[llvm] [X86] Support EGPR (R16-R31) for APX (PR #70958)

[Shengchen Kan via llvm-commits]

KanRobert wrote:

> I reran compile-time tests after #70222:
> 
> [Original results](https://llvm-compile-time-tracker.com/compare.php?from=b5dffd4957dfb58c73e168a3d9b6967f03b23a6c&to=feea5db01360b477b8cf2df03abfa9fc986633d5&stat=instructions%3Au) [New results](https://llvm-compile-time-tracker.com/compare.php?from=6c320b434d249c0d5692cbecc0e67ec22d659b7e&to=502ddafd0342160aefe92f3dced65865fad20746&stat=instructions:u)
> 
> It looks like the regression for optimized builds has been partially mitigated, but the regression for unoptimized builds is still about the same (e.g 0.7% on sqlite3).

Hi @nikic, From my side, the regression is smaller for O3 build. For example, the data in [New results](https://llvm-compile-time-tracker.com/compare.php?from=6c320b434d249c0d5692cbecc0e67ec22d659b7e&to=502ddafd0342160aefe92f3dced65865fad20746&stat=instructions:u) for O3 is 

```
CMakeFiles/sqlite3.dir/shell.c.o	1585M	1591M (+0.39%)
```
But when I run
```
bash$ valgrind --tool=callgrind clang-18 -DNDEBUG  -O3   -w -Werror=date-time -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1 -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1 -DSQLITE_OMIT_LOAD_EXTENSION=1 -DSQLITE_THREADSAFE=0 -I. -MD -MT MultiSource/Applications/sqlite3/CMakeFiles/sqlite3.dir/shell.c.o -MF MultiSource/Applications/sqlite3/CMakeFiles/sqlite3.dir/shell.c.o.d -o MultiSource/Applications/sqlite3/CMakeFiles/sqlite3.dir/shell.c.o -c /export/users/skan/test-suite/MultiSource/Applications/sqlite3/shell.c
```
the data is
```
CMakeFiles/sqlite3.dir/shell.c.o	1573M	1575M (+0.12%)
```

For O0 build, I can see 0.7% regression on sqlite3, too.



https://github.com/llvm/llvm-project/pull/70958