[i] = vec_add_16(vec_sub_16(accIn[i], columnR0[i]), columnA[i]);
+ }
+ else
+ {
+ const IndexType offsetR1 = HalfDimensions * removed[1];
+ auto columnR1 = reinterpret_cast<const vec_t*>(&weights[offsetR1]);
- out[j] = _mm_packs_epi16(pa, pb);
- }
+ for (IndexType i = 0; i < HalfDimensions * sizeof(WeightType) / sizeof(vec_t); ++i)
+ accOut[i] = vec_sub_16(vec_add_16(accIn[i], columnA[i]),
+ vec_add_16(columnR0[i], columnR1[i]));
+ }
-#elif defined(USE_NEON)
+ auto accPsqtIn = reinterpret_cast<const psqt_vec_t*>(
+ &(computed->*accPtr).psqtAccumulation[Perspective][0]);
+ auto accPsqtOut =
+ reinterpret_cast<psqt_vec_t*>(&(next->*accPtr).psqtAccumulation[Perspective][0]);
- constexpr IndexType OutputChunkSize = 128 / 8;
- static_assert((HalfDimensions / 2) % OutputChunkSize == 0);
- constexpr IndexType NumOutputChunks = HalfDimensions / 2 / OutputChunkSize;
+ const IndexType offsetPsqtR0 = PSQTBuckets * removed[0];
+ auto columnPsqtR0 = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offsetPsqtR0]);
+ const IndexType offsetPsqtA = PSQTBuckets * added[0];
+ auto columnPsqtA = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offsetPsqtA]);
- const int16x8_t Zero = vdupq_n_s16(0);
- const int16x8_t One = vdupq_n_s16(127);
+ if (removed.size() == 1)
+ {
+ for (std::size_t i = 0;
+ i < PSQTBuckets * sizeof(PSQTWeightType) / sizeof(psqt_vec_t); ++i)
+ accPsqtOut[i] = vec_add_psqt_32(vec_sub_psqt_32(accPsqtIn[i], columnPsqtR0[i]),
+ columnPsqtA[i]);
+ }
+ else
+ {
+ const IndexType offsetPsqtR1 = PSQTBuckets * removed[1];
+ auto columnPsqtR1 = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offsetPsqtR1]);
+
+ for (std::size_t i = 0;
+ i < PSQTBuckets * sizeof(PSQTWeightType) / sizeof(psqt_vec_t); ++i)
+ accPsqtOut[i] =
+ vec_sub_psqt_32(vec_add_psqt_32(accPsqtIn[i], columnPsqtA[i]),
+ vec_add_psqt_32(columnPsqtR0[i], columnPsqtR1[i]));
+ }
+ }
+ else
+ {
+ for (IndexType i = 0; i < HalfDimensions / TileHeight; ++i)
+ {
+ // Load accumulator
+ auto accTileIn = reinterpret_cast<const vec_t*>(
+ &(computed->*accPtr).accumulation[Perspective][i * TileHeight]);
+ for (IndexType j = 0; j < NumRegs; ++j)
+ acc[j] = vec_load(&accTileIn[j]);
+
+ // Difference calculation for the deactivated features
+ for (const auto index : removed)
+ {
+ const IndexType offset = HalfDimensions * index + i * TileHeight;
+ auto column = reinterpret_cast<const vec_t*>(&weights[offset]);
+ for (IndexType j = 0; j < NumRegs; ++j)
+ acc[j] = vec_sub_16(acc[j], column[j]);
+ }
+
+ // Difference calculation for the activated features
+ for (const auto index : added)
+ {
+ const IndexType offset = HalfDimensions * index + i * TileHeight;
+ auto column = reinterpret_cast<const vec_t*>(&weights[offset]);
+ for (IndexType j = 0; j < NumRegs; ++j)
+ acc[j] = vec_add_16(acc[j], column[j]);
+ }
+
+ // Store accumulator
+ auto accTileOut = reinterpret_cast<vec_t*>(
+ &(next->*accPtr).accumulation[Perspective][i * TileHeight]);
+ for (IndexType j = 0; j < NumRegs; ++j)
+ vec_store(&accTileOut[j], acc[j]);
+ }
- const int16x8_t* in0 = reinterpret_cast<const int16x8_t*>(&(accumulation[perspectives[p]][0]));
- const int16x8_t* in1 = reinterpret_cast<const int16x8_t*>(&(accumulation[perspectives[p]][HalfDimensions / 2]));
- int8x16_t* out = reinterpret_cast< int8x16_t*>(output + offset);
