vector_algorithms.cpp, *minmax*: invert the condition to improve *_element cases a bit more

[AlexGuteniev]

The possibility was realized when working on #4384.

I observed that the compiler swaps branches to make the _Minmax loop more local.
Applying the same change to the _Minmax_element loop gives significant improvement: 1.6 or 1.8 times for some cases.
Also made that change to _Minmax loop to maintain consistency (with both its counterpart and how it being compiled).

Benchmark results

--------------------------------------------------------------------------------------------
Benchmark                              Old Time        Old CPU    New Time         New CPU  
--------------------------------------------------------------------------------------------
bm<uint8_t, 8021, Op::Min>              647 ns          642 ns      361 ns          361 ns  
bm<uint8_t, 8021, Op::Max>              598 ns          586 ns      368 ns          369 ns  
bm<uint8_t, 8021, Op::Both>             679 ns          684 ns      556 ns          562 ns  
bm<uint8_t, 8021, Op::Min_val>          278 ns          279 ns      272 ns          273 ns  
bm<uint8_t, 8021, Op::Max_val>          285 ns          285 ns      279 ns          276 ns  
bm<uint8_t, 8021, Op::Both_val>       20891 ns        20996 ns    20642 ns        20403 ns  
bm<uint16_t, 8021, Op::Min>            1093 ns         1099 ns      791 ns          785 ns  
bm<uint16_t, 8021, Op::Max>            1104 ns         1099 ns      659 ns          670 ns  
bm<uint16_t, 8021, Op::Both>           1595 ns         1604 ns     1050 ns         1050 ns  
bm<uint16_t, 8021, Op::Min_val>         332 ns          330 ns      326 ns          322 ns  
bm<uint16_t, 8021, Op::Max_val>         331 ns          337 ns      324 ns          322 ns  
bm<uint16_t, 8021, Op::Both_val>      12045 ns        12207 ns    11841 ns        11719 ns  
bm<uint32_t, 8021, Op::Min>            2135 ns         2093 ns     1294 ns         1287 ns  
bm<uint32_t, 8021, Op::Max>            1356 ns         1350 ns     1304 ns         1311 ns  
bm<uint32_t, 8021, Op::Both>           2117 ns         2131 ns     2092 ns         2086 ns  
bm<uint32_t, 8021, Op::Min_val>         627 ns          641 ns      618 ns          625 ns  
bm<uint32_t, 8021, Op::Max_val>        1041 ns         1050 ns     1049 ns         1050 ns  
bm<uint32_t, 8021, Op::Both_val>        666 ns          670 ns      693 ns          698 ns  
bm<uint64_t, 8021, Op::Min>            4629 ns         4604 ns     4615 ns         4604 ns  
bm<uint64_t, 8021, Op::Max>            4745 ns         4757 ns     4735 ns         4708 ns  
bm<uint64_t, 8021, Op::Both>           5065 ns         5156 ns     4925 ns         5000 ns  
bm<uint64_t, 8021, Op::Min_val>        4068 ns         4081 ns     4068 ns         4081 ns  
bm<uint64_t, 8021, Op::Max_val>        4068 ns         4081 ns     4058 ns         3990 ns  
bm<uint64_t, 8021, Op::Both_val>       4103 ns         4081 ns     4082 ns         4081 ns  
bm<int8_t, 8021, Op::Min>               551 ns          547 ns      358 ns          357 ns  
bm<int8_t, 8021, Op::Max>               578 ns          578 ns      364 ns          368 ns  
bm<int8_t, 8021, Op::Both>              708 ns          711 ns      553 ns          562 ns  
bm<int8_t, 8021, Op::Min_val>           275 ns          279 ns      268 ns          267 ns  
bm<int8_t, 8021, Op::Max_val>           274 ns          273 ns      269 ns          264 ns  
bm<int8_t, 8021, Op::Both_val>        20376 ns        20403 ns    20147 ns        19950 ns  
bm<int16_t, 8021, Op::Min>             1046 ns         1050 ns      783 ns          785 ns  
bm<int16_t, 8021, Op::Max>             1041 ns         1050 ns      675 ns          684 ns  
bm<int16_t, 8021, Op::Both>            1609 ns         1604 ns     1084 ns         1088 ns  
bm<int16_t, 8021, Op::Min_val>          535 ns          544 ns      530 ns          531 ns  
bm<int16_t, 8021, Op::Max_val>          571 ns          572 ns      545 ns          544 ns  
bm<int16_t, 8021, Op::Both_val>       13254 ns        13393 ns    13141 ns        13114 ns  
bm<int32_t, 8021, Op::Min>             2071 ns         2086 ns     1292 ns         1287 ns  
bm<int32_t, 8021, Op::Max>             1356 ns         1350 ns     1312 ns         1311 ns  
bm<int32_t, 8021, Op::Both>            2153 ns         2148 ns     2066 ns         2086 ns  
bm<int32_t, 8021, Op::Min_val>         1052 ns         1067 ns     1033 ns         1025 ns  
bm<int32_t, 8021, Op::Max_val>          622 ns          628 ns      620 ns          614 ns  
bm<int32_t, 8021, Op::Both_val>        1030 ns         1046 ns     1045 ns         1050 ns  
bm<int64_t, 8021, Op::Min>             4628 ns         4604 ns     4672 ns         4604 ns  
bm<int64_t, 8021, Op::Max>             4746 ns         4708 ns     4763 ns         4743 ns  
bm<int64_t, 8021, Op::Both>            5013 ns         5000 ns     5013 ns         5000 ns  
bm<int64_t, 8021, Op::Min_val>         4065 ns         4081 ns     4157 ns         4171 ns  
bm<int64_t, 8021, Op::Max_val>         4071 ns         4098 ns     4136 ns         4143 ns  
bm<int64_t, 8021, Op::Both_val>        4090 ns         4081 ns     4159 ns         4143 ns  
bm<float, 8021, Op::Min>               2219 ns         2197 ns     2423 ns         2459 ns  
bm<float, 8021, Op::Max>               2469 ns         2455 ns     2409 ns         2407 ns  
bm<float, 8021, Op::Both>              2558 ns         2567 ns     3202 ns         3209 ns  
bm<float, 8021, Op::Min_val>           2001 ns         1995 ns     2031 ns         1995 ns  
bm<float, 8021, Op::Max_val>           1998 ns         1995 ns     2038 ns         2040 ns  
bm<float, 8021, Op::Both_val>          2034 ns         2040 ns     2039 ns         2040 ns  
bm<double, 8021, Op::Min>              4206 ns         4262 ns     4084 ns         4081 ns  
bm<double, 8021, Op::Max>              4378 ns         4395 ns     4347 ns         4395 ns  
bm<double, 8021, Op::Both>             5433 ns         5301 ns     5139 ns         5162 ns  
bm<double, 8021, Op::Min_val>          4027 ns         3990 ns     4028 ns         3990 ns  
bm<double, 8021, Op::Max_val>          4032 ns         4081 ns     4037 ns         4081 ns  
bm<double, 8021, Op::Both_val>         4113 ns         4081 ns     4065 ns         4081 ns  

[StephanTLavavej]

I'm mirroring this to the MSVC-internal repo - please notify me if any further changes are pushed.

[StephanTLavavej]

Maximum warp! 🛸 🚀 ⚡