Mingming Liu 34f0edd509 [TypeProf][PGO]Support skipping vtable comparisons for a class and its derived ones (#110575) ...

Performance critical core libraries could be highly-optimized for arch
or micro-arch features. For instance, the absl crc library specializes
different templated classes among different hardwares [1]. In a
practical setting, it's likely that instrumented profiles are collected
on one type of machine and used to optimize binaries that run on
multiple types of hardwares.

While this kind of specialization is rare in terms of lines of code,
compiler can do a better job to skip vtable-based ICP.
* The per-class `Extend` implementation is arch-specific as well. If an
instrumented profile is collected on one arch and applied to another
arch where `Extend` implementation is different, `Extend` might be
regarded as unlikely function in the latter case. `ABSL_ATTRIBUTE_HOT`
annotation alleviates the problem by putting all `Extend` implementation
into the hot text section [2]

This change introduces a comma-separated list to specify the mangled
vtable names, and ICP pass will skip vtable-based comparison if a vtable
variable definition is shown to be in its class hierarchy (per LLVM type
metadata).

[1]
c6b27359c3/absl/crc/internal/crc_x86_arm_combined.cc (L621-L650)
[2]
c6b27359c3/absl/crc/internal/crc_x86_arm_combined.cc (L370C3-L370C21)

2024-10-02 10:23:54 -07:00

42 KiB

Raw Blame History

View Raw