https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114729
--- Comment #14 from Vineet Gupta <vineetg at gcc dot gnu.org> --- Interim update: Per discussions [1] [2] with Richard Sandiford, some of the behavior is fundamental to the "model" heuristics of -fsched-pressure, specially for in-order cores which benefit from little more spilling. If the ECC clamping is disabled as hack in [1] things improve a little. We go down from 2.4 trillion icounts to 2 trillion, while still excess 0.8 trillion if sched1 is disabled altogether. [1] https://gcc.gnu.org/pipermail/gcc-patches/2024-August/659776.html [2] https://gcc.gnu.org/pipermail/gcc-patches/2024-August/659847.html
