diff --git a/develop/cmake/SetupTargets.cmake b/develop/cmake/SetupTargets.cmake
index fd5eac3..c1aa6e8 100644
--- a/develop/cmake/SetupTargets.cmake
+++ b/develop/cmake/SetupTargets.cmake
@@ -1,9 +1,6 @@
 add_library(${PROJECT_NAME}.setup_targets INTERFACE)
 add_library(${PROJECT_NAME}::setup_targets ALIAS ${PROJECT_NAME}.setup_targets)
 
-target_link_libraries(${PROJECT_NAME}.setup_targets INTERFACE
-    vmath.hpp.vendors::doctest)
-
 target_compile_options(${PROJECT_NAME}.setup_targets INTERFACE
     $<$<CXX_COMPILER_ID:MSVC>:
         /WX /W4>
@@ -20,26 +17,13 @@ target_compile_options(${PROJECT_NAME}.setup_targets INTERFACE
         -Wno-unknown-warning-option
         >)
 
-if(BUILD_WITH_COVERAGE)
-    target_link_libraries(${PROJECT_NAME}.setup_targets INTERFACE
-        vmath.hpp::enable_gcov)
-endif()
-
-if(BUILD_WITH_SANITIZERS)
-    target_link_libraries(${PROJECT_NAME}.setup_targets INTERFACE
+target_link_libraries(${PROJECT_NAME}.setup_targets INTERFACE
+    $<$<BOOL:${BUILD_WITH_COVERAGE}>:
+        vmath.hpp::enable_gcov>
+    $<$<BOOL:${BUILD_WITH_SANITIZERS}>:
         vmath.hpp::enable_asan
-        vmath.hpp::enable_ubsan)
-endif()
-
-if(BUILD_WITH_NO_EXCEPTIONS)
-    target_link_libraries(${PROJECT_NAME}.setup_targets INTERFACE
-        vmath.hpp::disable_exceptions)
-
-    target_compile_definitions(${PROJECT_NAME}.setup_targets INTERFACE
-        DOCTEST_CONFIG_NO_EXCEPTIONS_BUT_WITH_ALL_ASSERTS)
-endif()
-
-if(BUILD_WITH_NO_RTTI)
-    target_link_libraries(${PROJECT_NAME}.setup_targets INTERFACE
-        vmath.hpp::disable_rtti)
-endif()
+        vmath.hpp::enable_ubsan>
+    $<$<BOOL:${BUILD_WITH_NO_EXCEPTIONS}>:
+        vmath.hpp::disable_exceptions>
+    $<$<BOOL:${BUILD_WITH_NO_RTTI}>:
+        vmath.hpp::disable_rtti>)
diff --git a/develop/singles/CMakeLists.txt b/develop/singles/CMakeLists.txt
index 930b461..ea1551e 100644
--- a/develop/singles/CMakeLists.txt
+++ b/develop/singles/CMakeLists.txt
@@ -13,10 +13,13 @@ set(VMATH_HPP_SINGLES_SCRIPT "${VMATH_HPP_ROOT_DIR}/develop/singles/scripts/buil
 file(GLOB_RECURSE VMATH_SINGLES_DEPENDS CONFIGURE_DEPENDS "${VMATH_HPP_ROOT_DIR}/headers/*.hpp")
 
 add_custom_command(OUTPUT "${VMATH_HPP_SINGLES_OUTPUT}"
-    COMMAND "${Python3_EXECUTABLE}" "${VMATH_HPP_SINGLES_SCRIPT}"
-            "${VMATH_HPP_SINGLES_INPUT}" "${VMATH_HPP_SINGLES_OUTPUT}"
-    DEPENDS ${VMATH_SINGLES_DEPENDS}
-    WORKING_DIRECTORY "${VMATH_HPP_ROOT_DIR}")
+    COMMAND
+        "${Python3_EXECUTABLE}" "${VMATH_HPP_SINGLES_SCRIPT}"
+        "${VMATH_HPP_SINGLES_INPUT}" "${VMATH_HPP_SINGLES_OUTPUT}"
+    DEPENDS
+        "${VMATH_HPP_SINGLES_SCRIPT}" ${VMATH_SINGLES_DEPENDS}
+    WORKING_DIRECTORY
+        "${VMATH_HPP_ROOT_DIR}")
 
 add_custom_target(${PROJECT_NAME}.generate
     DEPENDS "${VMATH_HPP_SINGLES_OUTPUT}")
diff --git a/develop/singles/headers/vmath.hpp/vmath_all.hpp b/develop/singles/headers/vmath.hpp/vmath_all.hpp
index f9265d1..c1d12d9 100644
--- a/develop/singles/headers/vmath.hpp/vmath_all.hpp
+++ b/develop/singles/headers/vmath.hpp/vmath_all.hpp
@@ -4,6 +4,8 @@
  * Copyright (C) 2020-2023, by Matvey Cherevko (blackmatov@gmail.com)
  ******************************************************************************/
 
+#pragma once
+
 #include <cmath>
 #include <cstddef>
 #include <cstdlib>
@@ -148,49 +150,49 @@ namespace vmath_hpp
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    floor(T x) noexcept {
+    constexpr floor(T x) noexcept {
         return std::floor(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    trunc(T x) noexcept {
+    constexpr trunc(T x) noexcept {
         return std::trunc(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    round(T x) noexcept {
+    constexpr round(T x) noexcept {
         return std::round(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    ceil(T x) noexcept {
+    constexpr ceil(T x) noexcept {
         return std::ceil(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    fract(T x) noexcept {
+    constexpr fract(T x) noexcept {
         return x - floor(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    fmod(T x, T y) noexcept {
+    constexpr fmod(T x, T y) noexcept {
         return std::fmod(x, y);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    modf(T x, T* y) noexcept {
+    constexpr modf(T x, T* y) noexcept {
         return std::modf(x, y);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    copysign(T x, T s) noexcept {
+    constexpr copysign(T x, T s) noexcept {
         return std::copysign(x, s);
     }
 
@@ -264,91 +266,91 @@ namespace vmath_hpp
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    sin(T x) noexcept {
+    constexpr sin(T x) noexcept {
         return std::sin(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    cos(T x) noexcept {
+    constexpr cos(T x) noexcept {
         return std::cos(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    tan(T x) noexcept {
+    constexpr tan(T x) noexcept {
         return std::tan(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    asin(T x) noexcept {
+    constexpr asin(T x) noexcept {
         return std::asin(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    acos(T x) noexcept {
+    constexpr acos(T x) noexcept {
         return std::acos(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    atan(T x) noexcept {
+    constexpr atan(T x) noexcept {
         return std::atan(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    atan2(T y, T x) noexcept {
+    constexpr atan2(T y, T x) noexcept {
         return std::atan2(y, x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    sinh(T x) noexcept {
+    constexpr sinh(T x) noexcept {
         return std::sinh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    cosh(T x) noexcept {
+    constexpr cosh(T x) noexcept {
         return std::cosh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    tanh(T x) noexcept {
+    constexpr tanh(T x) noexcept {
         return std::tanh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    asinh(T x) noexcept {
+    constexpr asinh(T x) noexcept {
         return std::asinh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    acosh(T x) noexcept {
+    constexpr acosh(T x) noexcept {
         return std::acosh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    atanh(T x) noexcept {
+    constexpr atanh(T x) noexcept {
         return std::atanh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, std::pair<T, T>>
-    sincos(T x) noexcept {
+    constexpr sincos(T x) noexcept {
         return { sin(x), cos(x) };
     }
 
     template < typename T >
     std::enable_if_t<std::is_floating_point_v<T>, void>
-    sincos(T x, T* s, T* c) noexcept {
+    constexpr sincos(T x, T* s, T* c) noexcept {
         *s = sin(x);
         *c = cos(x);
     }
@@ -362,43 +364,43 @@ namespace vmath_hpp
 {
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    pow(T x, T y) noexcept {
+    constexpr pow(T x, T y) noexcept {
         return std::pow(x, y);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    exp(T x) noexcept {
+    constexpr exp(T x) noexcept {
         return std::exp(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    log(T x) noexcept {
+    constexpr log(T x) noexcept {
         return std::log(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    exp2(T x) noexcept {
+    constexpr exp2(T x) noexcept {
         return std::exp2(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    log2(T x) noexcept {
+    constexpr log2(T x) noexcept {
         return std::log2(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    sqrt(T x) noexcept {
+    constexpr sqrt(T x) noexcept {
         return std::sqrt(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    rsqrt(T x) noexcept {
+    constexpr rsqrt(T x) noexcept {
         return rcp(sqrt(x));
     }
 }
@@ -481,7 +483,7 @@ namespace vmath_hpp
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    refract(T i, T n, T eta) noexcept {
+    constexpr refract(T i, T n, T eta) noexcept {
         const T d = dot(n, i);
         const T k = T{1} - sqr(eta) * (T{1} - sqr(d));
         return k < T{0} ? T{0} : (eta * i - (eta * d + sqrt(k)) * n);
@@ -2874,226 +2876,139 @@ namespace vmath_hpp
     // transpose
     //
 
-    namespace impl
-    {
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 2> transpose_2x2_impl(
-            T a, T c,
-            T b, T d)
-        {
-            return {
-                a, b,
-                c, d};
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 3> transpose_3x3_impl(
-            T a, T d, T g,
-            T b, T e, T h,
-            T c, T f, T i)
-        {
-            return {
-                a, b, c,
-                d, e, f,
-                g, h, i};
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 4> transpose_4x4_impl(
-            T a, T e, T i, T m,
-            T b, T f, T j, T n,
-            T c, T g, T k, T o,
-            T d, T h, T l, T p)
-        {
-            return {
-                a, b, c, d,
-                e, f, g, h,
-                i, j, k, l,
-                m, n, o, p};
-        }
+    template < typename T >
+    [[nodiscard]] constexpr mat<T, 2> transpose(const mat<T, 2>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1],
+                c = _m[1][0], d = _m[1][1];
+        return {
+            a, c,
+            b, d};
     }
 
     template < typename T >
-    [[nodiscard]] constexpr mat<T, 2> transpose(const mat<T, 2>& m) {
-        return impl::transpose_2x2_impl(
-            m[0][0], m[0][1],
-            m[1][0], m[1][1]);
+    [[nodiscard]] constexpr mat<T, 3> transpose(const mat<T, 3>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2],
+                d = _m[1][0], e = _m[1][1], f = _m[1][2],
+                g = _m[2][0], h = _m[2][1], i = _m[2][2];
+        return {
+            a, d, g,
+            b, e, h,
+            c, f, i};
     }
 
     template < typename T >
-    [[nodiscard]] constexpr mat<T, 3> transpose(const mat<T, 3>& m) {
-        return impl::transpose_3x3_impl(
-            m[0][0], m[0][1], m[0][2],
-            m[1][0], m[1][1], m[1][2],
-            m[2][0], m[2][1], m[2][2]);
-    }
-
-    template < typename T >
-    [[nodiscard]] constexpr mat<T, 4> transpose(const mat<T, 4>& m) {
-        return impl::transpose_4x4_impl(
-            m[0][0], m[0][1], m[0][2], m[0][3],
-            m[1][0], m[1][1], m[1][2], m[1][3],
-            m[2][0], m[2][1], m[2][2], m[2][3],
-            m[3][0], m[3][1], m[3][2], m[3][3]);
+    [[nodiscard]] constexpr mat<T, 4> transpose(const mat<T, 4>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2], d = _m[0][3],
+                e = _m[1][0], f = _m[1][1], g = _m[1][2], h = _m[1][3],
+                i = _m[2][0], j = _m[2][1], k = _m[2][2], l = _m[2][3],
+                m = _m[3][0], n = _m[3][1], o = _m[3][2], p = _m[3][3];
+        return {
+            a, e, i, m,
+            b, f, j, n,
+            c, g, k, o,
+            d, h, l, p};
     }
 
     //
     // adjugate
     //
 
-    namespace impl
-    {
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 2> adjugate_2x2_impl(
-            T a, T b,
-            T c, T d)
-        {
-            return {
-                +d, -b,
-                -c, +a};
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 3> adjugate_3x3_impl(
-            T a, T b, T c,
-            T d, T e, T f,
-            T g, T h, T i)
-        {
-            return {
-                e * i - f * h,
-                c * h - b * i,
-                b * f - c * e,
-                f * g - d * i,
-                a * i - c * g,
-                c * d - a * f,
-                d * h - e * g,
-                b * g - a * h,
-                a * e - b * d};
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 4> adjugate_4x4_impl(
-            T a, T b, T c, T d,
-            T e, T f, T g, T h,
-            T i, T j, T k, T l,
-            T m, T n, T o, T p)
-        {
-            return {
-                f * (k * p - l * o) + g * (l * n - j * p) + h * (j * o - k * n),
-                j * (c * p - d * o) + k * (d * n - b * p) + l * (b * o - c * n),
-                n * (c * h - d * g) + o * (d * f - b * h) + p * (b * g - c * f),
-                b * (h * k - g * l) + c * (f * l - h * j) + d * (g * j - f * k),
-                g * (i * p - l * m) + h * (k * m - i * o) + e * (l * o - k * p),
-                k * (a * p - d * m) + l * (c * m - a * o) + i * (d * o - c * p),
-                o * (a * h - d * e) + p * (c * e - a * g) + m * (d * g - c * h),
-                c * (h * i - e * l) + d * (e * k - g * i) + a * (g * l - h * k),
-                h * (i * n - j * m) + e * (j * p - l * n) + f * (l * m - i * p),
-                l * (a * n - b * m) + i * (b * p - d * n) + j * (d * m - a * p),
-                p * (a * f - b * e) + m * (b * h - d * f) + n * (d * e - a * h),
-                d * (f * i - e * j) + a * (h * j - f * l) + b * (e * l - h * i),
-                e * (k * n - j * o) + f * (i * o - k * m) + g * (j * m - i * n),
-                i * (c * n - b * o) + j * (a * o - c * m) + k * (b * m - a * n),
-                m * (c * f - b * g) + n * (a * g - c * e) + o * (b * e - a * f),
-                a * (f * k - g * j) + b * (g * i - e * k) + c * (e * j - f * i)};
-        }
+    template < typename T >
+    [[nodiscard]] constexpr mat<T, 2> adjugate(const mat<T, 2>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1],
+                c = _m[1][0], d = _m[1][1];
+        return {
+            +d,
+            -b,
+            -c,
+            +a};
     }
 
     template < typename T >
-    [[nodiscard]] constexpr mat<T, 2> adjugate(const mat<T, 2>& m) {
-        return impl::adjugate_2x2_impl(
-            m[0][0], m[0][1],
-            m[1][0], m[1][1]);
+    [[nodiscard]] constexpr mat<T, 3> adjugate(const mat<T, 3>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2],
+                d = _m[1][0], e = _m[1][1], f = _m[1][2],
+                g = _m[2][0], h = _m[2][1], i = _m[2][2];
+        return {
+            e * i - f * h,
+            c * h - b * i,
+            b * f - c * e,
+            f * g - d * i,
+            a * i - c * g,
+            c * d - a * f,
+            d * h - e * g,
+            b * g - a * h,
+            a * e - b * d};
     }
 
     template < typename T >
-    [[nodiscard]] constexpr mat<T, 3> adjugate(const mat<T, 3>& m) {
-        return impl::adjugate_3x3_impl(
-            m[0][0], m[0][1], m[0][2],
-            m[1][0], m[1][1], m[1][2],
-            m[2][0], m[2][1], m[2][2]);
-    }
-
-    template < typename T >
-    [[nodiscard]] constexpr mat<T, 4> adjugate(const mat<T, 4>& m) {
-        return impl::adjugate_4x4_impl(
-            m[0][0], m[0][1], m[0][2], m[0][3],
-            m[1][0], m[1][1], m[1][2], m[1][3],
-            m[2][0], m[2][1], m[2][2], m[2][3],
-            m[3][0], m[3][1], m[3][2], m[3][3]);
+    [[nodiscard]] constexpr mat<T, 4> adjugate(const mat<T, 4>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2], d = _m[0][3],
+                e = _m[1][0], f = _m[1][1], g = _m[1][2], h = _m[1][3],
+                i = _m[2][0], j = _m[2][1], k = _m[2][2], l = _m[2][3],
+                m = _m[3][0], n = _m[3][1], o = _m[3][2], p = _m[3][3];
+        return {
+            f * (k * p - l * o) + g * (l * n - j * p) + h * (j * o - k * n),
+            j * (c * p - d * o) + k * (d * n - b * p) + l * (b * o - c * n),
+            n * (c * h - d * g) + o * (d * f - b * h) + p * (b * g - c * f),
+            b * (h * k - g * l) + c * (f * l - h * j) + d * (g * j - f * k),
+            g * (i * p - l * m) + h * (k * m - i * o) + e * (l * o - k * p),
+            k * (a * p - d * m) + l * (c * m - a * o) + i * (d * o - c * p),
+            o * (a * h - d * e) + p * (c * e - a * g) + m * (d * g - c * h),
+            c * (h * i - e * l) + d * (e * k - g * i) + a * (g * l - h * k),
+            h * (i * n - j * m) + e * (j * p - l * n) + f * (l * m - i * p),
+            l * (a * n - b * m) + i * (b * p - d * n) + j * (d * m - a * p),
+            p * (a * f - b * e) + m * (b * h - d * f) + n * (d * e - a * h),
+            d * (f * i - e * j) + a * (h * j - f * l) + b * (e * l - h * i),
+            e * (k * n - j * o) + f * (i * o - k * m) + g * (j * m - i * n),
+            i * (c * n - b * o) + j * (a * o - c * m) + k * (b * m - a * n),
+            m * (c * f - b * g) + n * (a * g - c * e) + o * (b * e - a * f),
+            a * (f * k - g * j) + b * (g * i - e * k) + c * (e * j - f * i)};
     }
 
