Merge pull request #427 from jpcima/alignment

Aligned vector

cmake/SfizzConfig.cmake

@@ -60,6 +60,9 @@ if (CMAKE_CXX_COMPILER_ID MATCHES "GNU|Clang")
     add_compile_options(-Werror=return-type)
     if (SFIZZ_SYSTEM_PROCESSOR MATCHES "^(i.86|x86_64)$")
         add_compile_options(-msse2)
+    elseif (SFIZZ_SYSTEM_PROCESSOR MATCHES "^(armv.*)$")
+        add_compile_options(-mfloat-abi=hard)
+        add_compile_options(-mfpu=neon)
     endif()
 elseif (CMAKE_CXX_COMPILER_ID MATCHES "MSVC")
     set(CMAKE_CXX_STANDARD 17)
@@ -76,6 +79,10 @@ endfunction()
 # The sndfile library
 add_library(sfizz-sndfile INTERFACE)
 
+# The jsl utility library for C++
+add_library(sfizz-jsl INTERFACE)
+target_include_directories(sfizz-jsl INTERFACE "external/jsl/include")
+
 if (SFIZZ_USE_VCPKG OR CMAKE_CXX_COMPILER_ID MATCHES "MSVC")
     find_package(SndFile CONFIG REQUIRED)
     find_path(SNDFILE_INCLUDE_DIR sndfile.hh)

external/jsl/LICENSE.md

@@ -0,0 +1,23 @@
+# Boost Software License - Version 1.0 - August 17th, 2003
+
+Permission is hereby granted, free of charge, to any person or organization
+obtaining a copy of the software and accompanying documentation covered by
+this license (the "Software") to use, reproduce, display, distribute,
+execute, and transmit the Software, and to prepare derivative works of the
+Software, and to permit third-parties to whom the Software is furnished to
+do so, all subject to the following:
+
+The copyright notices in the Software and this entire statement, including
+the above license grant, this restriction and the following disclaimer,
+must be included in all copies of the Software, in whole or in part, and
+all derivative works of the Software, unless such copies or derivative
+works are solely in the form of machine-executable object code generated by
+a source language processor.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT
+SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE
+FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE,
+ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+DEALINGS IN THE SOFTWARE.

external/jsl/include/jsl/allocator

@@ -0,0 +1,73 @@
+// -*- C++ -*-
+#pragma once
+#include <limits>
+#include <utility>
+
+namespace jsl {
+
+template <class T, class Traits>
+struct ordinary_allocator {
+    typedef T value_type;
+    typedef T *pointer;
+    typedef const T *const_pointer;
+    typedef T &reference;
+    typedef const T &const_reference;
+    typedef std::size_t size_type;
+    typedef std::ptrdiff_t difference_type;
+    typedef std::true_type propagate_on_container_move_assignment;
+    template <class U> struct rebind { typedef ordinary_allocator<U, Traits> other; };
+    typedef std::true_type is_always_equal;
+
+    ordinary_allocator() noexcept {}
+    ordinary_allocator(const ordinary_allocator &) noexcept {}
+    template <class U> ordinary_allocator(const ordinary_allocator<U, Traits>&) noexcept {}
+
+    T *address(T &x) const noexcept;
+    const T *address(const T &x) const noexcept;
+
+    T *allocate(std::size_t n, const void * = nullptr);
+    void deallocate(T *p, std::size_t n) noexcept;
+    std::size_t max_size() const noexcept;
+
+    template <class U, class... Args> void construct(U *p, Args &&...args);
+    template <class U> void destroy(U *p);
+};
+
+template <class T, class Traits>
+inline bool operator==(const ordinary_allocator<T, Traits> &, const ordinary_allocator<T, Traits> &) noexcept
+{
+    return true;
+}
+
+template <class T, class Traits>
+inline bool operator!=(const ordinary_allocator<T, Traits> &, const ordinary_allocator<T, Traits> &) noexcept
+{
+    return false;
+}
+
+//------------------------------------------------------------------------------
+
+struct stdc_allocator_traits {
+    static void *allocate(std::size_t n);
+    static void deallocate(void *p, std::size_t = 0);
+};
+
+template <class T>
+using stdc_allocator = ordinary_allocator<T, stdc_allocator_traits>;
+
+//------------------------------------------------------------------------------
+
+template <std::size_t Al>
+struct aligned_allocator_traits {
+    static void *allocate(std::size_t n);
+    static void deallocate(void *p, std::size_t = 0);
+};
+
+template <class T, std::size_t Al>
+using aligned_allocator = ordinary_allocator<T, aligned_allocator_traits<Al>>;
+
+}  // namespace jsl
+
+#include "bits/allocator/ordinary_allocator.tcc"
+#include "bits/allocator/stdc_allocator.tcc"
+#include "bits/allocator/aligned_allocator.tcc"

