I think some people are knocking these benchmarks for the wrong reason.
I’m a .NET dev primarily. And .net has historically been a terrible performer on these benchmarks. With .net core the asp.net team and community has been working on performance.
While the raw plain text benchmark is not really realistic of a real world scenario that we would use on a day to day basis. It is an indicator of the performance baseline that .net core can achieve before you begin adding all the fluff on top.
This coupled with sites like stack overflow showing the gains they get using asp.net and moving from 1 version to the next gives me confidence that .net is not a bad choice these days.
Prior to .net core I was more or less ready to go switch full time to something else. But now I’m happy using .net.
Agree, that MS is working on performance and can show it on TechEmpower benchmarks is great. I also think some top performers are using special ways to be that fast but if you move down a bit in the list there are probably more realistic implementations.
Always look at the source. Some of them are very representative of high-performance C#, but others cheat a little. E.g. regex-redux just calls out to PCRE, although .NET 5's Regex would easily be competitive with Java as well.
I’m a .NET dev primarily. And .net has historically been a terrible performer on these benchmarks. With .net core the asp.net team and community has been working on performance.
While the raw plain text benchmark is not really realistic of a real world scenario that we would use on a day to day basis. It is an indicator of the performance baseline that .net core can achieve before you begin adding all the fluff on top.
This coupled with sites like stack overflow showing the gains they get using asp.net and moving from 1 version to the next gives me confidence that .net is not a bad choice these days.
Prior to .net core I was more or less ready to go switch full time to something else. But now I’m happy using .net.
/my 2c