     //
     // determinant
     //
 
-    namespace impl
-    {
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        T determinant_2x2_impl(
-            T a, T b,
-            T c, T d)
-        {
-            return a * d - b * c;
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        T determinant_3x3_impl(
-            T a, T b, T c,
-            T d, T e, T f,
-            T g, T h, T i)
-        {
-            return
-                + a * (e * i - f * h)
-                - b * (d * i - f * g)
-                + c * (d * h - e * g);
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        T determinant_4x4_impl(
-            T a, T b, T c, T d,
-            T e, T f, T g, T h,
-            T i, T j, T k, T l,
-            T m, T n, T o, T p)
-        {
-            return
-                + a * (f * (k * p - l * o) - (j * (g * p - h * o)) + (n * (g * l - h * k)))
-                - b * (e * (k * p - l * o) - (i * (g * p - h * o)) + (m * (g * l - h * k)))
-                + c * (e * (j * p - l * n) - (i * (f * p - h * n)) + (m * (f * l - h * j)))
-                - d * (e * (j * o - k * n) - (i * (f * o - g * n)) + (m * (f * k - g * j)));
-        }
+    template < typename T >
+    [[nodiscard]] constexpr T determinant(const mat<T, 2>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1],
+                c = _m[1][0], d = _m[1][1];
+        return
+            a * d - b * c;
     }
 
     template < typename T >
-    [[nodiscard]] constexpr T determinant(const mat<T, 2>& m) {
-        return impl::determinant_2x2_impl(
-            m[0][0], m[0][1],
-            m[1][0], m[1][1]);
+    [[nodiscard]] constexpr T determinant(const mat<T, 3>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2],
+                d = _m[1][0], e = _m[1][1], f = _m[1][2],
+                g = _m[2][0], h = _m[2][1], i = _m[2][2];
+        return
+            + a * (e * i - f * h)
+            - b * (d * i - f * g)
+            + c * (d * h - e * g);
     }
 
     template < typename T >
-    [[nodiscard]] constexpr T determinant(const mat<T, 3>& m) {
-        return impl::determinant_3x3_impl(
-            m[0][0], m[0][1], m[0][2],
-            m[1][0], m[1][1], m[1][2],
-            m[2][0], m[2][1], m[2][2]);
-    }
-
-    template < typename T >
-    [[nodiscard]] constexpr T determinant(const mat<T, 4>& m) {
-        return impl::determinant_4x4_impl(
-            m[0][0], m[0][1], m[0][2], m[0][3],
-            m[1][0], m[1][1], m[1][2], m[1][3],
-            m[2][0], m[2][1], m[2][2], m[2][3],
-            m[3][0], m[3][1], m[3][2], m[3][3]);
+    [[nodiscard]] constexpr T determinant(const mat<T, 4>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2], d = _m[0][3],
+                e = _m[1][0], f = _m[1][1], g = _m[1][2], h = _m[1][3],
+                i = _m[2][0], j = _m[2][1], k = _m[2][2], l = _m[2][3],
+                m = _m[3][0], n = _m[3][1], o = _m[3][2], p = _m[3][3];
+        return
+            + a * (f * (k * p - l * o) - (j * (g * p - h * o)) + (n * (g * l - h * k)))
+            - b * (e * (k * p - l * o) - (i * (g * p - h * o)) + (m * (g * l - h * k)))
+            + c * (e * (j * p - l * n) - (i * (f * p - h * n)) + (m * (f * l - h * j)))
+            - d * (e * (j * o - k * n) - (i * (f * o - g * n)) + (m * (f * k - g * j)));
     }
 
     //
@@ -3661,12 +3576,12 @@ namespace vmath_hpp::detail
         }
     };
 
-    template < typename T, size_t Size >
+    template < typename T, std::size_t Size >
     [[nodiscard]] std::size_t hash(const vec<T, Size>& v) noexcept {
         return fold_join(hash_combiner{}, std::size_t{}, v);
     }
 
-    template < typename T, size_t Size >
+    template < typename T, std::size_t Size >
     [[nodiscard]] std::size_t hash(const mat<T, Size>& m) noexcept {
         return fold_join(hash_combiner{}, std::size_t{}, m);
     }
diff --git a/develop/singles/scripts/build_singles.py b/develop/singles/scripts/build_singles.py
index b3af5f8..a4bff29 100755
--- a/develop/singles/scripts/build_singles.py
+++ b/develop/singles/scripts/build_singles.py
@@ -5,10 +5,6 @@ import os
 import re
 import sys
 
-#
-#
-#
-
 EMPTY_MATCHER = re.compile(r'^\s*$')
 C_COMMENT_MATCHER = re.compile(r'^/\*.*\*/', re.S)
 
@@ -19,9 +15,6 @@ USER_INCLUDE_MATCHER = re.compile(r'#\s*include\s*\"(.*)\"')
 SYSTEM_INCLUDE_MATCHER = re.compile(r'#\s*include\s*<(.*)>')
 PRAGMA_ONCE_MATCHER = re.compile(r'#\s*pragma\s+once')
 
-#
-#
-#
 
 def CollectLicenseComment(headerPath):
     with open(headerPath, "r") as headerStream:
@@ -29,7 +22,8 @@ def CollectLicenseComment(headerPath):
         commentMatch = re.match(C_COMMENT_MATCHER, headerContent)
         return commentMatch.group() if commentMatch else ""
 
-def CollectSystemIncludes(headerPath, parsedHeaders = set()):
+
+def CollectSystemIncludes(headerPath, parsedHeaders=set()):
     with open(headerPath, "r") as headerStream:
         headerContent = headerStream.read().strip()
 
@@ -50,8 +44,10 @@ def CollectSystemIncludes(headerPath, parsedHeaders = set()):
 
             includeMatch = USER_INCLUDE_MATCHER.findall(headerLine)
             if includeMatch and ifdefScopeLevel == 0:
-                internalHeaderPath = os.path.abspath(os.path.join(os.path.dirname(headerPath), includeMatch[0]))
-                headerIncludes = headerIncludes.union(CollectSystemIncludes(internalHeaderPath, parsedHeaders))
+                internalHeaderPath = os.path.abspath(os.path.join(
+                    os.path.dirname(headerPath), includeMatch[0]))
+                headerIncludes = headerIncludes.union(
+                    CollectSystemIncludes(internalHeaderPath, parsedHeaders))
 
             includeMatch = SYSTEM_INCLUDE_MATCHER.findall(headerLine)
             if includeMatch and ifdefScopeLevel == 0:
@@ -59,7 +55,8 @@ def CollectSystemIncludes(headerPath, parsedHeaders = set()):
 
