Reduction performance tweaks

[andrews05]

Two key changes here:

• Removed some defunct reduction benches and added some more alpha reduction benches.
• Changed use of chunks() to chunks_exact() where appropriate to improve performance. (As a byproduct, this also fixes a potential crash if a malformed file has a PLTE chunk that isn't a multiple of 3.)

[andrews05]

The chunks_exact makes a big difference to the 16-to-8 benchmarks. (Still not appreciable in actual use though.)
Before:

test reductions_16_to_8_bits                        ... bench:     352,770.85 ns/iter (+/- 11,464.83)
test reductions_16_to_8_bits_scaled                 ... bench:     748,591.70 ns/iter (+/- 7,528.53)

After

test reductions_16_to_8_bits                        ... bench:     200,931.25 ns/iter (+/- 4,184.77)
test reductions_16_to_8_bits_scaled                 ... bench:     116,423.44 ns/iter (+/- 652.00)

[AlexTMjugador]

LGTM, thank you!