https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88784
--- Comment #6 from Fredrik Hederstierna <fredrik.hederstie...@securitas-direct.com> --- Created attachment 46397 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=46397&action=edit Some more patterns Looking into this I found some more places where it seems to be non-optimal code, maybe separate issue, but these are also example of equal evaluation for unsigned types ? Test 1 (x > y) || (x > (y / N)) equal to (x > (y / N)) Test 2 (x > y) || (x > (y >> N)) equal to (x > (y >> N)) Test 3 (x > y) && (x > (y / N)) equal to (x > y) Test 4 (x > y) && (x > (y >> N)) equal to (x > y) One thing to consider here maybe that depending on optimizing for size or speed, then the order of evaluation can be changed, so like if some operation is costy, then it could be avoided to obtain higher speed possibly assuming it will accept arguments prior in list I guess. But when optimizing for size, then I think always the more simplified expression would apply? Example for arm using above expressions, (code attached) 00000000 <test_xy_1_org>: 0: b510 push {r4, lr} 2: 000b movs r3, r1 4: 0004 movs r4, r0 6: 2001 movs r0, #1 8: 428c cmp r4, r1 a: d807 bhi.n 1c <test_xy_1_org+0x1c> c: 2103 movs r1, #3 e: 0018 movs r0, r3 10: f7ff fffe bl 0 <__aeabi_uidiv> 14: b2c0 uxtb r0, r0 16: 42a0 cmp r0, r4 18: 4180 sbcs r0, r0 1a: 4240 negs r0, r0 1c: bd10 pop {r4, pc} 0000001e <test_xy_1_equ>: 1e: b510 push {r4, lr} 20: 0004 movs r4, r0 22: 0008 movs r0, r1 24: 2103 movs r1, #3 26: f7ff fffe bl 0 <__aeabi_uidiv> 2a: b2c0 uxtb r0, r0 2c: 42a0 cmp r0, r4 2e: 4180 sbcs r0, r0 30: 4240 negs r0, r0 32: bd10 pop {r4, pc} 00000034 <test_xy_2_org>: 34: 0003 movs r3, r0 36: 2001 movs r0, #1 38: 428b cmp r3, r1 3a: d803 bhi.n 44 <test_xy_2_org+0x10> 3c: 08c9 lsrs r1, r1, #3 3e: 4299 cmp r1, r3 40: 4189 sbcs r1, r1 42: 4248 negs r0, r1 44: 4770 bx lr 00000046 <test_xy_2_equ>: 46: 08c9 lsrs r1, r1, #3 48: 4281 cmp r1, r0 4a: 4180 sbcs r0, r0 4c: 4240 negs r0, r0 4e: 4770 bx lr 00000050 <test_xy_3_org>: 50: b510 push {r4, lr} 52: 000b movs r3, r1 54: 0004 movs r4, r0 56: 2000 movs r0, #0 58: 428c cmp r4, r1 5a: d907 bls.n 6c <test_xy_3_org+0x1c> 5c: 2103 movs r1, #3 5e: 0018 movs r0, r3 60: f7ff fffe bl 0 <__aeabi_uidiv> 64: b2c0 uxtb r0, r0 66: 42a0 cmp r0, r4 68: 4180 sbcs r0, r0 6a: 4240 negs r0, r0 6c: bd10 pop {r4, pc} 0000006e <test_xy_3_equ>: 6e: 4281 cmp r1, r0 70: 4180 sbcs r0, r0 72: 4240 negs r0, r0 74: 4770 bx lr 00000076 <test_xy_4_org>: 76: 0003 movs r3, r0 78: 2000 movs r0, #0 7a: 428b cmp r3, r1 7c: d903 bls.n 86 <test_xy_4_org+0x10> 7e: 08c9 lsrs r1, r1, #3 80: 4299 cmp r1, r3 82: 4189 sbcs r1, r1 84: 4248 negs r0, r1 86: 4770 bx lr 00000088 <test_xy_4_equ>: 88: 4281 cmp r1, r0 8a: 4180 sbcs r0, r0 8c: 4240 negs r0, r0 8e: 4770 bx lr
