n

with precision 1e-10. Please use it if such accuracy is acceptable in your task.

If you need high precision tables for the such long list of different

n

, please post brief description of your project here. This is only fair price for the enormous chunk of work you are asking me to do.

n=32,64,100

to the table and source code.

Besides source code now includes quadratures for

n=96,128,256,512,1024

.

One note though. It could be pointless to apply high order quadratures using ‘double’ floating point numbers due its limited precision. Machine epsilon for double is around 1e-16. Usually Gaussian quadrature attain such error quickly even for relatively small

n\le20

. Thus higher orders will not improve error any further. In the same time rounding error will increase with higher order.

It has sense to use high-precision floating point numbers instead of ‘double’. Libraries GMP and MPFR are commonly used for this purpose. I’ve created C++ interface for MPFR to simplify its usage. If you interested please visit MPFR C++ page.