+ for (IndexType i = 0; i < PSQTBuckets / PsqtTileHeight; ++i)
+ {
+ // Load accumulator
+ auto accTilePsqtIn = reinterpret_cast<const psqt_vec_t*>(
+ &(computed->*accPtr).psqtAccumulation[Perspective][i * PsqtTileHeight]);
+ for (std::size_t j = 0; j < NumPsqtRegs; ++j)
+ psqt[j] = vec_load_psqt(&accTilePsqtIn[j]);
+
+ // Difference calculation for the deactivated features
+ for (const auto index : removed)
+ {
+ const IndexType offset = PSQTBuckets * index + i * PsqtTileHeight;
+ auto columnPsqt = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offset]);
+ for (std::size_t j = 0; j < NumPsqtRegs; ++j)
+ psqt[j] = vec_sub_psqt_32(psqt[j], columnPsqt[j]);
+ }
+
+ // Difference calculation for the activated features
+ for (const auto index : added)
+ {
+ const IndexType offset = PSQTBuckets * index + i * PsqtTileHeight;
+ auto columnPsqt = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offset]);
+ for (std::size_t j = 0; j < NumPsqtRegs; ++j)
+ psqt[j] = vec_add_psqt_32(psqt[j], columnPsqt[j]);
+ }
+
+ // Store accumulator
+ auto accTilePsqtOut = reinterpret_cast<psqt_vec_t*>(
+ &(next->*accPtr).psqtAccumulation[Perspective][i * PsqtTileHeight]);
+ for (std::size_t j = 0; j < NumPsqtRegs; ++j)
+ vec_store_psqt(&accTilePsqtOut[j], psqt[j]);
+ }
+ }
+#else
+ std::memcpy((next->*accPtr).accumulation[Perspective],
+ (computed->*accPtr).accumulation[Perspective],
+ HalfDimensions * sizeof(BiasType));
+ std::memcpy((next->*accPtr).psqtAccumulation[Perspective],
+ (computed->*accPtr).psqtAccumulation[Perspective],
+ PSQTBuckets * sizeof(PSQTWeightType));
+
+ // Difference calculation for the deactivated features
+ for (const auto index : removed)
+ {
+ const IndexType offset = HalfDimensions * index;
+ for (IndexType i = 0; i < HalfDimensions; ++i)
+ (next->*accPtr).accumulation[Perspective][i] -= weights[offset + i];
- for (IndexType j = 0; j < NumOutputChunks; j += 1)
- {
- const int16x8_t sum0a = vmaxq_s16(vminq_s16(in0[j * 2 + 0], One), Zero);
- const int16x8_t sum0b = vmaxq_s16(vminq_s16(in0[j * 2 + 1], One), Zero);
- const int16x8_t sum1a = vmaxq_s16(vminq_s16(in1[j * 2 + 0], One), Zero);
- const int16x8_t sum1b = vmaxq_s16(vminq_s16(in1[j * 2 + 1], One), Zero);
+ for (std::size_t i = 0; i < PSQTBuckets; ++i)
+ (next->*accPtr).psqtAccumulation[Perspective][i] -=
+ psqtWeights[index * PSQTBuckets + i];
+ }
- const int8x8_t pa = vshrn_n_s16(vmulq_s16(sum0a, sum1a), 7);
- const int8x8_t pb = vshrn_n_s16(vmulq_s16(sum0b, sum1b), 7);
+ // Difference calculation for the activated features
+ for (const auto index : added)
+ {
+ const IndexType offset = HalfDimensions * index;
+ for (IndexType i = 0; i < HalfDimensions; ++i)
+ (next->*accPtr).accumulation[Perspective][i] += weights[offset + i];
- out[j] = vcombine_s8(pa, pb);
- }
+ for (std::size_t i = 0; i < PSQTBuckets; ++i)
+ (next->*accPtr).psqtAccumulation[Perspective][i] +=
+ psqtWeights[index * PSQTBuckets + i];
+ }
+#endif
-#else
+ (next->*accPtr).computed[Perspective] = true;
- for (IndexType j = 0; j < HalfDimensions / 2; ++j) {
- BiasType sum0 = accumulation[static_cast<int>(perspectives[p])][j + 0];
- BiasType sum1 = accumulation[static_cast<int>(perspectives[p])][j + HalfDimensions / 2];
- sum0 = std::max<int>(0, std::min<int>(127, sum0));
- sum1 = std::max<int>(0, std::min<int>(127, sum1));
