What is ur datasource. I find DTNIQ to be causing heavy cpu cycles.
the
same scans that i was running last week is now slowing my system.
if i
shut down the connector and run it connected to a different
datasource
like IB, it is normal.
If anyother are having same issues with DTNIQ, i
would appreciate
response because i am going to complain to them
now.
For the last 2 days, it is so bad that i cannot use the DTN data
source at all. Maybe the 75 dollar savings compared to esignal is not
worth it.
In terms AB optimizations,
1. i run 2 scans, i run a
EOD scan to filter out potential candidates
adn put them in a watchlist.
then i run a realtime scan on the
watchlist to get me my results.
obviously, loops would add to the
execution times.
2. i have noticed
that repeat scans on the same formula returns data
quicker. i guess, the
first time, it has to backfill lot more. so i
start running my realtime
scans before 9:15 am so that by the time
the market opens, the scans are
runnign faster.
Finally, this week DTNIQ is a resource hog. really big
time.
--- In amibroker@xxxxxxxxxps.com,
"Padhu" <ccie8340@xx.> wrote:
>
> Hello,
>
> I am running Amibroker ver 4.9 on Dell single 2GHZ CPU with 768mb
ram.
> Just checking to see if anyone can share some
> Best
practices on tuning up Amibroker for faster scanning.
>
>
>
> 1. What are the things that can be be done on the system to
improve
scanning time?. My scan interval is set for a min, but sometimes
elapsed time is about 90 seconds.
>
> 2. Any recommendations
on CPU/Mem usage?.
>
> 3. Anything in AFL that is known to be a
resource hog and should
not be used or coded better shall we say?.
>
> 4. If scanning is done every min, do I still need backfill?. Isn't
backfill required only when previous bars data is missing in the
database?.
>
>
>
> Thanks.
>
>
>
> Cheers,Padhu
>