This PR is to solve the issue with solve and inv due to problems in paralellization.
For solve: Disable the parallelism and use normal loop approach for the moment until the issue is resolved.
For inv: The approach to calculate inverse is to solve the system AB=I, where I is an identity matrix, with the function solve. This is also what Numpy does.
For matmul: It seems to be related to ASAP destruction policy and can be resolved by using the variables after the paralellization.
This PR is to solve the issue with
solve
andinv
due to problems in paralellization.For
solve
: Disable the parallelism and use normal loop approach for the moment until the issue is resolved.For
inv
: The approach to calculate inverse is to solve the systemAB=I
, whereI
is an identity matrix, with the functionsolve
. This is also what Numpy does.For
matmul
: It seems to be related to ASAP destruction policy and can be resolved by using the variables after the paralellization.