Adds bench for writing accounts to append vecs and hot storage

[brooksprumo]

Problem

There's not a bench in place to compare the perf between append vecs and tiered storage.

Summary of Changes

Add a bench for writing accounts. Subsequent PRs can add benches for reading accounts.

Here's what a run looks like:

write_accounts_file/append_vec/1
                        time:   [966.62 ns 983.87 ns 999.72 ns]
                        thrpt:  [1.0003 Melem/s 1.0164 Melem/s 1.0345 Melem/s]
write_accounts_file/hot_storage/1
                        time:   [277.20 ns 303.86 ns 342.69 ns]
                        thrpt:  [2.9181 Melem/s 3.2910 Melem/s 3.6075 Melem/s]
write_accounts_file/append_vec/100
                        time:   [16.793 µs 16.946 µs 17.084 µs]
                        thrpt:  [5.8533 Melem/s 5.9010 Melem/s 5.9550 Melem/s]
write_accounts_file/hot_storage/100
                        time:   [21.794 µs 21.994 µs 22.174 µs]
                        thrpt:  [4.5098 Melem/s 4.5468 Melem/s 4.5885 Melem/s]
write_accounts_file/append_vec/1000
                        time:   [160.72 µs 161.89 µs 162.82 µs]
                        thrpt:  [6.1419 Melem/s 6.1771 Melem/s 6.2220 Melem/s]
write_accounts_file/hot_storage/1000
                        time:   [222.95 µs 223.76 µs 224.42 µs]
                        thrpt:  [4.4559 Melem/s 4.4691 Melem/s 4.4854 Melem/s]
write_accounts_file/append_vec/10000
                        time:   [1.6278 ms 1.6291 ms 1.6305 ms]
                        thrpt:  [6.1331 Melem/s 6.1384 Melem/s 6.1432 Melem/s]
write_accounts_file/hot_storage/10000
                        time:   [2.2862 ms 2.2894 ms 2.2933 ms]
                        thrpt:  [4.3605 Melem/s 4.3679 Melem/s 4.3741 Melem/s]

[jeffwashington]

lgtm

[yhchiang-sol]

Looks good! Thanks for working on this!