---
title: Sub
description: "向量与向量相减。"
url: https://www.hikunpeng.com/document/detail/zh/kunpengboostkithistory/2300/accel/kunpengaccel_hmpp_06_0095.html
sourcePath: /source/zh/kunpengboostkithistory/2300/accel/kunpengaccel_hmpp_06_0095.html
indexId: 7ec3714cd2ad6ffb3412815877b5e46ba1284a9319228eed321e86b7d9b13f3b75
---
# Sub

向量与向量相减。

函数接口声明如下：

- 整型数的操作：
  HmppResult HMPPS_Sub_16s(const int16_t *src1, const int16_t *src2, int16_t *dst, int32_t len);

- 浮点数的操作：
  HmppResult HMPPS_Sub_32f(const float *src1, const float *src2, float *dst, int32_t len);

  HmppResult HMPPS_Sub_64f(const double *src1, const double *src2, double *dst, int32_t len);

  HmppResult HMPPS_Sub_32fc(const Hmpp32fc *src1, const Hmpp32fc *src2, Hmpp32fc *dst, int32_t len);

  HmppResult HMPPS_Sub_16s32f(const int16_t *src1, const int16_t *src2, float *dst, int32_t len);

  HmppResult HMPPS_Sub_64fc(const Hmpp64fc *src1, const Hmpp64fc *src2, Hmpp64fc *dst, int32_t len);

- 有缩放的整型数操作：
  HmppResult HMPPS_Sub_16s_S(const int16_t *src1, const int16_t *src2, int16_t *dst, int32_t len, double scale);

  HmppResult HMPPS_Sub_32s_S(const int32_t *src1, const int32_t *src2, int32_t *dst, int32_t len, double scale);

  HmppResult HMPPS_Sub_8u_S(const uint8_t *src1, const uint8_t *src2, uint8_t *dst, int32_t len, double scale);

  HmppResult HMPPS_Sub_16u_S(const uint16_t *src1, const uint16_t *src2, uint16_t *dst, int32_t len, double scale);

  HmppResult HMPPS_Sub_16sc_S(const Hmpp16sc *src1, const Hmpp16sc *src2, Hmpp16sc *dst, int32_t len, double scale);

  HmppResult HMPPS_Sub_32sc_S(const Hmpp32sc *src1, const Hmpp32sc *src2, Hmpp32sc *dst, int32_t len, int32_t scale);

- 整型数的原址操作：
  HmppResult HMPPS_Sub_16s_I(const int16_t *src, int16_t *srcDst, int32_t len);

- 浮点数的原址操作：
  HmppResult HMPPS_Sub_32f_I(const float *src, float *srcDst, int32_t len);

  HmppResult HMPPS_Sub_64f_I(const double *src, double *srcDst, int32_t len);

  HmppResult HMPPS_Sub_32fc_I(const Hmpp32fc *src, Hmpp32fc *srcDst, int32_t len);

  HmppResult HMPPS_Sub_64fc_I(const Hmpp64fc *src, Hmpp64fc *srcDst, int32_t len);

- 有缩放的整型数原址操作：
  HmppResult HMPPS_Sub_16s_IS(const int16_t *src1, int16_t *srcDst, int32_t len, double scale);

  HmppResult HMPPS_Sub_32s_IS(const int32_t *src1, int32_t *srcDst, int32_t len, doublescale);

  HmppResult HMPPS_Sub_8u_IS(const uint8_t *src, uint8_t *srcDst, int32_t len, double scale);

  HmppResult HMPPS_Sub_16u_IS(const uint16_t *src, uint16_t *srcDst, int32_t len, doublescale);

  HmppResult HMPPS_Sub_16sc_IS(const Hmpp16sc *src, Hmpp16sc *srcDst, int32_t len, double scale);

  HmppResult HMPPS_Sub_32sc_IS(const Hmpp32sc *src, Hmpp32sc *srcDst, int32_t len, double scale);


#### 参数

| 参数名 | 描述 | 取值范围 | 输入/输出 |
| --- | --- | --- | --- |
| src1 | 指向减数的指针。 | 非空 | 输入 |
| src2 | 指向被减数的指针。 | 非空 | 输入 |
| dst | 指向目的向量的指针。 | 非空 | 输出 |
| src | 指向原址操作中减数的指针。 | 非空 | 输入 |
| srcDst | 指向原址操作向量中被减数的指针。 | 非空 | 输入/输出 |
| len | 向量长度。 | (0,INT\_MAX] | 输入 |
| scale | 缩放因数。 | (0,INF)且输入为2^n | 输入 |


#### 返回值

- 成功：返回HMPP_STS_NO_ERR。
- 失败：返回错误码。


#### 错误码

| 错误码 | 描述 |
| --- | --- |
| HMPP\_STS\_NULL\_PTR\_ERR | src1、src2、dst、src、srcDst这几个入参中存在空指针。 |
| HMPP\_STS\_SIZE\_ERR | len小于或等于0。 |
| HMPP\_STS\_SCALE\_ERR | scale不在(0,INF)范围内或输入为nan。 |


#### 示例

```
#define BUFFER_SIZE_T 10
void SubExample(void)
{
float src1[BUFFER_SIZE_T] = { -1.40, 10.85, 7.26, 8.98, 0.26, 6.47, 10.42, 8.64, 3.22, 10.90};
float src2[BUFFER_SIZE_T] = { 10.11, 4.27, 0.53, 4.00, 4.73, -1.23, 6.44, 8.88, 0.78, 5.33};
float dst[BUFFER_SIZE_T];
(void)HMPPS_Zero_32f(dst, BUFFER_SIZE_T);
HmppResult result = HMPPS_Sub_32f(src1, src2, dst, BUFFER_SIZE_T);
printf("result = %d\n", result);
if (result != HMPP_STS_NO_ERR) {
return;
}
printf("dst =");
for (int i = 0; i < BUFFER_SIZE_T; i++) {
printf(" %.2f", dst[i]);
}
printf("\n");
}
```

运行结果：

```
sub: result = 0.
dst = 11.51    -6.58    -6.73    -4.98    4.47    -7.70    -3.98    0.24    -2.44    -5.57
```