Add 'external/ircolib/' from commit 'ce3cd726c8df8388d554abf8bb55d55020eb4450'

git-subtree-dir: external/ircolib
git-subtree-mainline: e64eb40b38
git-subtree-split: ce3cd726c8

external/ircolib/README.md

@@ -0,0 +1,7 @@
+# ircolib
+
+Collections of useful functions I started copy-pasting in various projects and thus decided to gather all in one place.
+
+## Flags
+
+To enable SIMD for `floats.hpp`, add a `#define SIMD_SUPPORT` before `#include "floats.hpp"`.

external/ircolib/file.hpp

@@ -0,0 +1,43 @@
+#pragma once
+#include <types.hpp>
+#include <fstream>
+#include <vector>
+#include <filesystem>
+
+namespace fs = std::filesystem;
+
+namespace ircolib {
+static inline std::vector<u8> ReadFileBinary(const std::string &path) {
+  std::ifstream file(path, std::ios::binary);
+  return {std::istreambuf_iterator{file}, {}};
+}
+
+static inline void WriteFileBinary(const std::vector<u8> &data, const std::string &path) {
+  std::ofstream file(path, std::ios::binary);
+  std::copy(data.begin(), data.end(), std::ostreambuf_iterator{file});
+}
+
+static inline void WriteFileBinary(const u8 *data, const size_t size, const std::string &path) {
+  FILE *out = fopen(path.c_str(), "wb");
+  fwrite(data, size, 1, out);
+  fclose(out);
+}
+
+template <size_t Size>
+static inline void WriteFileBinary(const std::array<u8, Size> &data, const std::string &path) {
+  std::ofstream file(path, std::ios::binary);
+  std::copy(data.begin(), data.end(), std::ostreambuf_iterator{file});
+}
+
+static inline size_t NextPow2(size_t num) {
+  // Taken from "Bit Twiddling Hacks" by Sean Anderson:
+  // https://graphics.stanford.edu/~seander/bithacks.html#RoundUpPowerOf2
+  --num;
+  num |= num >> 1;
+  num |= num >> 2;
+  num |= num >> 4;
+  num |= num >> 8;
+  num |= num >> 16;
+  return num + 1;
+}
+} // namespace Util

external/ircolib/floats.hpp

@@ -0,0 +1,106 @@
+#pragma once
+#include <cmath>
+#include <types.hpp>
+
+namespace ircolib {
+static inline auto roundCeil(float f) {
+#ifdef SIMD_SUPPORT
+  __m128 t = _mm_set_ss(f);
+  t = _mm_round_ss(t, t, _MM_FROUND_TO_POS_INF);
+  return _mm_cvtss_f32(t);
+#else
+  return ceilf(f);
+#endif
+}
+
+static inline auto roundCeil(double f) {
+#ifdef SIMD_SUPPORT
+  __m128d t = _mm_set_sd(f);
+  t = _mm_round_sd(t, t, _MM_FROUND_TO_POS_INF);
+  return _mm_cvtsd_f64(t);
+#else
+  return ceil(f);
+#endif
+}
+
+static inline auto roundNearest(float f) {
+#ifdef SIMD_SUPPORT
+  __m128 t = _mm_set_ss(f);
+  t = _mm_round_ss(t, t, _MM_FROUND_TO_NEAREST_INT);
+  return _mm_cvtss_f32(t);
+#else
+  return roundf(f);
+#endif
+}
+
+static inline auto roundNearest(double f) {
+#ifdef SIMD_SUPPORT
+  __m128d t = _mm_set_sd(f);
+  t = _mm_round_sd(t, t, _MM_FROUND_TO_NEAREST_INT);
+  return _mm_cvtsd_f64(t);
+#else
+  return round(f);
+#endif
+}
+
+static inline auto roundCurrent(float f) {
+#ifdef SIMD_SUPPORT
+  auto t = _mm_set_ss(f);
+  t = _mm_round_ss(t, t, _MM_FROUND_CUR_DIRECTION);
+  return _mm_cvtss_f32(t);
+#else
+  return rint(f);
+#endif
+}
+
+static inline auto roundCurrent(double f) {
+#ifdef SIMD_SUPPORT
+  auto t = _mm_set_sd(f);
+  t = _mm_round_sd(t, t, _MM_FROUND_CUR_DIRECTION);
+  return _mm_cvtsd_f64(t);
+#else
+  return rint(f);
+#endif
+}
+
+
+static inline auto roundFloor(float f) {
+#ifdef SIMD_SUPPORT
+  __m128 t = _mm_set_ss(f);
+  t = _mm_round_ss(t, t, _MM_FROUND_TO_NEG_INF);
+  return _mm_cvtss_f32(t);
+#else
+  return floor(f);
+#endif
+}
+
+static inline auto roundFloor(double f) {
+#ifdef SIMD_SUPPORT
+  __m128d t = _mm_set_sd(f);
+  t = _mm_round_sd(t, t, _MM_FROUND_TO_NEG_INF);
+  return _mm_cvtsd_f64(t);
+#else
+  return floor(f);
+#endif
+}
+
+static inline auto roundTrunc(float f) {
+#ifdef SIMD_SUPPORT
+  __m128 t = _mm_set_ss(f);
+  t = _mm_round_ss(t, t, _MM_FROUND_TO_ZERO);
+  return _mm_cvtss_f32(t);
+#else
+  return trunc(f);
+#endif
+}
+
+static inline auto roundTrunc(double f) {
+#ifdef SIMD_SUPPORT
+  __m128d t = _mm_set_sd(f);
+  t = _mm_round_sd(t, t, _MM_FROUND_TO_ZERO);
+  return _mm_cvtsd_f64(t);
+#else
+  return trunc(f);
+#endif
+}
+} // namespace Util

