https://gcc.gnu.org/g:76f29e6239e47a3b0d61816f02804ac44c7679f0
commit 76f29e6239e47a3b0d61816f02804ac44c7679f0 Author: Michael Meissner <[email protected]> Date: Tue Sep 15 18:52:25 2026 -0400 Update ChangeLog.* Diff: --- gcc/ChangeLog.float | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) diff --git a/gcc/ChangeLog.float b/gcc/ChangeLog.float index 424da12e0bfd..c240e4e472a6 100644 --- a/gcc/ChangeLog.float +++ b/gcc/ChangeLog.float @@ -1,3 +1,24 @@ +==================== Branch work263-float, patch #212 ==================== + +Do BF to SF converts on p9; Remove int conversions; Add BF debug switches. + +2026-09-15 Michael Meissner <[email protected]> + +gcc/ + + * config/rs6000/float16.md (extendbf<mode>2): On power9, convert BFmode + to SFmode or DFmode by shifting. + (extendbf<mode>2_p1): Likewise. + (extendbf<mode>2_not_p10): Likewise. + (extendbfsf2_memory): Rename from convert_bf_to_sf_store. + (float<GPR:mode><FP16_HW:mode>2): Delete integer conversion insns. + (floatuns<GPR:mode><FP16_HW:mode>2): Likewise. + (fix_trunc<FP16_HW:mode><GPR:mode): Likewise. + (fixuns_trunc<FP16_HW:mode><GPR:mode>2): Likewise. + * config/rs6000/rs6000.h (TARGET_BFLOAT16_CVT): New macro. + * config/rs6000/rs6000.opt (-mbfloat16-convert): New debug switch. + (mbfloat16-p10-convert): Likewise. + ==================== Branch work263-float, patch #211 ==================== Limit 16-bit floating point to power9 or later.
