android - 优化 NEON 装配功能

Question

我正在开发一个应该在 ARMv7 处理器设备上运行的原生 Android 应用程序。由于某些原因，我需要对 vector (短和/或浮点)进行一些繁重的计算。我使用 NEON 命令实现了一些汇编函数来提高计算量。我获得了 1.5 的速度系数，这还不错。我想知道我是否可以改进这些功能以使其运行得更快。

那么问题是:我可以做哪些改变来改进这些功能？

    //add to float vectors.
//the result could be put in scr1 instead of dst
void add_float_vector_with_neon3(float* dst, float* src1, float* src2, int count)
{

    asm volatile (
           "1:                                                        \n"
           "vld1.32         {q0}, [%[src1]]!                          \n"
           "vld1.32         {q1}, [%[src2]]!                          \n"
           "vadd.f32        q0, q0, q1                                \n"
           "subs            %[count], %[count], #4                    \n"
           "vst1.32         {q0}, [%[dst]]!                           \n"
           "bgt             1b                                        \n"
           : [dst] "+r" (dst)
           : [src1] "r" (src1), [src2] "r" (src2), [count] "r" (count)
           : "memory", "q0", "q1"
      );
}

//multiply a float vector by a scalar.
//the result could be put in scr1 instead of dst
void mul_float_vector_by_scalar_with_neon3(float* dst, float* src1, float scalar, int count)
{

    asm volatile (

            "vdup.32         q1, %[scalar]                              \n"
            "2:                                                         \n"
            "vld1.32         {q0}, [%[src1]]!                           \n"
            "vmul.f32        q0, q0, q1                                 \n"
            "subs            %[count], %[count], #4                     \n"
            "vst1.32         {q0}, [%[dst]]!                            \n"
            "bgt             2b                                         \n"
            : [dst] "+r" (dst)
            : [src1] "r" (src1), [scalar] "r" (scalar), [count] "r" (count)
            : "memory", "q0", "q1"
      );
}

//add to short vector -> no problem of coding limits
//the result should be put in in a dest different from src1 and scr2
void add_short_vector_with_neon3(short* dst, short* src1, short* src2, int count)
{

    asm volatile (
           "3:                                                        \n"
           "vld1.16         {q0}, [%[src1]]!                          \n"
           "vld1.16         {q1}, [%[src2]]!                          \n"
           "vadd.i16        q0, q0, q1                                \n"
           "subs            %[count], %[count], #8                    \n"
           "vst1.16         {q0}, [%[dst]]!                           \n"
           "bgt             3b                                        \n"
           : [dst] "+r" (dst)
           : [src1] "r" (src1), [src2] "r" (src2), [count] "r" (count)
           : "memory", "q0", "q1"
      );
}

//multiply a short vector by a float vector and put the result bach into a short vector
//the result should be put in in a dest different from src1
void mul_short_vector_by_float_vector_with_neon3(short* dst, short* src1, float* src2, int count)
{
    asm volatile (
        "4:                                                         \n"
        "vld1.16        {d0}, [%[src1]]!                            \n"
        "vld1.32        {q1}, [%[src2]]!                            \n"
        "vmovl.s16      q0, d0                                      \n"
        "vcvt.f32.s32   q0, q0                                      \n"
        "vmul.f32       q0, q0, q1                                  \n"
        "vcvt.s32.f32   q0, q0                                      \n"
        "vmovn.s32      d0, q0                                      \n"
        "subs            %[count], %[count], #4                     \n"
        "vst1.16         {d0}, [%[dst]]!                            \n"
        "bgt             4b                                         \n"
        : [dst] "+r" (dst)
        : [src1] "r" (src1), [src2] "r" (src2), [count] "r" (count)
        : "memory", "d0", "q0", "q1"

    );
}

提前致谢!

Answer 1

您可以尝试展开循环以在每个循环中处理更多元素。

您的 add_float_vector_with_neon3 代码每 4 个元素需要 10 个周期(因为停顿)，而展开到 16 个元素需要 21 个周期。 http://pulsar.webshaker.net/ccc/sample-34e5f701

虽然有开销因为你需要处理余数(或者你可以填充你的数据为 16 的倍数)，但如果你有很多数据，与实际总和相比开销应该相当低。