pika lpop 命令占用cpu比redis 高几倍

[hmyhelaku]

Is this a regression?

Yes

Description

pika lpop 命令占用cpu比redis 高几倍
比如同样的查询qps,redis占用不到 10%cpu，换pika 要占用80%的cpu
v3.5.5 v4.0.2都有这个问题，系统是centos7

Please provide a link to a minimal reproduction of the bug

No response

Screenshots or videos

No response

Please provide the version you discovered this bug in (check about page for version information)

Anything else?

No response

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

Title: =pika lpop command occupies several times higher CPU than redis

[Mixficsol]

Pika 和 Redis 的 CPU 占用差异主要源于底层存储引擎的不同。Redis 使用完全基于内存的存储，所有操作都在内存中进行，CPU 占用低。而 Pika 则是基于 RocksDB 存储引擎，虽然也有内存缓存，但很多操作需要依赖磁盘 I/O，尤其是在数据量较大的情况下，这就会导致更高的 CPU 占用。此外，Pika 是多线程的，虽然可以并行处理请求，但也带来了更多的上下文切换和调度开销。因此，相比 Redis，Pika 的 CPU 占用通常会更高，尤其是在处理大量 I/O 操作时，总的来说，Pika 做复杂数据结构的操作链路会更长

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

The difference in CPU usage between Pika and Redis mainly stems from the differences in the underlying storage engine. Redis uses fully memory-based storage, all operations are performed in memory and has a low CPU footprint. Pika is based on the RocksDB storage engine. Although it also has memory caches, many operations need to rely on disk I/O, especially when the data volume is large, which will lead to higher CPU usage. In addition, Pika is multi-threaded and although it can handle requests in parallel, it also brings more context switching and scheduling overhead. Therefore, Pika usually consumes more CPU than Redis, especially when dealing with large amounts of I/O operations. In general, Pika operates complex data structures with longer operation links.