No, your tests shows that TC2000 itself accounts for 90% of run time (5 minutes with TC compared to less than half minute without link to TC).
And yes if you are using TC2000 you can switch back and forth - select TC2000 as data source, run one AA over all symbols to copy data to AB local database, then switch data source to (local database) to avoid querying TC2000 for new data - this gives 10x speedup because slow TC2000 is not contacted anymore. Best regards, Tomasz Janeczko amibroker.com ----- Original Message ----- From: "polomorabe" <[email protected]> To: <[email protected]> Sent: Sunday, September 13, 2009 8:28 AM Subject: [amibroker] Re: Backtest execution speed. > Thanks Bruce, Tomasz and Keith. > > Bruce, I checked with the simple AFL you suggested Buy=Sell=Short=Cover=0, > and my results were: > - First run after starting AB: 5'30" > - Subsequent runs: 2'10" > - Using local database: 0'23" > So indeed, the AFL itself accounts for most of the run-time. Thanks for the > tip. > > Keith, until he confirms himself, I believe Tomasz confirmed my > understanding, ie, that with the source set to "local database", > no accesses from the TC2000 database are made. That is, after starting AB, > the source must be set to TC2000 to ensure the local > database is updated. Afterwards, within the same AB session, the source can > be set "local database" to maximise performance. > > Paul > > > > ------------------------------------ > > **** IMPORTANT PLEASE READ **** > This group is for the discussion between users only. > This is *NOT* technical support channel. > > TO GET TECHNICAL SUPPORT send an e-mail directly to > SUPPORT {at} amibroker.com > > TO SUBMIT SUGGESTIONS please use FEEDBACK CENTER at > http://www.amibroker.com/feedback/ > (submissions sent via other channels won't be considered) > > For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG: > http://www.amibroker.com/devlog/ > > Yahoo! Groups Links > > >
