在使用bang语言编写一个relu算子的时候报错：

其中device侧的relu算子声明在头文件relu_api中，定义在relu.mlu文件中 并且在relu_test.cpp引入这个头文件中进行调用，具体的代码和makefile如下：

relu_test.cpp:

#include <stdio.h>

#include <stdlib.h>

#include "relu_api.h"

#include "cnrt.h"

void relu_mlu(float* dst, float* source, int len) {

    cnrtQueue_t queue;

    CNRT_CHECK(cnrtSetDevice(0));

    CNRT_CHECK(cnrtQueueCreate(&queue));

    cnrtDim3_t dim = {1,1,1};

    cnrtFunctionType_t ktype = CNRT_FUNC_TYPE_BLOCK;

    float* mlu_dst;

    float* mlu_src;

    CNRT_CHECK(cnrtMalloc((void**)&mlu_dst, len * sizeof(float)));

    CNRT_CHECK(cnrtMalloc((void**)&mlu_src, len * sizeof(float)));

    CNRT_CHECK(cnrtMemcpy(mlu_src, source, len * sizeof(float), cnrtMemcpyHostToDev));

    relu_kernel<<<dim, ktype, queue>>>(mlu_dst, mlu_src, len);

    cnrtQueueSync(queue);

    CNRT_CHECK(cnrtMemcpy(dst, mlu_dst, len * sizeof(float), cnrtMemcpyDevToHost));

    CNRT_CHECK(cnrtQueueDestroy(queue));

    cnrtFree(mlu_dst);

    cnrtFree(mlu_src);

    return;

}

int main() {

    int test_len = 5;

    float* host_dst = (float*)malloc(test_len * sizeof(float));

    float* host_src = (float*)malloc(test_len * sizeof(float));

    host_src[0] = -2.0;host_src[1]= -1.0;host_src[2]=0.0;host_src[3]=1.0;host_src[4]=2.0;

    relu_mlu(host_dst, host_src, test_len);

    for (int i = 0; i < test_len; i++) {

        printf("%f ", host_dst[i]);

    }

}

感觉是编译选项或者某些头文件没设置好的问题 求解答