Closed natalie-o-perret closed 4 years ago
Hi @kerry-perret,
It's much better to have such examples as a github repo which compiles and works ;)
Anyway, without a launch, I could advice two things:
I don't see reason except (2) explaining why code behaves differently.
It's much better to have such examples as a github repo which compiles and works ;)
Sure, will create a .NET Core solution for that along with the dockerfile.
It's much better to have such examples as a github repo which compiles and works ;)
Sure, will create a .NET Core solution for that along with the dockerfile.
Did you try to tune thread pool settings?
Not yet, actually I plan to have everything in that solution first, there is a huge lack of resources about performances when it comes to code with .NET.
I mean there are like general guidelines and some part of the RabbitMQ tutorial but that's pretty much about it:
Don't worry I will also check out the thread starvation.
@Pliner If you want, you can check this out: https://github.com/kerry-perret/CSharp.RabbitMQ.Benchmarks
@kerry-perret Have you had a chance to adjust thread pool and benchmark again?
@Pliner sorry I didnt have much time lately, I had 6 medical appointment in a row + work.
Will try to double check things this coming weekend or next week.
Going thru a hectic schedule at the moment: lawyer + tax residency issues + incoming surgeries + daily job + OTing.
I haven't fogotten that issue, will get back to this asap.
Feel free to reopen when you will have results.
I wanted to code my own publishing method from scratch with the official RabbitMQ .NET client, but for some reason I can't achieve the same performances as with EasyNetQ.
I was wondering if the code in
TestSameChannelSeqAsync
is roughly equivalent to the code inTestEasyNetQAsync
with the given extension methods:
and how it is called:
The data in "File.xml":
It turns out that my version coded from scratch takes about 2 min while EasyNetQ takes only about 40 seconds.
I'm not really sure to understand why, since I've basically followed the path taken in the EasyNetQ.
I just would like to know why EasyNetQ is blazing fast in that particular case?