> On Fri, Jun 11, 2010 at 11:22 PM, nonlinear5 <
eugene.kono...@gmail.com>wrote:
>
> > We should use the same benchmark for meaningful comparison. Let's
> > settle on this optimization job as a performance benchmark:
>
> > 1. JBT version 7.05
> > 2. Data file:
> >
http://code.google.com/p/jbooktrader/downloads/detail?name=ES-GLOBEX.zip
> > 3. Strategy: Equalizer
> > 4. Optimization options: brute force as search method, PI as selection
> > criteria, 300 minimum trades
>
> > This will run 29,400 strategies using the data file with 2,037,873
> > data points. For verification, the top parameters returned by
> > optimizer are 34-3900-14, net profit is $6757 on 311 trades.
>
> > My results on a 64-bit Vista, i7-920 processor, 6Gb RAM: 688 seconds.
> > This can also be expressed as a raw speed of 87,083,526 lines per
> > second:
> > (29,400 * 2,037,873) / 688 = 87,083,526
>
> > Just so that we are not overwhelmed with numbers, just report your OS,
> > processor, RAM size, and the number of seconds to complete the
> > benchmark.
>
> > --
> > You received this message because you are subscribed to the Google Groups
> > "JBookTrader" group.
> > To post to this group, send email to
jbook...@googlegroups.com.
> > To unsubscribe from this group, send email to