- output[offset + j] = static_cast<OutputType>(sum0 * sum1 / 128);
- }
+ if (!CurrentOnly && next != pos.state())
+ update_accumulator_incremental<Perspective, false>(pos, next);
+ }
-#endif
- }
+ template<Color Perspective>
+ void update_accumulator_refresh_cache(const Position& pos,
+ AccumulatorCaches::Cache<HalfDimensions>* cache) const {
+ assert(cache != nullptr);
- return psqt;
+ Square ksq = pos.square<KING>(Perspective);
+ auto& entry = (*cache)[ksq][Perspective];
+ FeatureSet::IndexList removed, added;
- } // end of function transform()
+ for (Color c : {WHITE, BLACK})
+ {
+ for (PieceType pt = PAWN; pt <= KING; ++pt)
+ {
+ const Piece piece = make_piece(c, pt);
+ const Bitboard oldBB = entry.byColorBB[c] & entry.byTypeBB[pt];
+ const Bitboard newBB = pos.pieces(c, pt);
+ Bitboard toRemove = oldBB & ~newBB;
+ Bitboard toAdd = newBB & ~oldBB;
+
+ while (toRemove)
+ {
+ Square sq = pop_lsb(toRemove);
+ removed.push_back(FeatureSet::make_index<Perspective>(sq, piece, ksq));
+ }
+ while (toAdd)
+ {
+ Square sq = pop_lsb(toAdd);
+ added.push_back(FeatureSet::make_index<Perspective>(sq, piece, ksq));
+ }
+ }
+ }
+ auto& accumulator = pos.state()->*accPtr;
+ accumulator.computed[Perspective] = true;
+#ifdef VECTOR
+ vec_t acc[NumRegs];
+ psqt_vec_t psqt[NumPsqtRegs];
- private:
- void update_accumulator(const Position& pos, const Color perspective) const {
-
- // The size must be enough to contain the largest possible update.
- // That might depend on the feature set and generally relies on the
- // feature set's update cost calculation to be correct and never
- // allow updates with more added/removed features than MaxActiveDimensions.
-
- #ifdef VECTOR
- // Gcc-10.2 unnecessarily spills AVX2 registers if this array
- // is defined in the VECTOR code below, once in each branch
- vec_t acc[NumRegs];
- psqt_vec_t psqt[NumPsqtRegs];
- #endif
-
- // Look for a usable accumulator of an earlier position. We keep track
- // of the estimated gain in terms of features to be added/subtracted.
- StateInfo *st = pos.state(), *next = nullptr;
- int gain = FeatureSet::refresh_cost(pos);
- while (st->previous && !st->accumulator.computed[perspective])
- {
- // This governs when a full feature refresh is needed and how many
- // updates are better than just one full refresh.
- if ( FeatureSet::requires_refresh(st, perspective)
- || (gain -= FeatureSet::update_cost(st) + 1) < 0)
- break;
- next = st;
- st = st->previous;
- }
-
- if (st->accumulator.computed[perspective])
- {
- if (next == nullptr)
- return;
-
- // Update incrementally in two steps. First, we update the "next"
- // accumulator. Then, we update the current accumulator (pos.state()).
-
- // Gather all features to be updated.
- const Square ksq = pos.square<KING>(perspective);
- FeatureSet::IndexList removed[2], added[2];
- FeatureSet::append_changed_indices(
- ksq, next->dirtyPiece, perspective, removed[0], added[0]);
- for (StateInfo *st2 = pos.state(); st2 != next; st2 = st2->previous)
- FeatureSet::append_changed_indices(
- ksq, st2->dirtyPiece, perspective, removed[1], added[1]);
-
- // Mark the accumulators as computed.
- next->accumulator.computed[perspective] = true;
- pos.state()->accumulator.computed[perspective] = true;
-
- // Now update the accumulators listed in states_to_update[], where the last element is a sentinel.
