https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127420
--- Comment #9 from Haochen Jiang <haochen.jiang at intel dot com> ---
(In reply to H.J. Lu from comment #8)
> (In reply to Haochen Jiang from comment #7)
> > > What are typical cases? Are they C++ programs?
> >
> > All SPEC2026 cases are CXX + C or CXX. For SPEC2017, 511 is CXX + C, the
> > remaining for mentioned is C only.
> >
> > For 735, the case is a 20-case switch in a loop:
> >
> > c3c090: inc %rdx
> > c3c093: mov %ecx,(%rax)
> > c3c095: movq $0x0,0x8(%rax)
> > c3c09d: cmp %rsi,%rdx
> > c3c0a0: jae c3c118
> > c3c0a2: cmp $0x13,%edx
> > c3c0a5: ja c3c2ff
> > c3c0ab: mov %edx,%ecx
> > c3c0ad: jmp *0x1021900(,%rcx,8) # 20-entry tablejump <- 7
> > bytes
> > c3c0b4: nopl 0x0(%rax)
> > c3c0b8: xor %ecx,%ecx # <-- 932,626 hits land here
> > c3c0ba: add $0x10,%rax
> > c3c0be: jmp c3c090
> >
> > The 84% c3c0ad indirect jmp FALLTHRU to c3c0b8, leading to total 26.7% for
> > 735.
>
> My patch only applies to sibcall, not table jump:
>
> [hjl@gnu-zen4-1 pr127420]$ cat t.c
> extern void bar0 (void);
> extern void bar1 (void);
> extern void bar2 (void);
> extern void bar3 (void);
> extern void bar4 (void);
>
> void
> foo (int i)
> {
> switch (i)
> {
> case 0: bar0 (); break;
> case 1: bar1 (); break;
> case 2: bar2 (); break;
> case 3: bar3 (); break;
> case 4: bar4 (); break;
> }
> }
> [hjl@gnu-zen4-1 pr127420]$ make
> /export/build/gnu/tools-build/gcc-gitlab-test-debug/build-x86_64-linux/gcc/
> xgcc
> -B/export/build/gnu/tools-build/gcc-gitlab-test-debug/build-x86_64-linux/gcc/
> -O2 -fno-plt -S t.c
> [hjl@gnu-zen4-1 pr127420]$ cat t.s
> .file "t.c"
> .text
> .p2align 4
> .globl foo
> .type foo, @function
> foo:
> .LFB0:
> .cfi_startproc
> cmpl $4, %edi
> ja .L1
> movl %edi, %edi
> jmp *.L4(,%rdi,8)
> .section .rodata
> .align 8
> .align 4
> .L4:
> .quad .L8
> .quad .L7
> .quad .L6
> .quad .L5
> .quad .L3
> .text
> .p2align 4,,10
> .p2align 3
> .L5:
> jmp *bar3@GOTPCREL(%rip)
> ud2
> .p2align 4,,10
> .p2align 3
> .L3:
> jmp *bar4@GOTPCREL(%rip)
> ud2
> .p2align 4,,10
> .p2align 3
> .L8:
> jmp *bar0@GOTPCREL(%rip)
> ud2
> .p2align 4,,10
> .p2align 3
> .L7:
> jmp *bar1@GOTPCREL(%rip)
> ud2
> .p2align 4,,10
> .p2align 3
> .L6:
> jmp *bar2@GOTPCREL(%rip)
> ud2
> .L1:
> ret
> .cfi_endproc
> .LFE0:
> .size foo, .-foo
> .ident "GCC: (GNU) 17.0.0 20260918 (experimental)"
> .section .note.GNU-stack,"",@progbits
> [hjl@gnu-zen4-1 pr127420]$
>
> Can you measure its performance impact?
I am currently mearsuring on EMR.