         return headerIncludes
 
-def ParseHeader(headerPath, parsedHeaders = set()):
+
+def ParseHeader(headerPath, parsedHeaders=set()):
     with open(headerPath, "r") as headerStream:
         headerContent = headerStream.read().strip()
         headerContent = re.sub(C_COMMENT_MATCHER, '', headerContent)
@@ -89,8 +86,10 @@ def ParseHeader(headerPath, parsedHeaders = set()):
 
             includeMatch = USER_INCLUDE_MATCHER.findall(headerLine)
             if includeMatch and ifdefScopeLevel == 0:
-                internalHeaderPath = os.path.abspath(os.path.join(os.path.dirname(headerPath), includeMatch[0]))
-                internalHeaderContent = ParseHeader(internalHeaderPath, parsedHeaders)
+                internalHeaderPath = os.path.abspath(os.path.join(
+                    os.path.dirname(headerPath), includeMatch[0]))
+                internalHeaderContent = ParseHeader(
+                    internalHeaderPath, parsedHeaders)
 
                 outputContent += internalHeaderContent
                 shouldSkipNextEmptyLines = True
@@ -106,9 +105,6 @@ def ParseHeader(headerPath, parsedHeaders = set()):
 
         return "{}\n".format(outputContent)
 
-#
-#
-#
 
 inputHeaderPath = os.path.abspath(sys.argv[1])
 outputHeaderPath = os.path.abspath(sys.argv[2])
@@ -118,13 +114,20 @@ os.makedirs(os.path.dirname(outputHeaderPath), exist_ok=True)
 with open(outputHeaderPath, "w") as outputHeaderStream:
     licenseComment = CollectLicenseComment(inputHeaderPath)
     systemIncludes = CollectSystemIncludes(inputHeaderPath)
+    headersContent = ParseHeader(inputHeaderPath)
 
-    outputHeaderStream.write("{}\n".format(licenseComment))
+    if licenseComment:
+        outputHeaderStream.write("{}\n".format(licenseComment))
+        outputHeaderStream.write("\n")
+
+    outputHeaderStream.write("#pragma once\n")
     outputHeaderStream.write("\n")
 
-    for systemInclude in sorted(systemIncludes):
-        outputHeaderStream.write("#include <{}>\n".format(systemInclude))
-    outputHeaderStream.write("\n")
+    if systemIncludes:
+        for systemInclude in sorted(systemIncludes):
+            outputHeaderStream.write("#include <{}>\n".format(systemInclude))
+        outputHeaderStream.write("\n")
 
-    outputHeaderStream.write(ParseHeader(inputHeaderPath).strip())
-    outputHeaderStream.write("\n")
+    if headersContent:
+        outputHeaderStream.write(headersContent.strip())
+        outputHeaderStream.write("\n")
diff --git a/develop/untests/CMakeLists.txt b/develop/untests/CMakeLists.txt
index af6f006..1cb65c8 100644
--- a/develop/untests/CMakeLists.txt
+++ b/develop/untests/CMakeLists.txt
@@ -8,11 +8,13 @@ add_executable(${PROJECT_NAME}.singles ${UNTESTS_SOURCES})
 
 target_link_libraries(${PROJECT_NAME} PRIVATE
     vmath.hpp::vmath.hpp
-    vmath.hpp::setup_targets)
+    vmath.hpp::setup_targets
+    vmath.hpp.vendors::doctest)
 
 target_link_libraries(${PROJECT_NAME}.singles PRIVATE
     vmath.hpp::singles
-    vmath.hpp::setup_targets)
+    vmath.hpp::setup_targets
+    vmath.hpp.vendors::doctest)
 
 add_test(${PROJECT_NAME} ${PROJECT_NAME})
 add_test(${PROJECT_NAME} ${PROJECT_NAME}.singles)
diff --git a/develop/untests/vmath_mat_fun_tests.cpp b/develop/untests/vmath_mat_fun_tests.cpp
index 7edc62f..2e5b1bd 100644
--- a/develop/untests/vmath_mat_fun_tests.cpp
+++ b/develop/untests/vmath_mat_fun_tests.cpp
@@ -322,34 +322,38 @@ TEST_CASE("vmath/mat_fun") {
             constexpr fmat4 rm1 = inverse(m1);
             STATIC_CHECK(all(approx(
                 unit4_z<float> * m1 * rm1,
-                unit4_z<float>)));
+                unit4_z<float>,
+                0.00001f)));
         }
 
         {
             const fvec3 axis2 = normalize(fvec3(1.f, 2.f, 3.f));
-            const fmat4 m2 = fmat4(rotate(0.5f,axis2));
+            const fmat4 m2 = perspective_lh(800.f, 600.f, -1.f, 1.f) * translate(fvec3(1.f, 2.f, 3.f)) * scale4(fvec3(2.f, 3.f, 4.f)) * fmat4(rotate(0.5f,axis2));
             const fmat4 rm2 = inverse(m2);
             CHECK(all(approx(
                 unit4_z<float> * m2 * rm2,
-                unit4_z<float>)));
+                unit4_z<float>,
+                0.00001f)));
         }
 
         {
             const fvec3 axis3 = normalize(fvec3(1.f, 2.f, 3.f));
-            const fmat3 m3 = fmat3(rotate(0.5f,axis3));
+            const fmat3 m3 = translate(fvec2(1.f, 2.f)) * scale(fvec3(2.f, 3.f, 4.f)) * fmat3(rotate(0.5f,axis3));
             const fmat3 rm3 = inverse(m3);
             CHECK(all(approx(
                 unit3_z<float> * m3 * rm3,
-                unit3_z<float>)));
+                unit3_z<float>,
+                0.00001f)));
         }
 
         {
             const fvec3 axis4 = normalize(fvec3(0.f, 0.f, 3.f));
-            const fmat2 m4 = fmat2(rotate(0.5f,axis4));
+            const fmat2 m4 = scale(fvec2(2.f, 3.f)) * fmat2(rotate(0.5f,axis4));
             const fmat2 rm4 = inverse(m4);
             CHECK(all(approx(
                 unit2_y<float> * m4 * rm4,
-                unit2_y<float>)));
+                unit2_y<float>,
+                0.00001f)));
         }
     }
 }
diff --git a/headers/vmath.hpp/vmath_ext.hpp b/headers/vmath.hpp/vmath_ext.hpp
index f85e8a1..6a41393 100644
--- a/headers/vmath.hpp/vmath_ext.hpp
+++ b/headers/vmath.hpp/vmath_ext.hpp
@@ -68,12 +68,12 @@ namespace vmath_hpp::detail
         }
     };
 
