Skip to main content
Greg Hurst's user avatar
Greg Hurst's user avatar
Greg Hurst's user avatar
Greg Hurst
  • Member for 8 years, 6 months
  • Last seen more than a week ago
awarded
revised
Loading…
revised
Loading…
revised
Loading…
revised
Loading…
revised
Loading…
comment
Reference for accelerated sum to compute the Meissel-Mertens constant
@mathworker21 But it's an interesting point. This series is a blend between equations $(7)$ and $(3)$ in the link I provided. When $N = 0$ we have equation $(7)$ and as $N \to \infty$ we have equation $(3)$.
comment
Reference for accelerated sum to compute the Meissel-Mertens constant
Yes, and that's a good point. If we let $N \to \infty$ in the formula in my question, we do get equation $(3)$. However there seems to be a 'sweet-spot' for the value of $N$. If $N$ is too small or too large the sum converges slower (implementation dependent values of $N$ course). For my implementation I used $N = 16$.
Loading…
awarded
awarded
Loading…
Loading…
revised
Loading…
Loading…
revised
Loading…
awarded
Loading…
comment
Is there a known asymptotic for $A(X):= \sum_{1 \leq i,j \leq X} \frac{1}{\mathrm{lcm}(i,j)}$?
Your estimate for $A_0$ leads to division by $0$. I think you mean to have $A_j = \frac{1}{(j+1)\zeta(2)} \log^{j+1} X + O(\log^j X)$.