Hacker News Reader

Comment by megapoliss | Hacker News Reader

Run some examples, and it looks like this "High-performance, real-time optimized, super-fast" language is

  ~ 10 times slower than luajit
  ~ 3 times slower than lua 5.4

View on HN

Not bad for version 0.1.0. lua(jit) is no slowpoke and has had decades of performance improvements.

With this and https://github.com/Beariish/bolt/blob/main/doc/Bolt%20Perfor..., it is indeed confusing without testing it out myself.

That said, somehow I do not believe it is faster than LuaJIT. We will see.

I used brainfuck interpreter https://github.com/Beariish/bolt/blob/main/examples/bf.bolt vs lua / luajit implementations from https://github.com/kostya/benchmarks

Just checked with nbody:

  - still 10 times slower than luajit
  - 2 times slower than luajit -joff
  - but 20% faster than lua 5.4
  - but uses 47 Mb RAM vs 2.5 Mb for lua/luajit

I appreciate the followup here. The brainfuck interpreter isn't meant to be a benchmark notably, it's a naive implementation for the sake of the example.

I did spot some poor code in the Bolt version of nbody that can be changed (the usage of `.each()` in the hot loop is creating loads of temporary iterators, that's the memory difference.)

luajit -joff does perform better even with this change, but I observe closer to 15% than a 2x difference

for nbody 500000 on my i5-9300H CPU @ 2.40GHz

  - 487.41 millis / 2364 kb ram for luajit -joff
  - 770.17 millis / 41712 kb ram for bolt

  770.17 / 487.41 ~~ 1.58 cpu, not 2x, but not 15% either
  41712 / 2364 ~~ 17.64 ram

Where do you get these numbers from? Looking at https://github.com/Beariish/bolt/blob/main/doc/Bolt%20Perfor... doesn’t seem to support these numbers.

They at least clarified it by saying "outperforming other languages in its class". It's a slow class so the bar is low.

Reply on news.ycombinator.com