-    template < typename T, size_t Size >
+    template < typename T, std::size_t Size >
     [[nodiscard]] std::size_t hash(const vec<T, Size>& v) noexcept {
         return fold_join(hash_combiner{}, std::size_t{}, v);
     }
 
-    template < typename T, size_t Size >
+    template < typename T, std::size_t Size >
     [[nodiscard]] std::size_t hash(const mat<T, Size>& m) noexcept {
         return fold_join(hash_combiner{}, std::size_t{}, m);
     }
diff --git a/headers/vmath.hpp/vmath_fun.hpp b/headers/vmath.hpp/vmath_fun.hpp
index 8c66aac..48a24a5 100644
--- a/headers/vmath.hpp/vmath_fun.hpp
+++ b/headers/vmath.hpp/vmath_fun.hpp
@@ -44,49 +44,49 @@ namespace vmath_hpp
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    floor(T x) noexcept {
+    constexpr floor(T x) noexcept {
         return std::floor(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    trunc(T x) noexcept {
+    constexpr trunc(T x) noexcept {
         return std::trunc(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    round(T x) noexcept {
+    constexpr round(T x) noexcept {
         return std::round(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    ceil(T x) noexcept {
+    constexpr ceil(T x) noexcept {
         return std::ceil(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    fract(T x) noexcept {
+    constexpr fract(T x) noexcept {
         return x - floor(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    fmod(T x, T y) noexcept {
+    constexpr fmod(T x, T y) noexcept {
         return std::fmod(x, y);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    modf(T x, T* y) noexcept {
+    constexpr modf(T x, T* y) noexcept {
         return std::modf(x, y);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    copysign(T x, T s) noexcept {
+    constexpr copysign(T x, T s) noexcept {
         return std::copysign(x, s);
     }
 
@@ -160,91 +160,91 @@ namespace vmath_hpp
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    sin(T x) noexcept {
+    constexpr sin(T x) noexcept {
         return std::sin(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    cos(T x) noexcept {
+    constexpr cos(T x) noexcept {
         return std::cos(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    tan(T x) noexcept {
+    constexpr tan(T x) noexcept {
         return std::tan(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    asin(T x) noexcept {
+    constexpr asin(T x) noexcept {
         return std::asin(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    acos(T x) noexcept {
+    constexpr acos(T x) noexcept {
         return std::acos(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    atan(T x) noexcept {
+    constexpr atan(T x) noexcept {
         return std::atan(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    atan2(T y, T x) noexcept {
+    constexpr atan2(T y, T x) noexcept {
         return std::atan2(y, x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    sinh(T x) noexcept {
+    constexpr sinh(T x) noexcept {
         return std::sinh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    cosh(T x) noexcept {
+    constexpr cosh(T x) noexcept {
         return std::cosh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    tanh(T x) noexcept {
+    constexpr tanh(T x) noexcept {
         return std::tanh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    asinh(T x) noexcept {
+    constexpr asinh(T x) noexcept {
         return std::asinh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    acosh(T x) noexcept {
+    constexpr acosh(T x) noexcept {
         return std::acosh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    atanh(T x) noexcept {
+    constexpr atanh(T x) noexcept {
         return std::atanh(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, std::pair<T, T>>
-    sincos(T x) noexcept {
+    constexpr sincos(T x) noexcept {
         return { sin(x), cos(x) };
     }
 
     template < typename T >
     std::enable_if_t<std::is_floating_point_v<T>, void>
-    sincos(T x, T* s, T* c) noexcept {
+    constexpr sincos(T x, T* s, T* c) noexcept {
         *s = sin(x);
         *c = cos(x);
     }
@@ -258,43 +258,43 @@ namespace vmath_hpp
 {
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    pow(T x, T y) noexcept {
+    constexpr pow(T x, T y) noexcept {
         return std::pow(x, y);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    exp(T x) noexcept {
+    constexpr exp(T x) noexcept {
         return std::exp(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    log(T x) noexcept {
+    constexpr log(T x) noexcept {
         return std::log(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    exp2(T x) noexcept {
+    constexpr exp2(T x) noexcept {
         return std::exp2(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    log2(T x) noexcept {
+    constexpr log2(T x) noexcept {
         return std::log2(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    sqrt(T x) noexcept {
+    constexpr sqrt(T x) noexcept {
         return std::sqrt(x);
     }
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    rsqrt(T x) noexcept {
+    constexpr rsqrt(T x) noexcept {
         return rcp(sqrt(x));
     }
 }
@@ -377,7 +377,7 @@ namespace vmath_hpp
 
     template < typename T >
     [[nodiscard]] std::enable_if_t<std::is_floating_point_v<T>, T>
-    refract(T i, T n, T eta) noexcept {
+    constexpr refract(T i, T n, T eta) noexcept {
         const T d = dot(n, i);
         const T k = T{1} - sqr(eta) * (T{1} - sqr(d));
         return k < T{0} ? T{0} : (eta * i - (eta * d + sqrt(k)) * n);
diff --git a/headers/vmath.hpp/vmath_mat_fun.hpp b/headers/vmath.hpp/vmath_mat_fun.hpp
index 4fcc33a..055a510 100644
--- a/headers/vmath.hpp/vmath_mat_fun.hpp
+++ b/headers/vmath.hpp/vmath_mat_fun.hpp
@@ -638,226 +638,139 @@ namespace vmath_hpp
     // transpose
     //
 
