But running BKZ-20 on such a high dimension is going to be very expensive, so it might be a better choice to switch to LLL and increase d to, say, 600, in which case we have 1.021l+dql/(l+d)≈2129.22, which is tight, but within range of being solved. The following graph plots...