Function runMultiplyBenchmark

examples/AutoResearch/AutoResearchSimd.h:1238–1270

};

inline BenchmarkResult runMultiplyBenchmark(int iters = 10000) {
    BenchmarkResult r;
    r.iterations = iters;

    // Add
    r.add_float_us   = benchFloat4(iters, OpAdd());
    r.add_s8x8_us    = benchS8x8_4(iters, OpAdd());
    r.add_s16x16_us  = benchS16x16_4(iters, OpAdd());
    r.add_u16x16_us  = benchU16x16_4(iters, OpAdd());
    r.add_simd_us    = benchSimd4(iters, OpAdd());

    // Sub
    r.sub_float_us   = benchFloat4(iters, OpSub());
    r.sub_s8x8_us    = benchS8x8_4(iters, OpSub());
    r.sub_s16x16_us  = benchS16x16_4(iters, OpSub());
    r.sub_u16x16_us  = benchU16x16_4(iters, OpSub());
    r.sub_simd_us    = benchSimd4(iters, OpSub());

    // Mul
    r.mul_float_us   = benchFloat4(iters, OpMul());
    r.mul_s8x8_us    = benchS8x8_4(iters, OpMul());
    r.mul_s16x16_us  = benchS16x16_4(iters, OpMul());
    r.mul_u16x16_us  = benchU16x16_4(iters, OpMul());
    r.mul_simd_us    = benchSimd4(iters, OpMul());

    // Div (no SIMD div for s16x16x4)
    r.div_float_us   = benchFloat4(iters, OpDivFloat());
    r.div_s8x8_us    = benchS8x8_4(iters, OpDivS8x8());
    r.div_s16x16_us  = benchS16x16_4(iters, OpDivS16x16());
    r.div_u16x16_us  = benchU16x16_4(iters, OpDivU16x16());

    return r;
}

// ============================================================================
// Test Runner