- StateInfo *states_to_update[3] =
- { next, next == pos.state() ? nullptr : pos.state(), nullptr };
- #ifdef VECTOR
for (IndexType j = 0; j < HalfDimensions / TileHeight; ++j)
{
- // Load accumulator
- auto accTile = reinterpret_cast<vec_t*>(
- &st->accumulator.accumulation[perspective][j * TileHeight]);
- for (IndexType k = 0; k < NumRegs; ++k)
- acc[k] = vec_load(&accTile[k]);
-
- for (IndexType i = 0; states_to_update[i]; ++i)
- {
- // Difference calculation for the deactivated features
- for (const auto index : removed[i])
+ auto accTile =
+ reinterpret_cast<vec_t*>(&accumulator.accumulation[Perspective][j * TileHeight]);
+ auto entryTile = reinterpret_cast<vec_t*>(&entry.accumulation[j * TileHeight]);
+
+ for (IndexType k = 0; k < NumRegs; ++k)
+ acc[k] = entryTile[k];
+
+ int i = 0;
+ for (; i < int(std::min(removed.size(), added.size())); ++i)
{
- const IndexType offset = HalfDimensions * index + j * TileHeight;
- auto column = reinterpret_cast<const vec_t*>(&weights[offset]);
- for (IndexType k = 0; k < NumRegs; ++k)
- acc[k] = vec_sub_16(acc[k], column[k]);
+ IndexType indexR = removed[i];
+ const IndexType offsetR = HalfDimensions * indexR + j * TileHeight;
+ auto columnR = reinterpret_cast<const vec_t*>(&weights[offsetR]);
+ IndexType indexA = added[i];
+ const IndexType offsetA = HalfDimensions * indexA + j * TileHeight;
+ auto columnA = reinterpret_cast<const vec_t*>(&weights[offsetA]);
+
+ for (unsigned k = 0; k < NumRegs; ++k)
+ acc[k] = vec_add_16(acc[k], vec_sub_16(columnA[k], columnR[k]));
}
+ for (; i < int(removed.size()); ++i)
+ {
+ IndexType index = removed[i];
+ const IndexType offset = HalfDimensions * index + j * TileHeight;
+ auto column = reinterpret_cast<const vec_t*>(&weights[offset]);
- // Difference calculation for the activated features
- for (const auto index : added[i])
+ for (unsigned k = 0; k < NumRegs; ++k)
+ acc[k] = vec_sub_16(acc[k], column[k]);
+ }
+ for (; i < int(added.size()); ++i)
{
- const IndexType offset = HalfDimensions * index + j * TileHeight;
- auto column = reinterpret_cast<const vec_t*>(&weights[offset]);
- for (IndexType k = 0; k < NumRegs; ++k)
- acc[k] = vec_add_16(acc[k], column[k]);
+ IndexType index = added[i];
+ const IndexType offset = HalfDimensions * index + j * TileHeight;
+ auto column = reinterpret_cast<const vec_t*>(&weights[offset]);
+
+ for (unsigned k = 0; k < NumRegs; ++k)
+ acc[k] = vec_add_16(acc[k], column[k]);
}
- // Store accumulator
- accTile = reinterpret_cast<vec_t*>(
- &states_to_update[i]->accumulator.accumulation[perspective][j * TileHeight]);
- for (IndexType k = 0; k < NumRegs; ++k)
- vec_store(&accTile[k], acc[k]);
- }
+ for (IndexType k = 0; k < NumRegs; k++)
+ vec_store(&entryTile[k], acc[k]);
+ for (IndexType k = 0; k < NumRegs; k++)
+ vec_store(&accTile[k], acc[k]);
}
for (IndexType j = 0; j < PSQTBuckets / PsqtTileHeight; ++j)
{
- // Load accumulator
- auto accTilePsqt = reinterpret_cast<psqt_vec_t*>(
- &st->accumulator.psqtAccumulation[perspective][j * PsqtTileHeight]);
- for (std::size_t k = 0; k < NumPsqtRegs; ++k)
- psqt[k] = vec_load_psqt(&accTilePsqt[k]);
-
- for (IndexType i = 0; states_to_update[i]; ++i)
- {
- // Difference calculation for the deactivated features
- for (const auto index : removed[i])
+ auto accTilePsqt = reinterpret_cast<psqt_vec_t*>(
+ &accumulator.psqtAccumulation[Perspective][j * PsqtTileHeight]);
+ auto entryTilePsqt =
+ reinterpret_cast<psqt_vec_t*>(&entry.psqtAccumulation[j * PsqtTileHeight]);
+
+ for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+ psqt[k] = entryTilePsqt[k];
+
+ for (int i = 0; i < int(removed.size()); ++i)
{
- const IndexType offset = PSQTBuckets * index + j * PsqtTileHeight;
- auto columnPsqt = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offset]);
- for (std::size_t k = 0; k < NumPsqtRegs; ++k)
- psqt[k] = vec_sub_psqt_32(psqt[k], columnPsqt[k]);
- }
+ IndexType index = removed[i];
+ const IndexType offset = PSQTBuckets * index + j * PsqtTileHeight;