external/jsl/include/jsl/bits/allocator/aligned_allocator.tcc

@@ -0,0 +1,41 @@
+#include "../../allocator"
+#include <new>
+#if defined(_WIN32)
+# include <malloc.h>
+#else
+# include <stdlib.h>
+#endif
+
+namespace jsl {
+
+template <std::size_t Al>
+void *aligned_allocator_traits<Al>::allocate(std::size_t n)
+{
+    static_assert(Al % sizeof(void *) == 0,
+                  "alignment must be a multiple of the pointer size");
+    static_assert((Al & (~Al + 1)) == Al,
+                  "alignment must be a power of two");
+#if defined(_WIN32)
+    void *p = ::_aligned_malloc(n, Al);
+    if (!p)
+        throw std::bad_alloc();
+    return p;
+#else
+    void *p;
+    if (::posix_memalign(&p, Al, n) != 0)
+        throw std::bad_alloc();
+    return p;
+#endif
+}
+
+template <std::size_t Al>
+void aligned_allocator_traits<Al>::deallocate(void *p, std::size_t)
+{
+#if defined(_WIN32)
+    ::_aligned_free(p);
+#else
+    ::free(p);
+#endif
+}
+
+}  // namespace jsl

external/jsl/include/jsl/bits/allocator/ordinary_allocator.tcc

@@ -0,0 +1,53 @@
+#include "../../allocator"
+#include <new>
+
+namespace jsl {
+
+template <class T, class Traits>
+inline T *ordinary_allocator<T, Traits>::address(T &x) const noexcept
+{
+    return &x;
+}
+
+template <class T, class Traits>
+inline const T *ordinary_allocator<T, Traits>::address(const T &x) const noexcept
+{
+    return &x;
+}
+
+template <class T, class Traits>
+T *ordinary_allocator<T, Traits>::allocate(std::size_t n, const void *)
+{
+    T *ptr = (T *)Traits::allocate(n * sizeof(T));
+    if (!ptr)
+        throw std::bad_alloc();
+    return ptr;
+}
+
+template <class T, class Traits>
+void ordinary_allocator<T, Traits>::deallocate(T *p, std::size_t n) noexcept
+{
+    Traits::deallocate(p, n * sizeof(T));
+}
+
+template <class T, class Traits>
+std::size_t ordinary_allocator<T, Traits>::max_size() const noexcept
+{
+    return std::numeric_limits<std::size_t>::max() / sizeof(T);
+}
+
+template <class T, class Traits>
+template <class U, class... Args>
+void ordinary_allocator<T, Traits>::construct(U *p, Args &&...args)
+{
+    ::new((void *)p) U(std::forward<Args>(args)...);
+}
+
+template <class T, class Traits>
+template <class U>
+void ordinary_allocator<T, Traits>::destroy(U *p)
+{
+    p->~U();
+}
+
+}  // namespace jsl

external/jsl/include/jsl/bits/allocator/stdc_allocator.tcc

@@ -0,0 +1,16 @@
+#include "../../allocator"
+#include <stdlib.h>
+
+namespace jsl {
+
+inline void *stdc_allocator_traits::allocate(std::size_t n)
+{
+    return ::malloc(n);
+}
+
+inline void stdc_allocator_traits::deallocate(void *p, std::size_t)
+{
+    ::free(p);
+}
+
+}  // namespace jsl

scripts/run_clang_tidy.sh

@@ -30,7 +30,7 @@ clang-tidy \
   vst/SfizzVstProcessor.cpp \
   vst/SfizzVstEditor.cpp \
   vst/SfizzVstState.cpp \
-  -- -Iexternal/abseil-cpp -Isrc/external -Isrc/external/pugixml/src \
+  -- -Iexternal/abseil-cpp -Iexternal/jsl/include -Isrc/external -Isrc/external/pugixml/src \
       -Isrc/sfizz -Isrc -Isrc/external/spline -Isrc/external/cpuid/src \
       -Ivst -Ivst/external/VST_SDK/VST3_SDK -Ivst/external/VST_SDK/VST3_SDK/vstgui4 -Ivst/external/ring_buffer \
       -Ieditor/src \

