import torch;
import habana_frameworks.torch.core as htcore
msize=2
A = torch.randn(msize,msize,dtype=torch.bfloat16).to("hpu")
B = torch.randn(msize,msize,dtype=torch.bfloat16).to("hpu")
C = torch.randn(msize,msize,dtype=torch.bfloat16).to("hpu")
torch.matmul(A,B,out=C)
torch.hpu.synchronize()
R = C.to("cpu")
print(A)
print(B)
print(C)
print(R)