external/ircolib/mem_access.hpp

@@ -0,0 +1,146 @@
+#pragma once
+#include <types.hpp>
+#include <cstring>
+#include <functional>
+#include <bit>
+#include <algorithm>
+#include <vector>
+#include <concepts>
+
+namespace ircolib {
+static inline std::vector<u8> IntegralToBuffer(const std::integral auto &val) {
+  std::vector<u8> ret{};
+  ret.resize(sizeof(val));
+
+  memcpy(ret.data(), &val, sizeof(val));
+
+  return ret;
+}
+
+static inline constexpr bool IsInsideRange(const std::integral auto& addr,
+                                                 const std::integral auto& start,
+                                                 const std::integral auto& end) {
+  return addr >= start && addr <= end;
+}
+
+template <typename T>
+static constexpr inline T ReadAccess(const u8 *data, const u32 index);
+template <typename T>
+static constexpr inline T ReadAccess(const std::vector<u8> &data, const u32 index);
+template <typename T, size_t Size>
+static constexpr inline T ReadAccess(const std::array<u8, Size> &data, const u32 index);
+
+template <typename T>
+static constexpr inline void WriteAccess(u8 *data, const u32 index, const T val);
+template <typename T>
+static constexpr inline void WriteAccess(std::vector<u8> &data, const u32 index, const T val);
+template <typename T, size_t Size>
+static constexpr inline void WriteAccess(std::array<u8, Size> &data, const u32 index, const T val);
+
+template <>
+constexpr inline u64 ReadAccess(const u8 *data, const u32 index) {
+  u32 hi = *reinterpret_cast<const u32 *>(&data[index + 0]);
+  u32 lo = *reinterpret_cast<const u32 *>(&data[index + 4]);
+  const auto& result = static_cast<u64>(hi) << 32 | static_cast<u64>(lo);
+  return result;
+}
+
+template <typename T>
+static constexpr inline T ReadAccess(const u8 *data, const u32 index) {
+  return *reinterpret_cast<const T *>(&data[index]);
+}
+
+template <>
+constexpr inline u64 ReadAccess(const std::vector<u8> &data, const u32 index) {
+  u32 hi = *reinterpret_cast<const u32 *>(&data[index + 0]);
+  u32 lo = *reinterpret_cast<const u32 *>(&data[index + 4]);
+  return (static_cast<u64>(hi) << 32) | static_cast<u64>(lo);
+}
+
+template <typename T>
+static constexpr inline T ReadAccess(const std::vector<u8> &data, const u32 index) {
+  return *reinterpret_cast<const T *>(&data[index]);
+}
+
+template <size_t Size>
+constexpr inline u64 ReadAccess(const std::array<u8, Size> &data, const u32 index) {
+  u32 hi = *reinterpret_cast<const u32 *>(&data[index + 0]);
+  u32 lo = *reinterpret_cast<const u32 *>(&data[index + 4]);
+  return static_cast<u64>(hi) << 32 | static_cast<u64>(lo);
+}
+
+template <typename T, size_t Size>