src/CMakeLists.txt

@@ -216,7 +216,7 @@ target_sources(sfizz_static PRIVATE
 target_include_directories (sfizz_static PUBLIC .)
 target_include_directories (sfizz_static PUBLIC external)
 target_link_libraries (sfizz_static PUBLIC absl::strings absl::span)
-target_link_libraries (sfizz_static PRIVATE sfizz_parser absl::flat_hash_map Threads::Threads sfizz-sndfile sfizz-pugixml sfizz-spline sfizz-tunings sfizz-kissfft sfizz-cpuid sfizz-atomic)
+target_link_libraries (sfizz_static PRIVATE sfizz_parser absl::flat_hash_map Threads::Threads sfizz-sndfile sfizz-pugixml sfizz-spline sfizz-tunings sfizz-kissfft sfizz-cpuid sfizz-jsl sfizz-atomic)
 set_target_properties (sfizz_static PROPERTIES OUTPUT_NAME sfizz PUBLIC_HEADER "sfizz.h;sfizz.hpp")
 if (WIN32)
     target_compile_definitions (sfizz_static PRIVATE _USE_MATH_DEFINES)
@@ -255,7 +255,7 @@ if (SFIZZ_SHARED)
         ${SFIZZ_HEADERS} ${SFIZZ_SOURCES} ${FAUST_FILES} sfizz/sfizz_wrapper.cpp sfizz/sfizz.cpp)
     target_include_directories (sfizz_shared PRIVATE .)
     target_include_directories (sfizz_shared PRIVATE external)
-    target_link_libraries (sfizz_shared PRIVATE absl::strings absl::span sfizz_parser absl::flat_hash_map Threads::Threads sfizz-sndfile sfizz-pugixml sfizz-spline sfizz-tunings sfizz-kissfft sfizz-cpuid sfizz-atomic)
+    target_link_libraries (sfizz_shared PRIVATE absl::strings absl::span sfizz_parser absl::flat_hash_map Threads::Threads sfizz-sndfile sfizz-pugixml sfizz-spline sfizz-tunings sfizz-kissfft sfizz-cpuid sfizz-jsl sfizz-atomic)
     if (WIN32)
         target_compile_definitions (sfizz_shared PRIVATE _USE_MATH_DEFINES)
     endif()

src/sfizz/Oversampler.cpp

@@ -9,6 +9,10 @@
 #include "AudioSpan.h"
 #include "AudioReader.h"
 #include "SIMDConfig.h"
+#include <jsl/allocator>
+
+template <class T, std::size_t A = sfz::config::defaultAlignment>
+using aligned_vector = std::vector<T, jsl::aligned_allocator<T, A>>;
 
 constexpr std::array<double, 12> coeffsStage2x {
     0.036681502163648017,
@@ -70,9 +74,9 @@ void sfz::Oversampler::stream(AudioSpan<float> input, AudioSpan<float> output, s
     const auto numFrames = input.getNumFrames();
     const auto numChannels = input.getNumChannels();
 
-    std::vector<Upsampler2x> upsampler2x;
+    aligned_vector<Upsampler2x> upsampler2x;
-    std::vector<Upsampler4x> upsampler4x;
+    aligned_vector<Upsampler4x> upsampler4x;
-    std::vector<Upsampler8x> upsampler8x;
+    aligned_vector<Upsampler8x> upsampler8x;
 
     switch(factor)
     {
@@ -145,9 +149,9 @@ void sfz::Oversampler::stream(AudioReader& input, AudioSpan<float> output, std::
     const auto numFrames = static_cast<size_t>(input.frames());
     const auto numChannels = input.channels();
 
-    std::vector<Upsampler2x> upsampler2x;
+    aligned_vector<Upsampler2x> upsampler2x;
-    std::vector<Upsampler4x> upsampler4x;
+    aligned_vector<Upsampler4x> upsampler4x;
-    std::vector<Upsampler8x> upsampler8x;
+    aligned_vector<Upsampler8x> upsampler8x;
 
     switch(factor)
     {