[erlang-questions] trading systems
Thu Oct 1 22:26:22 CEST 2009
> I'm assuming it's an Erlang-specific suggestion, right? http://www.kx.com
> to 20% faster with 64 bits on some benchmarks. There are more
> registers available for the compiler to use, etc. In short, I
> disagree since your suggestion is quite general and is not backed up
> by known use cases.
Actually, it's only mildly supported by Erlang, but heavily supported
in Python and Ruby. Unless you're generating machine code, I've never
seen 64-bit to be a big boost. While HIPE might make that a non-
issue, I've generally observed that VMs don't benefit from 64-bit in
> I assume you have not used Haskell for heavy-duty projects but heard
> of it and maybe tinkered with it. I used Haskell for a networking
> project I extensively wrote about (google for Erlang vs Haskell). I
> switched to Erlang in the end and felt much better. Haskell is lazy
> which makes performance unpredictable.
I've used it and loathed that lack of libraries. I was only pointing
out that it's type system is cleaner. Just trying to round out the
spectrum of options.
> Why? I interviewed with a hedge fund. They use templates extensively
> and it makes sense. Partial template specialization is just the
> ticket to optimize generic code at compile time.
Templates can drastically inflate code size and cause an extra
indirection at function call time. I've always found that, for
optimization, templates are middling, doing your own indirection in
straight C is king--especially when you give branching hints. Of
course, that might be premature optimization at this point. ;P
> Yes, thanks. I've been using tracing for years now.
I'll assume from your OpenPoker statement that you were doing tracing
with Erlang. I've found the documentation a bit lacking, so I like to
throw that example around to make sure that people can actually figure
out how to use it. Let's just say that dbg:tpl is not exactly an
> And faster time to market.
More information about the erlang-questions