I wonder if some of the implementations support keep alive by default while others don’t, and if that for instance could explain such a big difference between Nim and the other ones.
It seems a bit silly to not look into why each of the resulting numbers aren’t an order of magnitude higher, and work out what the limiting factor is in each case.
Someone like lemire shouldn’t be running “hello world” benchmarks. They are dumb and not productive. It is basically a low-effort shitpost to stoke an arm chair perf war.
Even root causing the differences is dumb. Do something useful in your benchmarks! This is a nothing burger to drive traffic to his site. Booo!
As soon as you add ANY AMOUNT OF LOGIC, the shape of perf changes. This shit belongs on the middle school internet.
If you look at techempower benchmarks[0], not really. Add a database call and it becomes a test of how well the database library uses asynchronous IO operations (I’m guessing).
I wonder if some of the implementations support keep alive by default while others don’t, and if that for instance could explain such a big difference between Nim and the other ones.
It seems a bit silly to not look into why each of the resulting numbers aren’t an order of magnitude higher, and work out what the limiting factor is in each case.
Benchmarks like this are useless. The numbers don’t mean anything.
Someone like lemire shouldn’t be running “hello world” benchmarks. They are dumb and not productive. It is basically a low-effort shitpost to stoke an arm chair perf war.
Even root causing the differences is dumb. Do something useful in your benchmarks! This is a nothing burger to drive traffic to his site. Booo!
As soon as you add ANY AMOUNT OF LOGIC, the shape of perf changes. This shit belongs on the middle school internet.
Yup. Add a database call and they all balance out to the same number of requests / second.
Talk to me about the devops side of the frameworks. Or security. Or developer experience. I find that much more interesting.
If you look at techempower benchmarks[0], not really. Add a database call and it becomes a test of how well the database library uses asynchronous IO operations (I’m guessing).
[0] https://www.techempower.com/benchmarks/#section=data-r21&test=fortune
why go1.19, when 1.21 is available and 1.19 isn’t supported anymore?
And why not fasthttp if they’re going to put their finger on the scales with httpbeast?
fasthttp isn’t spec-compliant and doesn’t support HTTP/2. Is httpbeast doing the same kind of shortcuts?
httpbeast seems to only implement HTTP 1.1.
From their GitHub page: