static StatusCode RunReference2(const Arguments<U> &args, BuffersHost<T> &buffers_host, Queue&) {
cblasXhpr(convertToCBLAS(args.layout),
convertToCBLAS(args.triangle),
args.n, args.alpha,
buffers_host.x_vec, args.x_offset, args.x_inc,
buffers_host.ap_mat, args.ap_offset);
return StatusCode::kSuccess;
}
static StatusCode RunReference2(const Arguments<T> &args, BuffersHost<T> &buffers_host, Queue &) {
cblasXgbmv(convertToCBLAS(args.layout),
convertToCBLAS(args.a_transpose),
args.m, args.n, args.kl, args.ku, args.alpha,
buffers_host.a_mat, args.a_offset, args.a_ld,
buffers_host.x_vec, args.x_offset, args.x_inc, args.beta,
buffers_host.y_vec, args.y_offset, args.y_inc);
return StatusCode::kSuccess;
}
static StatusCode RunReference2(const Arguments<T> &args, BuffersHost<T> &buffers_host, Queue &) {
cblasXtrmv(convertToCBLAS(args.layout),
convertToCBLAS(args.triangle),
convertToCBLAS(args.a_transpose),
convertToCBLAS(args.diagonal),
args.n,
buffers_host.a_mat, args.a_offset, args.a_ld,
buffers_host.x_vec, args.x_offset, args.x_inc);
return StatusCode::kSuccess;
}
static StatusCode RunReference2(const Arguments<U> &args, Buffers<T> &buffers, Queue &queue) {
std::vector<T> ap_mat_cpu(args.ap_size, static_cast<T>(0));
std::vector<T> x_vec_cpu(args.x_size, static_cast<T>(0));
buffers.ap_mat.Read(queue, args.ap_size, ap_mat_cpu);
buffers.x_vec.Read(queue, args.x_size, x_vec_cpu);
cblasXhpr(convertToCBLAS(args.layout),
convertToCBLAS(args.triangle),
args.n, args.alpha,
x_vec_cpu, args.x_offset, args.x_inc,
ap_mat_cpu, args.ap_offset);
buffers.ap_mat.Write(queue, args.ap_size, ap_mat_cpu);
return StatusCode::kSuccess;
}
static StatusCode RunReference2(const Arguments<T> &args, Buffers<T> &buffers, Queue &queue) {
std::vector<T> a_mat_cpu(args.a_size, static_cast<T>(0));
std::vector<T> x_vec_cpu(args.x_size, static_cast<T>(0));
std::vector<T> y_vec_cpu(args.y_size, static_cast<T>(0));
buffers.a_mat.Read(queue, args.a_size, a_mat_cpu);
buffers.x_vec.Read(queue, args.x_size, x_vec_cpu);
buffers.y_vec.Read(queue, args.y_size, y_vec_cpu);
cblasXsbmv(convertToCBLAS(args.layout),
convertToCBLAS(args.triangle),
args.n, args.kl, args.alpha,
a_mat_cpu, args.a_offset, args.a_ld,
x_vec_cpu, args.x_offset, args.x_inc, args.beta,
y_vec_cpu, args.y_offset, args.y_inc);
buffers.y_vec.Write(queue, args.y_size, y_vec_cpu);
return StatusCode::kSuccess;
}
static StatusCode RunReference2(const Arguments<T> &args, Buffers<T> &buffers, Queue &queue) {
std::vector<T> a_mat_cpu(args.a_size, static_cast<T>(0));
std::vector<T> b_mat_cpu(args.b_size, static_cast<T>(0));
std::vector<T> c_mat_cpu(args.c_size, static_cast<T>(0));
buffers.a_mat.Read(queue, args.a_size, a_mat_cpu);
buffers.b_mat.Read(queue, args.b_size, b_mat_cpu);
buffers.c_mat.Read(queue, args.c_size, c_mat_cpu);
cblasXhemm(convertToCBLAS(args.layout),
convertToCBLAS(args.side),
convertToCBLAS(args.triangle),
args.m, args.n, args.alpha,
a_mat_cpu, args.a_offset, args.a_ld,
b_mat_cpu, args.b_offset, args.b_ld, args.beta,
c_mat_cpu, args.c_offset, args.c_ld);
buffers.c_mat.Write(queue, args.c_size, c_mat_cpu);
return StatusCode::kSuccess;
}