Why is it that a CPU running at full capacity only has less than 2000RPS when I test throughput on Linux x86 system?

[eaufavor]

What is your test setup? Benchmark numbers can vary a lot depending on your test setup.

[dosens]

What is your test setup? Benchmark numbers can vary a lot depending on your test setup.

I am not sure if the following information is sufficient. 1、The source code of load_balance is copied from the quick_start.md file；

2、Pingora uses a yaml configuration file as follows： version: 1 threads: 1 pid_file: /tmp/load_balancer.pid error_log: /tmp/load_balancer_err.log upgrade_sock: /tmp/load_balancer.sock upstream_keepalive_pool_size: 10240

3、The backend is an nginx server with a performance of up to 80000 RPS

[styrowolf]

Are you sure that you're building the program in release mode?

[dosens]

Are you sure that you're building the program in release mode?

Oh, I understand. I used to compile in debug mode, thank you for the reminder!

[dosens]

Are you sure that you're building the program in release mode?

Oh, I understand. I used to compile in debug mode, thank you for the reminder!

I tested using release mode and found that one CPU has 2.2W RPS. Does this meet expectations?

[bestgopher]

What is your test setup? Benchmark numbers can vary a lot depending on your test setup.

I am not sure if the following information is sufficient. 1、The source code of load_balance is copied from the quick_start.md file；

2、Pingora uses a yaml configuration file as follows： version: 1 threads: 1 pid_file: /tmp/load_balancer.pid error_log: /tmp/load_balancer_err.log upgrade_sock: /tmp/load_balancer.sock upstream_keepalive_pool_size: 10240

3、The backend is an nginx server with a performance of up to 80000 RPS

threads: 1 means single thread?

[eaufavor]

I tested using release mode and found that one CPU has 2.2W RPS. Does this meet expectations?

The raw RPS highly depends on your hardware/OS. To compare, try setting up a nginx/enovy proxy and run your benchmark on the same machine.

BTW I saw from https://github.com/envoyproxy/envoy/issues/19103 that envoy has RPS in the same order of magnitude (hardware spec undisclosed).

[dosens]

What is your test setup? Benchmark numbers can vary a lot depending on your test setup.

I am not sure if the following information is sufficient. 1、The source code of load_balance is copied from the quick_start.md file； 2、Pingora uses a yaml configuration file as follows： version: 1 threads: 1 pid_file: /tmp/load_balancer.pid error_log: /tmp/load_balancer_err.log upgrade_sock: /tmp/load_balancer.sock upstream_keepalive_pool_size: 10240 3、The backend is an nginx server with a performance of up to 80000 RPS

threads: 1 means single thread?

yes.

[dosens]

I tested multiple threads and one thread to obtain RPS throughput, which is not a multiple relationship. Does it meet expectations?

1 thread - Pingora echo http server - 7,9000 RPS 2 thread - Pingora echo http server - 12,4000 RPS 3 thread - Pingora echo http server - 17,0000 RPS

1 worker - nginx http server - 9,8000 RPS 2 worker - nginx http server - 18,1000 RPS 3 worker - nginx http server - 22,1000 RPS

The Pingora echo HTTP server and nginx http server work on the same operating system, using the same binding core strategy.

[eaufavor]

Does it meet expectations?

These numbers looks fine to me. Please be mindful that the performance numbers on these synthetic benchmarks are not the direct indicators of real world performance.

[github-actions[bot]]

This question has been stale for a week. It will be closed in an additional day if not updated.

[github-actions[bot]]

This issue has been closed because it has been stalled with no activity.