-    namespace impl
-    {
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 2> transpose_2x2_impl(
-            T a, T c,
-            T b, T d)
-        {
-            return {
-                a, b,
-                c, d};
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 3> transpose_3x3_impl(
-            T a, T d, T g,
-            T b, T e, T h,
-            T c, T f, T i)
-        {
-            return {
-                a, b, c,
-                d, e, f,
-                g, h, i};
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 4> transpose_4x4_impl(
-            T a, T e, T i, T m,
-            T b, T f, T j, T n,
-            T c, T g, T k, T o,
-            T d, T h, T l, T p)
-        {
-            return {
-                a, b, c, d,
-                e, f, g, h,
-                i, j, k, l,
-                m, n, o, p};
-        }
+    template < typename T >
+    [[nodiscard]] constexpr mat<T, 2> transpose(const mat<T, 2>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1],
+                c = _m[1][0], d = _m[1][1];
+        return {
+            a, c,
+            b, d};
     }
 
     template < typename T >
-    [[nodiscard]] constexpr mat<T, 2> transpose(const mat<T, 2>& m) {
-        return impl::transpose_2x2_impl(
-            m[0][0], m[0][1],
-            m[1][0], m[1][1]);
+    [[nodiscard]] constexpr mat<T, 3> transpose(const mat<T, 3>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2],
+                d = _m[1][0], e = _m[1][1], f = _m[1][2],
+                g = _m[2][0], h = _m[2][1], i = _m[2][2];
+        return {
+            a, d, g,
+            b, e, h,
+            c, f, i};
     }
 
     template < typename T >
-    [[nodiscard]] constexpr mat<T, 3> transpose(const mat<T, 3>& m) {
-        return impl::transpose_3x3_impl(
-            m[0][0], m[0][1], m[0][2],
-            m[1][0], m[1][1], m[1][2],
-            m[2][0], m[2][1], m[2][2]);
-    }
-
-    template < typename T >
-    [[nodiscard]] constexpr mat<T, 4> transpose(const mat<T, 4>& m) {
-        return impl::transpose_4x4_impl(
-            m[0][0], m[0][1], m[0][2], m[0][3],
-            m[1][0], m[1][1], m[1][2], m[1][3],
-            m[2][0], m[2][1], m[2][2], m[2][3],
-            m[3][0], m[3][1], m[3][2], m[3][3]);
+    [[nodiscard]] constexpr mat<T, 4> transpose(const mat<T, 4>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2], d = _m[0][3],
+                e = _m[1][0], f = _m[1][1], g = _m[1][2], h = _m[1][3],
+                i = _m[2][0], j = _m[2][1], k = _m[2][2], l = _m[2][3],
+                m = _m[3][0], n = _m[3][1], o = _m[3][2], p = _m[3][3];
+        return {
+            a, e, i, m,
+            b, f, j, n,
+            c, g, k, o,
+            d, h, l, p};
     }
 
     //
     // adjugate
     //
 
-    namespace impl
-    {
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 2> adjugate_2x2_impl(
-            T a, T b,
-            T c, T d)
-        {
-            return {
-                +d, -b,
-                -c, +a};
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 3> adjugate_3x3_impl(
-            T a, T b, T c,
-            T d, T e, T f,
-            T g, T h, T i)
-        {
-            return {
-                e * i - f * h,
-                c * h - b * i,
-                b * f - c * e,
-                f * g - d * i,
-                a * i - c * g,
-                c * d - a * f,
-                d * h - e * g,
-                b * g - a * h,
-                a * e - b * d};
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        mat<T, 4> adjugate_4x4_impl(
-            T a, T b, T c, T d,
-            T e, T f, T g, T h,
-            T i, T j, T k, T l,
-            T m, T n, T o, T p)
-        {
-            return {
-                f * (k * p - l * o) + g * (l * n - j * p) + h * (j * o - k * n),
-                j * (c * p - d * o) + k * (d * n - b * p) + l * (b * o - c * n),
-                n * (c * h - d * g) + o * (d * f - b * h) + p * (b * g - c * f),
-                b * (h * k - g * l) + c * (f * l - h * j) + d * (g * j - f * k),
-                g * (i * p - l * m) + h * (k * m - i * o) + e * (l * o - k * p),
-                k * (a * p - d * m) + l * (c * m - a * o) + i * (d * o - c * p),
-                o * (a * h - d * e) + p * (c * e - a * g) + m * (d * g - c * h),
-                c * (h * i - e * l) + d * (e * k - g * i) + a * (g * l - h * k),
-                h * (i * n - j * m) + e * (j * p - l * n) + f * (l * m - i * p),
-                l * (a * n - b * m) + i * (b * p - d * n) + j * (d * m - a * p),
-                p * (a * f - b * e) + m * (b * h - d * f) + n * (d * e - a * h),
-                d * (f * i - e * j) + a * (h * j - f * l) + b * (e * l - h * i),
-                e * (k * n - j * o) + f * (i * o - k * m) + g * (j * m - i * n),
-                i * (c * n - b * o) + j * (a * o - c * m) + k * (b * m - a * n),
-                m * (c * f - b * g) + n * (a * g - c * e) + o * (b * e - a * f),
-                a * (f * k - g * j) + b * (g * i - e * k) + c * (e * j - f * i)};
-        }
+    template < typename T >
+    [[nodiscard]] constexpr mat<T, 2> adjugate(const mat<T, 2>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1],
+                c = _m[1][0], d = _m[1][1];
+        return {
+            +d,
+            -b,
+            -c,
+            +a};
     }
 
     template < typename T >
-    [[nodiscard]] constexpr mat<T, 2> adjugate(const mat<T, 2>& m) {
-        return impl::adjugate_2x2_impl(
-            m[0][0], m[0][1],
-            m[1][0], m[1][1]);
+    [[nodiscard]] constexpr mat<T, 3> adjugate(const mat<T, 3>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2],
+                d = _m[1][0], e = _m[1][1], f = _m[1][2],
+                g = _m[2][0], h = _m[2][1], i = _m[2][2];
+        return {
+            e * i - f * h,
+            c * h - b * i,
+            b * f - c * e,
+            f * g - d * i,
+            a * i - c * g,
+            c * d - a * f,
+            d * h - e * g,
+            b * g - a * h,
+            a * e - b * d};
     }
 
