[Haskell-cafe] binary IO

[Branimir Maksimovic]

>From: Joel Reymont <joelr1 at gmail.com>
>To: Tomasz Zielonka <tomasz.zielonka at gmail.com>
>CC: Jeremy Shaw <Jeremy.Shaw at linspireinc.com>,haskell-cafe at haskell.org
>Subject: Re: [Haskell-cafe] binary IO
>Date: Tue, 27 Dec 2005 09:18:54 +0000
>
>Tomasz,
>
>Try http://wagerlabs.com/timeleak.tgz. See the "Killer pickler  
>combinators" thread as well.
>
>My desired goal is to have 4k bots (threads?) running at the same  time. 
>At, say, 1k/s per bot I figure something like 4Mb/s round-trip.  Each bot 
>cannot spend more than a couple of seconds on pickling/ unpickling. I'm not 
>even close to hitting that goal even reading from  a file.
>
>I'm getting delays of 4s+ with just 100-200 bots reading from a file  and 
>even less than that in a networked environment. The more bots I  run the 
>higher the delays, to the point of frequent delays of 10s+.  The kicker is 
>that some packets come in at 4k compressed with Zlib  but become something 
>like 50k uncompressed and then expand to a list  of 500+ records, etc.
>

I have C++ concurrent server that performs 2600 reqs/sec on about 500
connections and dual Xeon 2.8Ghz, but no pickling /unpickling, just short 
text.
Has sepparate IO threads that divide descriptor sets (num descs / IO thread)
and worker threads as number of CPU's * 2, no locking of shared queue.
So with 4k connections I guess that would be maximum 2k requests on *dual* 
box
per second, without pickling / unpickling, just short textual protocol and
simple services.
I think that you will get hard time even with C to achieve your goal.

Greetings, Bane.

_________________________________________________________________
Don't just search. Find. Check out the new MSN Search! 
http://search.msn.com/