+ auto columnPsqt = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offset]);
- // Difference calculation for the activated features
- for (const auto index : added[i])
+ for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+ psqt[k] = vec_sub_psqt_32(psqt[k], columnPsqt[k]);
+ }
+ for (int i = 0; i < int(added.size()); ++i)
{
- const IndexType offset = PSQTBuckets * index + j * PsqtTileHeight;
- auto columnPsqt = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offset]);
- for (std::size_t k = 0; k < NumPsqtRegs; ++k)
- psqt[k] = vec_add_psqt_32(psqt[k], columnPsqt[k]);
+ IndexType index = added[i];
+ const IndexType offset = PSQTBuckets * index + j * PsqtTileHeight;
+ auto columnPsqt = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offset]);
+
+ for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+ psqt[k] = vec_add_psqt_32(psqt[k], columnPsqt[k]);
}
- // Store accumulator
- accTilePsqt = reinterpret_cast<psqt_vec_t*>(
- &states_to_update[i]->accumulator.psqtAccumulation[perspective][j * PsqtTileHeight]);
for (std::size_t k = 0; k < NumPsqtRegs; ++k)
- vec_store_psqt(&accTilePsqt[k], psqt[k]);
- }
+ vec_store_psqt(&entryTilePsqt[k], psqt[k]);
+ for (std::size_t k = 0; k < NumPsqtRegs; ++k)
+ vec_store_psqt(&accTilePsqt[k], psqt[k]);
}
- #else
- for (IndexType i = 0; states_to_update[i]; ++i)
- {
- std::memcpy(states_to_update[i]->accumulator.accumulation[perspective],
- st->accumulator.accumulation[perspective],
- HalfDimensions * sizeof(BiasType));
-
- for (std::size_t k = 0; k < PSQTBuckets; ++k)
- states_to_update[i]->accumulator.psqtAccumulation[perspective][k] = st->accumulator.psqtAccumulation[perspective][k];
-
- st = states_to_update[i];
+#else
- // Difference calculation for the deactivated features
- for (const auto index : removed[i])
- {
+ for (const auto index : removed)
+ {
const IndexType offset = HalfDimensions * index;
-
for (IndexType j = 0; j < HalfDimensions; ++j)
- st->accumulator.accumulation[perspective][j] -= weights[offset + j];
+ entry.accumulation[j] -= weights[offset + j];
for (std::size_t k = 0; k < PSQTBuckets; ++k)
- st->accumulator.psqtAccumulation[perspective][k] -= psqtWeights[index * PSQTBuckets + k];
- }
-
- // Difference calculation for the activated features
- for (const auto index : added[i])
- {
+ entry.psqtAccumulation[k] -= psqtWeights[index * PSQTBuckets + k];
+ }
+ for (const auto index : added)
+ {
const IndexType offset = HalfDimensions * index;
-
for (IndexType j = 0; j < HalfDimensions; ++j)
- st->accumulator.accumulation[perspective][j] += weights[offset + j];
+ entry.accumulation[j] += weights[offset + j];
for (std::size_t k = 0; k < PSQTBuckets; ++k)
- st->accumulator.psqtAccumulation[perspective][k] += psqtWeights[index * PSQTBuckets + k];
- }
- }
- #endif
- }
- else
- {
- // Refresh the accumulator
- auto& accumulator = pos.state()->accumulator;
- accumulator.computed[perspective] = true;
- FeatureSet::IndexList active;
- FeatureSet::append_active_indices(pos, perspective, active);
-
- #ifdef VECTOR
- for (IndexType j = 0; j < HalfDimensions / TileHeight; ++j)
- {
- auto biasesTile = reinterpret_cast<const vec_t*>(
- &biases[j * TileHeight]);
- for (IndexType k = 0; k < NumRegs; ++k)
- acc[k] = biasesTile[k];
-
- for (const auto index : active)
- {
- const IndexType offset = HalfDimensions * index + j * TileHeight;
- auto column = reinterpret_cast<const vec_t*>(&weights[offset]);
-
- for (unsigned k = 0; k < NumRegs; ++k)
- acc[k] = vec_add_16(acc[k], column[k]);
- }
-
- auto accTile = reinterpret_cast<vec_t*>(
- &accumulator.accumulation[perspective][j * TileHeight]);
- for (unsigned k = 0; k < NumRegs; k++)
- vec_store(&accTile[k], acc[k]);
+ entry.psqtAccumulation[k] += psqtWeights[index * PSQTBuckets + k];
}
- for (IndexType j = 0; j < PSQTBuckets / PsqtTileHeight; ++j)
- {
- for (std::size_t k = 0; k < NumPsqtRegs; ++k)
- psqt[k] = vec_zero_psqt();
+ // The accumulator of the refresh entry has been updated.
