Optimize zz_and() for small arguments and get rid of zz_and_sl()?

[skirpichev]

// like mpz_and()

[skirpichev]

benchmark from diofant/python-gmp#263

patch - no mixed mode for zz_and(), xxx - new zz (optimized zz_and), xxx2 - collatz1 with & one
gcc (Debian 14.2.0-19) 14.2.0

release (-O3):

Benchmark	gmpy2	ref	patch	xxx	xxx2
collatz0(97)	101 us	87.6 us: 1.15x faster	87.7 us: 1.15x faster	85.9 us: 1.17x faster	86.3 us: 1.17x faster
collatz0(871)	151 us	133 us: 1.13x faster	132 us: 1.14x faster	128 us: 1.18x faster	129 us: 1.17x faster
collatz0((1<<128)+31)	813 us	717 us: 1.13x faster	705 us: 1.15x faster	692 us: 1.17x faster	691 us: 1.18x faster
collatz1(97)	95.3 us	85.6 us: 1.11x faster	104 us: 1.09x slower	102 us: 1.07x slower	84.6 us: 1.13x faster
collatz1(871)	143 us	129 us: 1.10x faster	156 us: 1.10x slower	153 us: 1.07x slower	127 us: 1.12x faster
collatz1((1<<128)+31)	771 us	690 us: 1.12x faster	844 us: 1.09x slower	811 us: 1.05x slower	694 us: 1.11x faster
Geometric mean	(ref)	1.12x faster	1.02x faster	1.05x faster	1.14x faster

debug:

Benchmark	gmpy2	ref	patch	xxx	xxx2
collatz0(97)	134 us	138 us: 1.03x slower	144 us: 1.08x slower	143 us: 1.06x slower	142 us: 1.06x slower
collatz0(871)	202 us	207 us: 1.03x slower	216 us: 1.07x slower	212 us: 1.05x slower	215 us: 1.07x slower
collatz0((1<<128)+31)	1.06 ms	1.10 ms: 1.03x slower	1.16 ms: 1.09x slower	1.13 ms: 1.06x slower	1.13 ms: 1.06x slower
collatz1(97)	143 us	149 us: 1.04x slower	183 us: 1.28x slower	175 us: 1.23x slower	146 us: 1.02x slower
collatz1(871)	214 us	222 us: 1.04x slower	274 us: 1.28x slower	262 us: 1.23x slower	218 us: 1.02x slower
collatz1((1<<128)+31)	1.13 ms	1.17 ms: 1.04x slower	1.42 ms: 1.26x slower	1.38 ms: 1.22x slower	1.16 ms: 1.02x slower
Geometric mean	(ref)	1.03x slower	1.17x slower	1.14x slower	1.04x slower

with Debian clang version 19.1.7 (3+b1)

release (-O3):

Benchmark	gmpy2	ref	patch	xxx	xxx2
collatz0(97)	97.0 us	91.5 us: 1.06x faster	88.9 us: 1.09x faster	85.9 us: 1.13x faster	85.7 us: 1.13x faster
collatz0(871)	144 us	136 us: 1.06x faster	132 us: 1.09x faster	129 us: 1.12x faster	129 us: 1.12x faster
collatz0((1<<128)+31)	784 us	737 us: 1.06x faster	710 us: 1.10x faster	704 us: 1.11x faster	695 us: 1.13x faster
collatz1(97)	96.2 us	92.5 us: 1.04x faster	107 us: 1.11x slower	103 us: 1.07x slower	86.3 us: 1.11x faster
collatz1(871)	144 us	139 us: 1.03x faster	159 us: 1.11x slower	154 us: 1.07x slower	131 us: 1.09x faster
collatz1((1<<128)+31)	760 us	739 us: 1.03x faster	843 us: 1.11x slower	814 us: 1.07x slower	707 us: 1.08x faster
Geometric mean	(ref)	1.05x faster	1.01x slower	1.02x faster	1.11x faster

debug:

Benchmark	gmpy2	ref	patch	xxx	xxx2
collatz0(97)	124 us	136 us: 1.09x slower	135 us: 1.09x slower	132 us: 1.07x slower	133 us: 1.07x slower
collatz0(871)	186 us	203 us: 1.09x slower	204 us: 1.10x slower	199 us: 1.07x slower	198 us: 1.07x slower
collatz0((1<<128)+31)	988 us	1.07 ms: 1.08x slower	1.07 ms: 1.09x slower	1.05 ms: 1.07x slower	1.05 ms: 1.06x slower
collatz1(97)	132 us	144 us: 1.09x slower	175 us: 1.33x slower	168 us: 1.28x slower	141 us: 1.07x slower
collatz1(871)	198 us	216 us: 1.09x slower	262 us: 1.32x slower	252 us: 1.27x slower	213 us: 1.07x slower
collatz1((1<<128)+31)	1.04 ms	1.14 ms: 1.10x slower	1.39 ms: 1.33x slower	1.34 ms: 1.29x slower	1.11 ms: 1.07x slower
Geometric mean	(ref)	1.09x slower	1.20x slower	1.17x slower	1.07x slower