[Haskell-cafe] Re: Joels Time Leak
joelr1 at gmail.com
Tue Jan 3 16:42:24 EST 2006
The timeleak code is just a repro case. In real life I'm reading from
sockets as opposed to a file.
All I'm trying to do is run poker bots. They talk to the server and
play poker. Of course some events are more important than others, a
request to make a bet is more important than, say, a table update. I
do need to run as many poker bots as I can.
I think that my customer's goal of 4,000 bots is unattainable in a
single app. It's probably possible per machine. Overall, I find this
too complex to manage with Haskell as there are many factors that can
contribute to my delays and timeouts. There are also quite a few
unanswered questions at the moment (why is 'sequ' slow? does the
scheduler need to be tuned?) that leave me scratching my head.
On Jan 3, 2006, at 9:17 PM, S. Alexander Jacobson wrote:
> You should be grouping incoming events into queues by expected
> workload/event. Then you can give the client fairly reliable
> information about how long it will have to wait based on the size
> of the queue on which event is waiting.
> And if you have no way to differentiate between event workloads a
> priori then you really can't be giving clients response guarantees
> and need to rethink your business logic.
More information about the Haskell-Cafe