+ // Now copy its content to the actual accumulator we were refreshing.
- for (const auto index : active)
- {
- const IndexType offset = PSQTBuckets * index + j * PsqtTileHeight;
- auto columnPsqt = reinterpret_cast<const psqt_vec_t*>(&psqtWeights[offset]);
+ std::memcpy(accumulator.accumulation[Perspective], entry.accumulation,
+ sizeof(BiasType) * HalfDimensions);
- for (std::size_t k = 0; k < NumPsqtRegs; ++k)
- psqt[k] = vec_add_psqt_32(psqt[k], columnPsqt[k]);
- }
+ std::memcpy(accumulator.psqtAccumulation[Perspective], entry.psqtAccumulation,
+ sizeof(int32_t) * PSQTBuckets);
+#endif
- auto accTilePsqt = reinterpret_cast<psqt_vec_t*>(
- &accumulator.psqtAccumulation[perspective][j * PsqtTileHeight]);
- for (std::size_t k = 0; k < NumPsqtRegs; ++k)
- vec_store_psqt(&accTilePsqt[k], psqt[k]);
- }
+ for (Color c : {WHITE, BLACK})
+ entry.byColorBB[c] = pos.pieces(c);
- #else
- std::memcpy(accumulator.accumulation[perspective], biases,
- HalfDimensions * sizeof(BiasType));
+ for (PieceType pt = PAWN; pt <= KING; ++pt)
+ entry.byTypeBB[pt] = pos.pieces(pt);
+ }
- for (std::size_t k = 0; k < PSQTBuckets; ++k)
- accumulator.psqtAccumulation[perspective][k] = 0;
+ template<Color Perspective>
+ void hint_common_access_for_perspective(const Position& pos,
+ AccumulatorCaches::Cache<HalfDimensions>* cache) const {
- for (const auto index : active)
- {
- const IndexType offset = HalfDimensions * index;
+ // Works like update_accumulator, but performs less work.
+ // Updates ONLY the accumulator for pos.
- for (IndexType j = 0; j < HalfDimensions; ++j)
- accumulator.accumulation[perspective][j] += weights[offset + j];
+ // Look for a usable accumulator of an earlier position. We keep track
+ // of the estimated gain in terms of features to be added/subtracted.
+ // Fast early exit.
+ if ((pos.state()->*accPtr).computed[Perspective])
+ return;
- for (std::size_t k = 0; k < PSQTBuckets; ++k)
- accumulator.psqtAccumulation[perspective][k] += psqtWeights[index * PSQTBuckets + k];
- }
- #endif
- }
+ StateInfo* oldest = try_find_computed_accumulator<Perspective>(pos);
+
+ if ((oldest->*accPtr).computed[Perspective] && oldest != pos.state())
+ update_accumulator_incremental<Perspective, true>(pos, oldest);
+ else
+ update_accumulator_refresh_cache<Perspective>(pos, cache);
+ }
+
+ template<Color Perspective>
+ void update_accumulator(const Position& pos,
+ AccumulatorCaches::Cache<HalfDimensions>* cache) const {
- #if defined(USE_MMX)
- _mm_empty();
- #endif
+ StateInfo* oldest = try_find_computed_accumulator<Perspective>(pos);
+
+ if ((oldest->*accPtr).computed[Perspective] && oldest != pos.state())
+ // Start from the oldest computed accumulator, update all the
+ // accumulators up to the current position.
+ update_accumulator_incremental<Perspective, false>(pos, oldest);
+ else
+ update_accumulator_refresh_cache<Perspective>(pos, cache);
}
+ template<IndexType Size>
+ friend struct AccumulatorCaches::Cache;
+
alignas(CacheLineSize) BiasType biases[HalfDimensions];
alignas(CacheLineSize) WeightType weights[HalfDimensions * InputDimensions];
alignas(CacheLineSize) PSQTWeightType psqtWeights[InputDimensions * PSQTBuckets];
- };
+};
} // namespace Stockfish::Eval::NNUE
-#endif // #ifndef NNUE_FEATURE_TRANSFORMER_H_INCLUDED
+#endif // #ifndef NNUE_FEATURE_TRANSFORMER_H_INCLUDED