https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112510
--- Comment #7 from Vladimir Sadovnikov <sadko4u at gmail dot com> ---
Disassembly for GCC 11.4.0:
```
00000000000011e9 <_Z13gate_x1_curvePfPKfPKN3dsp11gate_knee_tEm>:
11e9: f3 0f 1e fa endbr64
11ed: 55 push %rbp
11ee: 48 89 e5 mov %rsp,%rbp
11f1: 41 57 push %r15
11f3: 41 56 push %r14
11f5: 41 55 push %r13
11f7: 41 54 push %r12
11f9: 53 push %rbx
11fa: 48 83 e4 c0 and $0xffffffffffffffc0,%rsp
11fe: 48 81 ec c0 03 00 00 sub $0x3c0,%rsp
1205: 49 89 fc mov %rdi,%r12
1208: 49 89 f5 mov %rsi,%r13
120b: 49 89 d7 mov %rdx,%r15
120e: 49 89 ce mov %rcx,%r14
1211: 48 8d 5c 24 20 lea 0x20(%rsp),%rbx
1216: 48 89 5c 24 18 mov %rbx,0x18(%rsp)
121b: 83 3d ee 2d 00 00 00 cmpl $0x0,0x2dee(%rip) # 4010
<__asan_option_detect_stack_use_after_return@@Base>
1222: 0f 85 65 01 00 00 jne 138d
<_Z13gate_x1_curvePfPKfPKN3dsp11gate_knee_tEm+0x1a4>
1228: 48 c7 03 b3 8a b5 41 movq $0x41b58ab3,(%rbx)
122f: 48 8d 05 d2 0d 00 00 lea 0xdd2(%rip),%rax # 2008
<_IO_stdin_used+0x8>
1236: 48 89 43 08 mov %rax,0x8(%rbx)
123a: 48 8d 05 a8 ff ff ff lea -0x58(%rip),%rax # 11e9
<_Z13gate_x1_curvePfPKfPKN3dsp11gate_knee_tEm>
1241: 48 89 43 10 mov %rax,0x10(%rbx)
1245: 48 89 d8 mov %rbx,%rax
1248: 48 c1 e8 03 shr $0x3,%rax
124c: c7 80 00 80 ff 7f f1 movl $0xf1f1f1f1,0x7fff8000(%rax)
1253: f1 f1 f1
1256: c7 80 08 80 ff 7f f2 movl $0xf2f2f2f2,0x7fff8008(%rax)
125d: f2 f2 f2
1260: c7 80 1c 80 ff 7f f2 movl $0xf2f2f2f2,0x7fff801c(%rax)
1267: f2 f2 f2
126a: c7 80 20 80 ff 7f f2 movl $0xf2f2f2f2,0x7fff8020(%rax)
1271: f2 f2 f2
1274: c7 80 64 80 ff 7f f3 movl $0xf3f3f3f3,0x7fff8064(%rax)
127b: f3 f3 f3
127e: c7 80 68 80 ff 7f f3 movl $0xf3f3f3f3,0x7fff8068(%rax)
1285: f3 f3 f3
1288: c7 80 6c 80 ff 7f f3 movl $0xf3f3f3f3,0x7fff806c(%rax)
128f: f3 f3 f3
1292: 64 48 8b 14 25 28 00 mov %fs:0x28,%rdx
1299: 00 00
129b: 48 89 94 24 b8 03 00 mov %rdx,0x3b8(%rsp)
12a2: 00
12a3: 31 d2 xor %edx,%edx
12a5: 62 d2 7d 48 18 07 vbroadcastss (%r15),%zmm0
12ab: 62 d2 7d 48 18 4f 01 vbroadcastss 0x4(%r15),%zmm1
12b2: 62 d2 7d 48 18 57 02 vbroadcastss 0x8(%r15),%zmm2
12b9: 62 d2 7d 48 18 5f 03 vbroadcastss 0xc(%r15),%zmm3
12c0: 62 d2 7d 48 18 67 04 vbroadcastss 0x10(%r15),%zmm4
12c7: 62 d2 7d 48 18 6f 05 vbroadcastss 0x14(%r15),%zmm5
12ce: 62 d2 7d 48 18 77 06 vbroadcastss 0x18(%r15),%zmm6
12d5: 62 d2 7d 48 18 7f 07 vbroadcastss 0x1c(%r15),%zmm7
12dc: 62 f1 7c 48 29 83 20 vmovaps %zmm0,0x120(%rbx)
12e3: 01 00 00
12e6: 62 f1 7c 48 29 8b 60 vmovaps %zmm1,0x160(%rbx)
12ed: 01 00 00
12f0: 62 f1 7c 48 29 93 a0 vmovaps %zmm2,0x1a0(%rbx)
12f7: 01 00 00
12fa: 62 f1 7c 48 29 9b e0 vmovaps %zmm3,0x1e0(%rbx)
1301: 01 00 00
1304: 62 f1 7c 48 29 a3 20 vmovaps %zmm4,0x220(%rbx)
130b: 02 00 00
130e: 62 f1 7c 48 29 ab 60 vmovaps %zmm5,0x260(%rbx)
1315: 02 00 00
1318: 62 f1 7c 48 29 b3 a0 vmovaps %zmm6,0x2a0(%rbx)
131f: 02 00 00
1322: 62 f1 7c 48 29 bb e0 vmovaps %zmm7,0x2e0(%rbx)
1329: 02 00 00
```
Here we have offsets multiple of 0x20 but not multiple of 0x40 but missing
strange load of %rbx from stack. It is reasonable because of this instruction:
```
1211: 48 8d 5c 24 20 lea 0x20(%rsp),%rbx
```
All works fine.
gcc --version
gcc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
Copyright (C) 2021 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
