Dear authors @juliusberner ,
Very nice work. Thank you for the code.
We are able to reproduce your result for the Black-Scholes model (Table 1), heat equation with paraboloid initial condition (Table 3), and Gaussian initial condition (Table 4). However, we have a problem with the Basket option (Table 2). The time for gradient step 4k is not within 811s as the table suggested. In fact, we didn't even get any results even though we kept the code run for 12 hours.
We guess the main reason perhaps due to the two inner loops in that the code for the simulation of the solution (
|
for t, x, sigma, mu, K in zip( |
)
Do you think we can remove this part and make it parallel?
I wonder if you already save some of the simulated data that we could directly use. Thank you very much!