Benchmarking hyperdrive v10

NANOBENCH version 2
> /usr/bin/node bench.js

# write fs 5000 * 4K sequentially
ok ~262 ms (0 s + 261710236 ns)

# write hyperdrive 5000 * 4K sequentially
ok ~3.65 s (3 s + 653388561 ns)

# write client 5000 * 4K sequentially
ok ~17 s (16 s + 794988825 ns)

# write fuse 5000 * 4K sequentially
ok ~10 s (10 s + 368266064 ns)



# write fs 100 * 4K in parallel
ok ~62 ms (0 s + 62275880 ns)

# write hyperdrive 100 * 4K in parallel
ok ~65 ms (0 s + 64558644 ns)

# write client 100 * 4K in parallel
ok ~161 ms (0 s + 161051725 ns)

# write fuse 100 * 4K in parallel
(does not finish, seems to break the daemon?)