+static constexpr inline T ReadAccess(const std::array<u8, Size> &data, const u32 index) {
+  return *reinterpret_cast<const T *>(&data[index]);
+}
+
+template <size_t Size>
+constexpr inline void WriteAccess(std::array<u8, Size> &data, const u32 index, const u64 val) {
+  const u32 hi = val >> 32;
+  const u32 lo = val;
+
+  *reinterpret_cast<u32 *>(&data[index + 0]) = hi;
+  *reinterpret_cast<u32 *>(&data[index + 4]) = lo;
+}
+
+template <typename T, size_t Size>
+static constexpr inline void WriteAccess(std::array<u8, Size> &data, const u32 index, const T val) {
+  *reinterpret_cast<T *>(&data[index]) = val;
+}
+
+template <>
+constexpr inline void WriteAccess(std::vector<u8> &data, const u32 index, const u64 val) {
+  const u32 hi = val >> 32;
+  const u32 lo = val;
+
+  *reinterpret_cast<u32 *>(&data[index + 0]) = hi;
+  *reinterpret_cast<u32 *>(&data[index + 4]) = lo;
+}
+
+template <typename T>
+static constexpr inline void WriteAccess(std::vector<u8> &data, const u32 index, const T val) {
+  *reinterpret_cast<T *>(&data[index]) = val;
+}
+
+template <>
+constexpr inline void WriteAccess(u8 *data, const u32 index, const u64 val) {
+  const u32 hi = val >> 32;
+  const u32 lo = val;
+
+  *reinterpret_cast<u32 *>(&data[index + 0]) = hi;
+  *reinterpret_cast<u32 *>(&data[index + 4]) = lo;
+}
+
+template <typename T>
+static constexpr inline void WriteAccess(u8 *data, const u32 index, const T val) {
+  *reinterpret_cast<T *>(&data[index]) = val;
+}
+
+template <typename T>
+static constexpr inline void SwapBuffer(std::vector<u8> &data) {
+  for (size_t i = 0; i < data.size(); i += sizeof(T)) {
+    const T original = *reinterpret_cast<T *>(&data[i]);
+    *reinterpret_cast<T *>(&data[i]) = std::byteswap(original);
+  }
+}
+
+template <typename T, size_t Size>
+static constexpr inline void SwapBuffer(std::array<u8, Size> &data) {
+  for (size_t i = 0; i < data.size(); i += sizeof(T)) {
+    const T original = *reinterpret_cast<T *>(&data[i]);
+    *reinterpret_cast<T *>(&data[i]) = std::byteswap(original);
+  }
+}
+
+#ifdef _WIN32
+inline void *aligned_alloc(const size_t alignment, const size_t size) { return _aligned_malloc(size, alignment); }
+
+inline void aligned_free(void *ptr) { _aligned_free(ptr); }
+#else
+inline void *aligned_alloc(const size_t alignment, const size_t size) {
+  return std::aligned_alloc(alignment, size);
+}
+
+inline void aligned_free(void *ptr) { std::free(ptr); }
+#endif
+} // namespace Util

external/ircolib/types.hpp

@@ -0,0 +1,13 @@
+#pragma once
+#include <cstdint>
+
+namespace ircolib {
+using u8 = uint8_t;
+using u16 = uint16_t;
+using u32 = uint32_t;
+using u64 = uint64_t;
+using s8 = int8_t;
+using s16 = int16_t;
+using s32 = int32_t;
+using s64 = int64_t;
+}