https://gcc.gnu.org/g:68d98f521b57588e586dfa6321d25307669eeaeb
commit 68d98f521b57588e586dfa6321d25307669eeaeb Author: Michael Meissner <[email protected]> Date: Fri Jun 19 19:36:52 2026 -0400 Update ChangeLog.* Diff: --- gcc/ChangeLog.float | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/gcc/ChangeLog.float b/gcc/ChangeLog.float index b69aa8b018aa..741cdbbcb8da 100644 --- a/gcc/ChangeLog.float +++ b/gcc/ChangeLog.float @@ -1,3 +1,21 @@ +==================== Branch work247-float, patch #231 ==================== + +Optimization convertion between __bfloat16 and _Float16. + +2026-06-19 Michael Meissner <[email protected]> + +gcc/ + + * config/rs6000/float16.md (cvt_fp16_to_v4sf): Rename from + cvt_fp16_to_v4sf_insn. + (cvt_v4sf_to_fp16): New mode attribute. + (UNSPEC_CVT_V4SF_TO_FP16): New unspec. + (trunchfbf2): Rework to do the conversion from vector to vector, and not + converting to double as an intermediate. Move ode down in float16.md. + (extendbfhf2): Likewise. + (trunchfbf2_mem): New insn to optimization load and conversion. + (extendbfhf2_mem): Likewise. + ==================== Branch work247-float, patch #230 ==================== Add _Float16 and __bfloat16 tests.
