docs: add `cva` to benchmarks

[mskelton]

Fixes #134

Adds CVA to the benchmarks since a lot of folks often want to compare the perf of the two libraries.

There are only two benchmarks for CVA since it doesn't support slots.

TV without slots & tw-merge (enabled) x 743,372 ops/sec ±0.62% (93 runs sampled)
TV without slots & tw-merge (disabled) x 1,006,953 ops/sec ±0.27% (95 runs sampled)
TV with slots & tw-merge (enabled) x 304,484 ops/sec ±1.21% (96 runs sampled)
TV with slots & tw-merge (disabled) x 343,062 ops/sec ±1.04% (98 runs sampled)
TV without slots & custom tw-merge config x 729,188 ops/sec ±1.28% (96 runs sampled)
TV with slots & custom tw-merge config x 404,001 ops/sec ±0.22% (98 runs sampled)
CVA without slots & tw-merge (enabled) x 1,039,518 ops/sec ±0.20% (100 runs sampled)
CVA without slots & tw-merge (disabled) x 1,915,606 ops/sec ±1.87% (97 runs sampled)
Fastest is CVA without slots & tw-merge (disabled)