Re: riscv64 GCC 14 -> 15 C++ build time regression
On 9/28/25 12:06 PM, Matthias Klose wrote:
[CCing Jeff]
I asked Jeff at the Cauldron, and he confirmed there is an issue with
floating-point code.
Right. Though it's far from certain the ext-dce problem is the cause of
the compile-time slowdown or not. In fact, it looks like Aurelien may
have concluded its not ext-dce, but more likely location views.
There's been some talk of disabling some of the debugging bits that
don't have consumers (statement frontiers, maybe others).
jeff
Reply to: