vadd/vsub/vmul/vdiv

功能说明

计算每个向量元素的加法、减法、乘法或除法，计算公式如下：

// vadd
[dst] = [src0] + [src1]
// vsub
[dst] = [src0] - [src1]
// vmul
[dst] = [src0] * [src1]
// vdiv
[dst] = [src0] / [src1]

vdiv 中的 src1 不应该为零，否则结果未知，产生异常。

以 block（32Byte）为单位完成计算，一次完成8个 block 的计算。

上述接口均支持通过MASK控制哪些元素参与计算。

接口原型

// 相同接口的不同原型区别在于源地址和目的地址的数据类型不同
// vadd
void vadd(__ubuf__ half *dst, __ubuf__ half *src0, __ubuf__ half *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
void vadd(__ubuf__ float *dst, __ubuf__ float *src0, __ubuf__ float *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
void vadd(__ubuf__ int16_t *dst, __ubuf__ int16_t *src0, __ubuf__ int16_t *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
void vadd(__ubuf__ int32_t *dst, __ubuf__ int32_t *src0, __ubuf__ int32_t *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride);

// vsub 
void vsub(__ubuf__ half *dst, __ubuf__ half *src0, __ubuf__ half *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
void vsub(__ubuf__ float *dst, __ubuf__ float *src0, __ubuf__ float *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
void vsub(__ubuf__ int16_t *dst, __ubuf__ int16_t *src0, __ubuf__ int16_t *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
void vsub(__ubuf__ int32_t *dst, __ubuf__ int32_t *src0, __ubuf__ int32_t *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
// vmul 
void vmul(__ubuf__ half *dst, __ubuf__ half *src0, __ubuf__ half *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
void vmul(__ubuf__ float *dst, __ubuf__ float *src0, __ubuf__ float *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
void vmul(__ubuf__ int16_t *dst, __ubuf__ int16_t *src0, __ubuf__ int16_t *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
void vmul(__ubuf__ int32_t *dst, __ubuf__ int32_t *src0, __ubuf__ int32_t *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
// vdiv 
void vdiv(__ubuf__ half *dst, __ubuf__ half *src0, __ubuf__ half *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride); 
 
void vdiv(__ubuf__ float *dst, __ubuf__ float *src0, __ubuf__ float *src1, uint8_t repeat, uint8_t dstBlockStride, uint8_t src0BlockStride, uint8_t src1BlockStride, uint8_t dstRepeatStride, uint8_t src0RepeatStride, uint8_t src1RepeatStride);

参数说明

参数含义见表2 双目运算参数说明。

流水类型

PIPE_V

父主题： 双目运算