ecsimsw

constructor ary = new int*[count_rows]; for (int i = 0; i

/// matmul_op.cc #endif // GOOGLE_CUDA template class MatMulOp : public OpKernel { public: explicit MatMulOp(OpKernelConstruction* ctx) : OpKernel(ctx), algorithms_set_already_(false) { OP_REQUIRES_OK(ctx, ctx->GetAttr("transpose_a", &transpose_a_)); OP_REQUIRES_OK(ctx, ctx->GetAttr("transpose_b", &transpose_b_)); LaunchMatMul::GetBlasGemmAlgorithm( ctx, &algorithms_, &algorithms_set_already_); ..

/// matmul_op.cc 1468 REGISTER_KERNEL_BUILDER #if defined(INTEL_MKL) // math kernel library TF_CALL_float(REGISTER_CPU_EIGEN); #else TF_CALL_float(REGISTER_CPU); #define REGISTER_CPU_EIGEN(T) \ REGISTER_KERNEL_BUILDER( \ Name("MatMul").Device(DEVICE_CPU).TypeConstraint("T").Label("eigen"), \ MatMulOp); #define REGISTER_CPU(T) \ REGISTER_KERNEL_BUILDER( \ Name("MatMul").Device(DEVICE_CPU).TypeCon..

암시적으로 생성자가 호출되면서 발생할 수 있는 의도치 않은 형 변환을 막기 위해 생성자의 명시적 호출을 강제하기 위한 키워드

/// REGISTER_KERNEL_BUILDER(Name("SparseApplyAdadelta") \ .Device(DEVICE_CPU) \ .TypeConstraint("T") \ .TypeConstraint("Tindices"), \ SparseApplyAdadeltaOp); /// #define REGISTER_KERNEL_BUILDER_UNIQ(ctr, kernel_builder, ...) \ constexpr bool should_register_##ctr##__flag = \ SHOULD_REGISTER_OP_KERNEL(#__VA_ARGS__); \ static ::tensorflow::kernel_factory::OpKernelRegistrar \ registrar__body__##ctr..

http://blog.naver.com/PostView.nhn?blogId=wkdghcjf1234&logNo=220210906503 형 변환 연산자: static_cast와 reinterpret_cast의 차이 이번에는 간단하게 static_cast와 reinterpret_cast의 차이점에 대하여 알아봅니다. 실제로 이 두가지는 비... blog.naver.com 다시 정리

stackexchange_ What's the difference between user registers and kernel registers? "Answer" It's simple - when each application program runs, it has access to its own set of registers. When you switch to other application, these register contents is saved to memory, and registers, saved from other application, loaded and this application continues its execution. Similarly, OS has its own register..