performance improvements for pipelining

[romange]

from running:

 ./dfly_bench  --command "zadd foo nx __score__ __data__" -d 1 -c 1 --qps 0 -n 50000000 --pipeline=5000

I saw we have:

1. some bottleneck on absl::GetCurrentTimeNanos()
2. Bottleneck on ZzlStrtod (i.e. strtod)
3. Bottleneck in Transaction::StoreKeysInArgs (stub transaction in multi squasher), and in general InitByKeys is quite significant (8-9%) because it probably reads the argument slices in another thread (cold ram). I do not know if we can do anything about it, just noting.

[romange]

Same command with valkey:

docker run --network="host" --rm   --cpuset-cpus="2-3"   valkey/valkey:9.0   --save "" --appendonly no    --protected-mode no  

reaches almost double throughput.

(this scenario is not common in prod, the workload is deliberately chosen to showcase the threading/networking part of both systems for a single connection)