Method Compute

en_ops/e_where_gequal.cc:111–159

        OP_REQUIRES(context, lib != NULL, errors::Unknown("Unable to load sgx.so!"));
    }
    void Compute(OpKernelContext *context) override
    {

        const Tensor &grad = context->input(0);
        auto grad_flat = grad.flat<float>();

        const Tensor &input1 = context->input(1);
        auto input1_flat = input1.flat<float>();

        const Tensor &input2 = context->input(2);
        auto input2_flat = input2.flat<float>();

        const Tensor &cond1 = context->input(3);
        auto cond1_flat = cond1.flat<float>();

        const Tensor &cond2 = context->input(4);
        auto cond2_flat = cond2.flat<float>();

        const TensorShape &input1_shape = input1.shape();
        const TensorShape &input2_shape = input2.shape();

        Tensor *output1 = NULL;
        OP_REQUIRES_OK(context, context->allocate_output(0, input1_shape, &output1));
        auto output1_flat = output1->flat<float>();

        Tensor *output2 = NULL;
        OP_REQUIRES_OK(context, context->allocate_output(1, input2_shape, &output2));
        auto output2_flat = output2->flat<float>();
        int N = input1_flat.size() / times_;

        int n1 = 1;
        int n2 = 1;
        if (cond1_flat.size() > 1)
        {
            n1 = cond1_flat.size() / times_;
        }
        if (cond2_flat.size() > 1)
        {
            n2 = cond2_flat.size() / times_;
        }

        unsigned long int eid_ = (eid_high_ << 32) + eid_low_;
        typedef void (*function)(unsigned long int eid, float *grad, float *input1, float *input2, float *cond1, float *cond2, int N, int n1, int n2, float *output1, float *output2);
        dlerror();
        function where_gequal_grad_kernel = (function)dlsym(lib, "where_gequal_grad");
        const char *dlsym_error = dlerror();
        OP_REQUIRES(context, !dlsym_error, errors::Unknown("loading of where_gequal_grad failed: ", dlsym_error));
        where_gequal_grad_kernel(eid_, (float *)grad_flat.data(), (float *)input1_flat.data(), (float *)input2_flat.data(), (float *)cond1_flat.data(), (float *)cond2_flat.data(), N, n1, n2, (float *)output1_flat.data(), (float *)output2_flat.data());
    };

private:
    void *lib;