I would probably try something like:
1. To avoid unparsing/parsing which just add/remove a lot of redundancy:
use byte-encoding of that info(i don't know what kind of info You have,
so I cannot say how, probably just to blit struct's that You have would
be enough). In other words: use fd456 instead of 2815000.
2. If (and only if) that doesn't help - try to integrate both programs
using tool such as HaskellDirect
Hope it helps
Michal Gajda
[EMAIL PROTECTED]
On 31 Dec 1999, Matthias Riese wrote:
> Hi!
>
> My system is made up of a non-haskell program producing data and a
> haskell programm consuming the data. Currently the non-haskell program
> writes out the data using a Haskell "data" type format. The haskell
> program uses "deriving Read" to read in the data.
>
> Now the data flow shows up as a critical performance bottleneck: It
> takes much much more time to read in the data than to process it. So
> I'd like to hear suggestions about how to speed it up. Any ideas?
>
> TIA, Best Regards,
> Matthias Riese
>