     template < typename T >
-    [[nodiscard]] constexpr mat<T, 3> adjugate(const mat<T, 3>& m) {
-        return impl::adjugate_3x3_impl(
-            m[0][0], m[0][1], m[0][2],
-            m[1][0], m[1][1], m[1][2],
-            m[2][0], m[2][1], m[2][2]);
-    }
-
-    template < typename T >
-    [[nodiscard]] constexpr mat<T, 4> adjugate(const mat<T, 4>& m) {
-        return impl::adjugate_4x4_impl(
-            m[0][0], m[0][1], m[0][2], m[0][3],
-            m[1][0], m[1][1], m[1][2], m[1][3],
-            m[2][0], m[2][1], m[2][2], m[2][3],
-            m[3][0], m[3][1], m[3][2], m[3][3]);
+    [[nodiscard]] constexpr mat<T, 4> adjugate(const mat<T, 4>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2], d = _m[0][3],
+                e = _m[1][0], f = _m[1][1], g = _m[1][2], h = _m[1][3],
+                i = _m[2][0], j = _m[2][1], k = _m[2][2], l = _m[2][3],
+                m = _m[3][0], n = _m[3][1], o = _m[3][2], p = _m[3][3];
+        return {
+            f * (k * p - l * o) + g * (l * n - j * p) + h * (j * o - k * n),
+            j * (c * p - d * o) + k * (d * n - b * p) + l * (b * o - c * n),
+            n * (c * h - d * g) + o * (d * f - b * h) + p * (b * g - c * f),
+            b * (h * k - g * l) + c * (f * l - h * j) + d * (g * j - f * k),
+            g * (i * p - l * m) + h * (k * m - i * o) + e * (l * o - k * p),
+            k * (a * p - d * m) + l * (c * m - a * o) + i * (d * o - c * p),
+            o * (a * h - d * e) + p * (c * e - a * g) + m * (d * g - c * h),
+            c * (h * i - e * l) + d * (e * k - g * i) + a * (g * l - h * k),
+            h * (i * n - j * m) + e * (j * p - l * n) + f * (l * m - i * p),
+            l * (a * n - b * m) + i * (b * p - d * n) + j * (d * m - a * p),
+            p * (a * f - b * e) + m * (b * h - d * f) + n * (d * e - a * h),
+            d * (f * i - e * j) + a * (h * j - f * l) + b * (e * l - h * i),
+            e * (k * n - j * o) + f * (i * o - k * m) + g * (j * m - i * n),
+            i * (c * n - b * o) + j * (a * o - c * m) + k * (b * m - a * n),
+            m * (c * f - b * g) + n * (a * g - c * e) + o * (b * e - a * f),
+            a * (f * k - g * j) + b * (g * i - e * k) + c * (e * j - f * i)};
     }
 
     //
     // determinant
     //
 
-    namespace impl
-    {
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        T determinant_2x2_impl(
-            T a, T b,
-            T c, T d)
-        {
-            return a * d - b * c;
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        T determinant_3x3_impl(
-            T a, T b, T c,
-            T d, T e, T f,
-            T g, T h, T i)
-        {
-            return
-                + a * (e * i - f * h)
-                - b * (d * i - f * g)
-                + c * (d * h - e * g);
-        }
-
-        template < typename T >
-        [[nodiscard]] constexpr VMATH_HPP_FORCE_INLINE
-        T determinant_4x4_impl(
-            T a, T b, T c, T d,
-            T e, T f, T g, T h,
-            T i, T j, T k, T l,
-            T m, T n, T o, T p)
-        {
-            return
-                + a * (f * (k * p - l * o) - (j * (g * p - h * o)) + (n * (g * l - h * k)))
-                - b * (e * (k * p - l * o) - (i * (g * p - h * o)) + (m * (g * l - h * k)))
-                + c * (e * (j * p - l * n) - (i * (f * p - h * n)) + (m * (f * l - h * j)))
-                - d * (e * (j * o - k * n) - (i * (f * o - g * n)) + (m * (f * k - g * j)));
-        }
+    template < typename T >
+    [[nodiscard]] constexpr T determinant(const mat<T, 2>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1],
+                c = _m[1][0], d = _m[1][1];
+        return
+            a * d - b * c;
     }
 
     template < typename T >
-    [[nodiscard]] constexpr T determinant(const mat<T, 2>& m) {
-        return impl::determinant_2x2_impl(
-            m[0][0], m[0][1],
-            m[1][0], m[1][1]);
+    [[nodiscard]] constexpr T determinant(const mat<T, 3>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2],
+                d = _m[1][0], e = _m[1][1], f = _m[1][2],
+                g = _m[2][0], h = _m[2][1], i = _m[2][2];
+        return
+            + a * (e * i - f * h)
+            - b * (d * i - f * g)
+            + c * (d * h - e * g);
     }
 
     template < typename T >
-    [[nodiscard]] constexpr T determinant(const mat<T, 3>& m) {
-        return impl::determinant_3x3_impl(
-            m[0][0], m[0][1], m[0][2],
-            m[1][0], m[1][1], m[1][2],
-            m[2][0], m[2][1], m[2][2]);
-    }
-
-    template < typename T >
-    [[nodiscard]] constexpr T determinant(const mat<T, 4>& m) {
-        return impl::determinant_4x4_impl(
-            m[0][0], m[0][1], m[0][2], m[0][3],
-            m[1][0], m[1][1], m[1][2], m[1][3],
-            m[2][0], m[2][1], m[2][2], m[2][3],
-            m[3][0], m[3][1], m[3][2], m[3][3]);
+    [[nodiscard]] constexpr T determinant(const mat<T, 4>& _m) {
+        // NOLINTNEXTLINE(*-isolate-declaration)
+        const T a = _m[0][0], b = _m[0][1], c = _m[0][2], d = _m[0][3],
+                e = _m[1][0], f = _m[1][1], g = _m[1][2], h = _m[1][3],
+                i = _m[2][0], j = _m[2][1], k = _m[2][2], l = _m[2][3],
+                m = _m[3][0], n = _m[3][1], o = _m[3][2], p = _m[3][3];
+        return
+            + a * (f * (k * p - l * o) - (j * (g * p - h * o)) + (n * (g * l - h * k)))
+            - b * (e * (k * p - l * o) - (i * (g * p - h * o)) + (m * (g * l - h * k)))
+            + c * (e * (j * p - l * n) - (i * (f * p - h * n)) + (m * (f * l - h * j)))
+            - d * (e * (j * o - k * n) - (i * (f * o - g * n)) + (m * (f * k - g * j)));
     }
 
     //