Image Details
 
                    Caption: Figure 20.
Number of accurate digits in 64-bit and 128-bit summations for ﹩\dot{Y}{(}^{4}\mathrm{He})﹩ as measured by the 100 digit sum calculated by the multiple precision packages. The x-axis gives the timestep number for the integration done with IEEE 128-bit summations. The number of accurate digits in ﹩\dot{Y}(p)﹩ and ﹩\dot{Y}(n)﹩ are within a few digits of ﹩\dot{Y}{(}^{4}\mathrm{He})﹩.
© 2015. The American Astronomical Society. All rights reserved.