Richard Biener of SUSE published a new status report on the state of GCC 16 development. Regression fixing has been going slow but they are hoping to publish a release candidate by mid-April.
Richard shared in today’s GCC 16 status report:
“We’re now nearly two months into Stage 4 of the development of GCC 16, making slow progress towards reducing the number of regressions, in particular P1 classified regressions to zero.
While there are only 14 P1 classified regressions there are another 14 unclassified (P3) regressions would be introduced by GCC 16, and 28 P4 classified of similar kind, meaning for not release critical languages or targets – you might want to prioritize those as to make GCC 16 not worse than GCC 15 for users.
Historically we expect a RC of GCC 16 to materialize around mid April.”
It’s that P1 regression count currently at 14 that they try to zero out for release time by either fixing them or demoting them to a lower priority regression. As of writing there are 586 P2 regressions and 190 P3 regressions.
GCC 16 brings many changes including the Algol 68 programming language front-end, C++20 standard by default, AMD Zen 6 “znver6” initial support, support for using the Picolibc embedded C library, AVX10.2 and APX support is ready for Intel Nova Lake, various performance optimizations, Intel Wildcat Lake targeting support, a higher default LTO partition count to deal with today’s higher core count processors, function multi-versioning on ARM64 is no longer experimental, and many other changes throughout this dominant open-source compiler stack.
