When you're running out of registers while writing a JIT, you might resort to more unconventional methods for memory access. You might choose to resort to segment registers if you need a fixed register for memory offsets.
Instructions such as:
lea rax,gs:[rcx+rdx*8]
mov rax,gs:[rcx+rdx*8]
would then be available for your use.
| // Note that our hearing is logarithmic. | |
| // 0 dB = 1.0 | |
| // For the below equations, ratio refers to the amplitude ratio. | |
| // 0 dB = same power = same amplitude = same psycho-acoustic volume | |
| // -3 dB = half power | |
| // -6 dB = half amplitude | |
| // -10 dB = half psycho-acoustic volume = 32% amplitude | |
| double dB_to_ratio(double dB) { |
| template <typename T> | |
| constexpr T pext(T x, T mask) { | |
| T ret = 0; | |
| for (T bb = 1; mask != 0; bb += bb) { | |
| if (x & mask & -mask) { | |
| ret |= bb; | |
| } | |
| mask &= (mask - 1); | |
| } | |
| return ret; |
| #include <array> | |
| #include <cstddef> | |
| #include <type_traits> | |
| template <typename T, std::size_t N, std::size_t... Ns> | |
| struct multi_array | |
| { | |
| using type = std::array<typename multi_array<T, Ns...>::type, N>; | |
| }; |
| #include <array> | |
| #include <functional> | |
| template <typename T, size_t N, size_t M, size_t... iota_N, size_t... iota_M> | |
| auto append_impl( | |
| std::index_sequence<iota_N...>, | |
| std::index_sequence<iota_M...>, | |
| const std::array<T, N>& a, | |
| const std::array<T, M>& b | |
| ) -> std::array<T, N+M> { |
| Was writing a tuple class for fun. | |
| template <typename... Ts> | |
| class tuple { | |
| template <typename T> | |
| T& get(); | |
| template <typename T> | |
| T const& get() const; | |
| }; |
| BbAa contains BBBBbbbbAAAAaaaa | |
| DdCc contains DDDDddddCCCCcccc | |
| we want: DDDDCCCCBBBBAAAA | |
| we want: ddddccccbbbbaaaa | |
| const auto process_pair = [](const __m256i BbAa, const __m256i DdCc) { | |
| const __m256i DbCa = _mm256_blend_epi32(BbAa, DdCc, 0b11001100); | |
| const __m256i dBcA = _mm256_blend_epi32(BbAa, DdCc, 0b00110011); | |
| const __m256i DCba = _mm256_permute4x64_epi64(DbCa, 0b11011000); |
| // As with usual x64 ABI, rdi contains pointer to string | |
| strlen: | |
| xor eax, eax | |
| test dil, 7 | |
| je .skip_align | |
| // rax contains our return value, as per usual x64 ABI | |
| xor eax, eax | |
| // Line up to 8-byte alignment | |
| .align_loop: |
| #include <type_traits> | |
| /// Internal implementation of this header. | |
| namespace mp_impl { | |
| template<class... Ls> | |
| struct mp_append_impl; | |
| template<> struct mp_append_impl<> { | |
| using type = mp_list<>; |