Decimal is Accepted and Next Steps

Matt Borland · Jan 15, 2026

After two reviews the Decimal (https://github.com/cppalliance/decimal) library has been accepted into Boost. Look for it to ship for the first time with Boost 1.91 in the Spring. For current and prospective users, a new release series (v6) is available on the releases page of the library. This major version change contains all of the bug fixes and addresses comments from the second review. We have once again overhauled the documentation based on the review to include a significant increase in the number of examples. Between the Basic Usage and Examples tabs on the website we believe there’s now enough information to quickly make good use of the library. One big quality of life worth highlighting for this version is that it ships with pretty printers for both GDB and LLDB. It is a huge release (1108 commits with a diff stat of >50k LOC), but is be better than ever. I expect that this is the last major version that will be released prior to moving to the Boost release cycle.

Where to go from here?

As I have mentioned in previous posts, the int128 (https://github.com/cppalliance/int128) library started life as the backend for portable arithmetic and representation in the Decimal library. It has since been expanded to include more of the standard library features that are unnecessary as a back-end, but useful to many people like <format> support. The last major update that I intend to make to the library prior to proposal for Boost is to add CUDA support. This would not only add portability to another platform for many users, it would open the door for Decimal to also have CUDA support. I will also be looking at a few of our performance measures as I think there are still places for improvement (such as signed 128-bit division).

Lastly, towards the end of this quarter (March 5 - March 15), I will be serving as the review manager for Alfredo Correa’s Multi (https://github.com/correaa/boost-multi) library. Multi is a modern C++ library that provides manipulation and access of data in multidimensional arrays for both CPU and GPU memory. Feel free to give the library a go now and comment on what you find. This is a very high quality library which should have an exciting review.