4 ms·

These kinds of benchmarks are not very meaningful. I think that pretty much any modern framework/language can handle at least 1 million idle WebSockets. It's mu

These kinds of benchmarks are not very meaningful. I think that pretty much any modern framework/language can handle at least 1 million idle WebSockets. It's much more interesting to measure performance when you start sending messages through them at regular intervals.

e-r-a-n 8y ago
They are not idling. Look at the code.
kjksf 8y ago
When people rewrite system from X where X in (Python, Ruby, Node, Clojure) to Go they usually see at least 10x improvement. This is just one recent example: https://www.infoq.com/articles/api-gateway-clojure-golang https://www.infoq.com/articles/api-gateway-clojure-golang The money quote: "The end result enabled us to reduce 25 instances (c4 xlarge) running Clojure code - able to process 60 concurrent requests, to two instances (c3.2xlarge) running Go code able to support ~5000 concurrent requests a minute" If you google around you'll find more stories like that. To do better than Go you would have to drop to C++ or Rust.
- Scarbutt 8y ago
  "The end result enabled us to reduce 25 instances (c4 xlarge) running Clojure code - able to process 60 concurrent requests, to two instances (c3.2xlarge) running Go code able to support ~5000 concurrent requests a minute" That doesn't make any sense though.
- karim 8y ago
  "When people rewrite system from X where X in (Python, Ruby, Node, Clojure) to Go they usually see at least 10x improvement." Sure but how much of the performance gains come from Go and how much come from just having a better understanding of the problem the second time around?
  - zepolen 8y ago
    Go is strongly typed and compiled, it literally does 10x less work than dynamic languages for most things. Pypy leads to a near 10x speed up over cPython as well.
- kasey_junk 8y ago
  Or Java or C# or basically any non-interpreted language.
  - rakenodiax 8y ago
    In this case, Clojure is already running on the JVM
    - kasey_junk 8y ago
      The Clojure compiler does not generally produce the same byte code as the Java one. Idiomatic clojure is routinely slower than idiomatic Java and that’s a well known and expected outcome.
- piranha 8y ago
  That article is anything but useful. They have spent significantly more time designing and writing app in Go than their prototype in Clojure and now it’s faster? What a surprise!
- apta 8y ago
  > To do better than Go you would have to drop to C++ or Rust. C++ and Rust yes, but also Java and .NET.
  - weberc2 8y ago
    In my experience Java, C#, and Go are roughly on par. Go is a bit easier to optimize though.
- basic6 8y ago
  I didn't think Go rewrites were common anymore since the results were mediocre.
- fulafel 8y ago
  Hmm, there's a slight disparity between the quote and their benchmark table, which shows Clojure and Go neck and neck in the request/sec dept. I guess all the system re-architecting explains most of the difference. In any case, as the article text also mentions, on the Go side they are using a complete reverse-proxy library that's included with the Go stdlib, which can be a significant advantage aside from the properties of the language itself. But then it seems they ended up reimplementing many things that the JVM provides: "In order to achieve out of the box functionality such as CPU and memory usage metrics, business logic counters and more - we needed to write basically this entire stack from scratch, which enabled us a much more rapid deep dive of the intricacies of Golang."

e-r-a-n 8y ago

They are not idling. Look at the code.

kjksf 8y ago

When people rewrite system from X where X in (Python, Ruby, Node, Clojure) to Go they usually see at least 10x improvement. This is just one recent example: https://www.infoq.com/articles/api-gateway-clojure-golang https://www.infoq.com/articles/api-gateway-clojure-golang The money quote: "The end result enabled us to reduce 25 instances (c4 xlarge) running Clojure code - able to process 60 concurrent requests, to two instances (c3.2xlarge) running Go code able to support ~5000 concurrent requests a minute" If you google around you'll find more stories like that. To do better than Go you would have to drop to C++ or Rust.

Scarbutt 8y ago
"The end result enabled us to reduce 25 instances (c4 xlarge) running Clojure code - able to process 60 concurrent requests, to two instances (c3.2xlarge) running Go code able to support ~5000 concurrent requests a minute" That doesn't make any sense though.
karim 8y ago
"When people rewrite system from X where X in (Python, Ruby, Node, Clojure) to Go they usually see at least 10x improvement." Sure but how much of the performance gains come from Go and how much come from just having a better understanding of the problem the second time around?
- zepolen 8y ago
  Go is strongly typed and compiled, it literally does 10x less work than dynamic languages for most things. Pypy leads to a near 10x speed up over cPython as well.
kasey_junk 8y ago
Or Java or C# or basically any non-interpreted language.
- rakenodiax 8y ago
  In this case, Clojure is already running on the JVM
  - kasey_junk 8y ago
    The Clojure compiler does not generally produce the same byte code as the Java one. Idiomatic clojure is routinely slower than idiomatic Java and that’s a well known and expected outcome.
piranha 8y ago
That article is anything but useful. They have spent significantly more time designing and writing app in Go than their prototype in Clojure and now it’s faster? What a surprise!
apta 8y ago
> To do better than Go you would have to drop to C++ or Rust. C++ and Rust yes, but also Java and .NET.
- weberc2 8y ago
  In my experience Java, C#, and Go are roughly on par. Go is a bit easier to optimize though.
basic6 8y ago
I didn't think Go rewrites were common anymore since the results were mediocre.
fulafel 8y ago
Hmm, there's a slight disparity between the quote and their benchmark table, which shows Clojure and Go neck and neck in the request/sec dept. I guess all the system re-architecting explains most of the difference. In any case, as the article text also mentions, on the Go side they are using a complete reverse-proxy library that's included with the Go stdlib, which can be a significant advantage aside from the properties of the language itself. But then it seems they ended up reimplementing many things that the JVM provides: "In order to achieve out of the box functionality such as CPU and memory usage metrics, business logic counters and more - we needed to write basically this entire stack from scratch, which enabled us a much more rapid deep dive of the intricacies of Golang."