More efficient Eq, Ord for Seq

[meooow25]

For #1016.

Benchmarks on GHC 9.6.3:

Before

All
  eq
    100/100:     OK
      1.43 μs ±  86 ns, 7.5 KB allocated,   1 B  copied,  10 MB peak memory
    10000/10000: OK
      152  μs ±  13 μs, 748 KB allocated, 237 B  copied,  10 MB peak memory
  compare
    100/100:     OK
      1.59 μs ±  88 ns, 7.5 KB allocated,   1 B  copied,  10 MB peak memory
    10000/10000: OK
      170  μs ±  14 μs, 748 KB allocated, 234 B  copied,  10 MB peak memory
    100/10000:   OK
      1.62 μs ± 158 ns, 7.8 KB allocated,   1 B  copied,  10 MB peak memory
    10000/100:   OK
      1.70 μs ± 166 ns, 7.7 KB allocated,   1 B  copied,  10 MB peak memory

After

All
  eq
    100/100:     OK
      843  ns ±  43 ns, 7.1 KB allocated,   0 B  copied,  10 MB peak memory, 40% less than baseline
    10000/10000: OK
      92.1 μs ± 5.4 μs, 719 KB allocated, 217 B  copied,  10 MB peak memory, 39% less than baseline
  compare
    100/100:     OK
      955  ns ±  90 ns, 7.1 KB allocated,   0 B  copied,  10 MB peak memory, 39% less than baseline
    10000/10000: OK
      98.0 μs ± 5.8 μs, 719 KB allocated, 216 B  copied,  10 MB peak memory, 42% less than baseline
    100/10000:   OK
      931  ns ±  55 ns, 7.4 KB allocated,   1 B  copied,  10 MB peak memory, 42% less than baseline
    10000/100:   OK
      941  ns ±  88 ns, 7.4 KB allocated,   1 B  copied,  10 MB peak memory, 44% less than baseline