Add new blas extension and update dpnp.matmul func

[vlad-perevezentsev]

This PR updates dpnp.matmul implementation with gemm function call from oneapi::mkl::blas added to dpnp.backend.extensions.blas.
In addition, gemm_batch function is also added which is the batched versions of gemm, performing multiple gemm operations in a single call.
The implementation is written as a pybind11 extension above required BLAS functions.

• Have you provided a meaningful PR description?
• Have you added a test, reproducer or referred to issue with a reproducer?
• Have you tested your changes locally for CPU and GPU devices?
• Have you made sure that new changes do not introduce compiler warnings?
• Have you checked performance impact of proposed changes?
• If this PR is a work in progress, are you filing the PR as a draft?

[github-actions]

View rendered docs @ https://intelpython.github.io/dpnp/index.html

[npolina4]

The result does not match with numpy

importnumpy, dpnpdefinit():
N=5A=numpy.empty((N, N), dtype=numpy.float32)
foriinrange(N):
A[i, : i+1] =numpy.fromfunction(lambdaj: (-j%N) /N+1, (i+1,))
A[i, i+1 :] =0.0A[i, i+1 :] =0.0returnAdefkernel(A):
foriinrange(A.shape[0]):
forjinrange(i):
A[i, j] -=A[i, :j] @ A[:j, j]
a=init()
b=dpnp.asarray(a)
kernel(a)
kernel(b)
print(a-b.asnumpy())

[[0. 0. 0. 0. 0. ]
[0. 0. 0. 0. 0. ]
[0. 0. 0. 0. 0. ]
[0. 0. 0. 0. 0. ]
[0. 0. 0. 1.6 0. ]]

[antonwolfy]

Thank you @vtavana !