Add LTO in configure, add more benchmarks

bench/tests/BenchVec2f.c

@@ -0,0 +1,204 @@
+#include "BH/Bench.h"
+#include "BH/Math/Vec2f.h"
+
+
+static float vecA[2] = {1.0f, 2.0f};
+static float vecB[2] = {3.0f, -1.0f};
+static float vecC[2] = {0.5f, 4.0f};
+static float vecZero[2] = {0.0f, 0.0f};
+static float vecLarge[2] = {1e6f, -1e6f};
+static float vecOut[2];
+
+
+BH_BENCH_TEST(Add)
+{
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fAdd(vecA, vecB, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(Sub)
+{
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fSub(vecA, vecB, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(Mul)
+{
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fMul(vecA, vecB, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(Scale)
+{
+    const float scale = 2.5f;
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fScale(vecA, scale, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(MulAdd)
+{
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fMulAdd(vecA, vecB, vecC, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(Negate)
+{
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fNegate(vecA, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(Dot)
+{
+    float result;
+    while (BH_BenchIter(state))
+    {
+        result = BH_Vec2fDot(vecA, vecB);
+        BH_BenchDoNotOptimize(result);
+    }
+}
+
+
+BH_BENCH_TEST(Cross)
+{
+    float result;
+    while (BH_BenchIter(state))
+    {
+        result = BH_Vec2fCross(vecA, vecB);
+        BH_BenchDoNotOptimize(result);
+    }
+}
+
+
+BH_BENCH_TEST(Length)
+{
+    float result;
+    while (BH_BenchIter(state))
+    {
+        result = BH_Vec2fLength(vecA);
+        BH_BenchDoNotOptimize(result);
+    }
+}
+
+
+BH_BENCH_TEST(Normal)
+{
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fNormal(vecA, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(NormalEx)
+{
+    float length;
+    while (BH_BenchIter(state))
+    {
+        length = BH_Vec2fNormalEx(vecA, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+        BH_BenchDoNotOptimize(length);
+    }
+}
+
+
+BH_BENCH_TEST(Min)
+{
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fMin(vecA, vecB, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(Max)
+{
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fMax(vecA, vecB, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(Lerp)
+{
+    const float t = 0.3f;
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fLerp(vecA, vecB, t, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(Project)
+{
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fProject(vecA, vecB, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+BH_BENCH_TEST(Barycentric)
+{
+    const float v = 0.2f, w = 0.3f;
+    while (BH_BenchIter(state))
+    {
+        BH_Vec2fBarycentric(vecA, vecB, vecC, v, w, vecOut);
+        BH_BenchDoNotOptimize(vecOut);
+    }
+}
+
+
+int main(int argc, char **argv)
+{
+    BH_UNUSED(argc);
+    BH_UNUSED(argv);
+
+    BH_BENCH_ADD(Add);
+    BH_BENCH_ADD(Sub);
+    BH_BENCH_ADD(Mul);
+    BH_BENCH_ADD(Scale);
+    BH_BENCH_ADD(MulAdd);
+    BH_BENCH_ADD(Negate);
+    BH_BENCH_ADD(Dot);
+    BH_BENCH_ADD(Cross);
+    BH_BENCH_ADD(Length);
+    BH_BENCH_ADD(Normal);
+    BH_BENCH_ADD(NormalEx);
+    BH_BENCH_ADD(Min);
+    BH_BENCH_ADD(Max);
+    BH_BENCH_ADD(Lerp);
+    BH_BENCH_ADD(Project);
+    BH_BENCH_ADD(Barycentric);
+
+    return BH_BenchRun();
+}