[llvm-dev] Converting between different but same-sized floating-point types

[Russell Wallace via llvm-dev]

If a narrowing floating-point conversion is fptrunc, and a widening one is
fpext, what instruction is used for a conversion between half and bfloat?
It is neither purely narrowing nor purely widening, nor is it a bitcast
that just reinterprets the bits.

Similarly, what happens in the case of fp128 and ppc_fp128?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20220108/6892d046/attachment.html>