Skip to content

Simplified oneMKL example by using usm_ndarray class for pybind11 #793

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Mar 21, 2022
Merged

Simplified oneMKL example by using usm_ndarray class for pybind11 #793

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
70 changes: 19 additions & 51 deletions examples/pybind11/onemkl_gemv/sycl_gemm/_onemkl.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,16 +6,6 @@

namespace py = pybind11;

/* DPCTL C-API for usm_ndarray
UsmNDArray_GetData
UsmNDArray_GetNDim
UsmNDArray_GetShape
UsmNDArray_GetStrides
UsmNDArray_GetTypenum
UsmNDArray_GetFlags
UsmNDArray_GetQueueRef
*/

sycl::event keep_args_alive(sycl::queue q,
py::object o1,
py::object o2,
Expand Down Expand Up @@ -47,51 +37,29 @@ sycl::event keep_args_alive(sycl::queue q,

std::pair<sycl::event, sycl::event>
gemv(sycl::queue q,
py::object matrix,
py::object vector,
py::object result,
dpctl::tensor::usm_ndarray matrix,
dpctl::tensor::usm_ndarray vector,
dpctl::tensor::usm_ndarray result,
const std::vector<sycl::event> &depends = {})
{
PyObject *m_src = matrix.ptr();
if (!PyObject_TypeCheck(m_src, &PyUSMArrayType)) {
throw std::runtime_error("Matrix is not a dpctl.tensor.usm_ndarray");
}

PyObject *v_src = vector.ptr();
if (!PyObject_TypeCheck(v_src, &PyUSMArrayType)) {
throw std::runtime_error("Vector is not a dpctl.tensor.usm_ndarray");
}

PyObject *r_src = result.ptr();
if (!PyObject_TypeCheck(r_src, &PyUSMArrayType)) {
throw std::runtime_error("Result is not a dpctl.tensor.usm_ndarray");
}

PyUSMArrayObject *m_usm_ary = reinterpret_cast<PyUSMArrayObject *>(m_src);
PyUSMArrayObject *v_usm_ary = reinterpret_cast<PyUSMArrayObject *>(v_src);
PyUSMArrayObject *r_usm_ary = reinterpret_cast<PyUSMArrayObject *>(r_src);

if (UsmNDArray_GetNDim(m_usm_ary) != 2 ||
UsmNDArray_GetNDim(v_usm_ary) != 1 ||
UsmNDArray_GetNDim(r_usm_ary) != 1)
{
if (matrix.get_ndim() != 2 || vector.get_ndim() != 1 ||
result.get_ndim() != 1) {
throw std::runtime_error(
"Inconsistent dimensions, expecting matrix and a vector");
}

py::ssize_t *m_sh = UsmNDArray_GetShape(m_usm_ary);
py::ssize_t n = m_sh[0];
py::ssize_t m = m_sh[1];
py::ssize_t n = matrix.get_shape(0); // get 0-th element of the shape
py::ssize_t m = matrix.get_shape(1);

py::ssize_t *v_sh = UsmNDArray_GetShape(v_usm_ary);
py::ssize_t *r_sh = UsmNDArray_GetShape(r_usm_ary);
if (v_sh[0] != m || r_sh[0] != n) {
py::ssize_t v_dim = vector.get_shape(0);
py::ssize_t r_dim = result.get_shape(0);
if (v_dim != m || r_dim != n) {
throw std::runtime_error("Inconsistent shapes.");
}

int mat_flags = UsmNDArray_GetFlags(m_usm_ary);
int v_flags = UsmNDArray_GetFlags(v_usm_ary);
int r_flags = UsmNDArray_GetFlags(r_usm_ary);
int mat_flags = matrix.get_flags();
int v_flags = vector.get_flags();
int r_flags = result.get_flags();

if (!((mat_flags & (USM_ARRAY_C_CONTIGUOUS | USM_ARRAY_F_CONTIGUOUS)) &&
(v_flags & (USM_ARRAY_C_CONTIGUOUS | USM_ARRAY_F_CONTIGUOUS)) &&
Expand All @@ -100,9 +68,9 @@ gemv(sycl::queue q,
throw std::runtime_error("Arrays must be contiguous.");
}

int mat_typenum = UsmNDArray_GetTypenum(m_usm_ary);
int v_typenum = UsmNDArray_GetTypenum(v_usm_ary);
int r_typenum = UsmNDArray_GetTypenum(r_usm_ary);
int mat_typenum = matrix.get_typenum();
int v_typenum = vector.get_typenum();
int r_typenum = result.get_typenum();

if ((mat_typenum != v_typenum) || (r_typenum != v_typenum) ||
!((v_typenum == UAR_DOUBLE) || (v_typenum == UAR_FLOAT) ||
Expand All @@ -116,9 +84,9 @@ gemv(sycl::queue q,
"Only real and complex floating point arrays are supported.");
}

char *mat_typeless_ptr = UsmNDArray_GetData(m_usm_ary);
char *v_typeless_ptr = UsmNDArray_GetData(v_usm_ary);
char *r_typeless_ptr = UsmNDArray_GetData(r_usm_ary);
char *mat_typeless_ptr = matrix.get_data();
char *v_typeless_ptr = vector.get_data();
char *r_typeless_ptr = result.get_data();

sycl::event res_ev;
if (v_typenum == UAR_DOUBLE) {
Expand Down