Tomasz, I'm sure you've answered this before but I couldn't find it. How do you best determine AB internal cache size? I see mine is set at 409MB. Other database settings are 10000 symbols (even though I never have over 200-300 in my database), number of bars to load is 1000, limiting number of saved quotations isn't checked.
I'm running an Athlon 3200+ processor, Windows XP SP2, and AB 5.09.0. I have 2GB and rarely use much over half of it. I'm mostly interested in speed while optimizing formula parameters with Quotes Plus EOD data using the Intelligent Optimizer (IO). I have one long run on the weekends that I'd be (very) interested in speeding up Thanks, Dan --- In [email protected], "Tomasz Janeczko" <[EMAIL PROTECTED]> wrote: > > Hello, > > As for 1) you would do better by increasing AB internal cache in preferences. > > Best regards, > Tomasz Janeczko > amibroker.com > ----- Original Message ----- > From: "akerstkh" <[EMAIL PROTECTED]> > To: <[email protected]> > Sent: Wednesday, May 14, 2008 6:09 PM > Subject: [amibroker] Re: Dual-core vs. quad-core > > > > Good morning, > > > > Just to share some experience in Optimizing the Speed... Mainly in > > compiling of data with the ATC instruction with loops... > > > > 1) Created a RAM Disk of 1G. Copy the Data Folder to the RAM Disk > > and then use the Data in the RAM Disk do the compiling. (Only worth > > doing if compile a couple of times a day.. It cut down from 16 mins > > to 11 mins. ) > > > > 2) MOVE the Data Folder to another physical drive. Then perform the > > disk defragmenter on the Amibroker Drive a couple of times. Then > > MOVE the Data Folder back to the original position. > > > > 3) Chop Down the number of bars in Data Base. > > > > > > > > > > The Current system is Intel Core 2 @2.13GHz with 4G RAM. > > > > Yet to try is to upgrade the system with RAID4 with 4 Disks and > > Q9450. > > > > Is there any one can comment on the experience with RAID5 system? > > > > > > Best Regards > > KH Tang > > > > > > --- In [email protected], Dennis Brown <see3d@> wrote: > >> > >> It also makes me wonder if there is another possibility for speed > >> optimization. I run more than 1 RT chart at the same time to get > >> multi-timeframe results that are not possible in a single chart - - > > then > >> pass results between the charts. It seems as though parallel > >> processing would be possible between multiple "live" indicator > > mode > >> charts. Maybe I will have to get the 8 core system next time :) > >> > >> Best regards, > >> Dennis > >> > >> On May 14, 2008, at 10:13 AM, dloyer123 wrote: > >> > >> > Interesting result. It is reasonable to expect newer and future > >> > processors to have even larger caches. 6MB on the E8400 > >> > > >> > So, there may be an opportunity to get useful work out of the > > other > >> > cores during long optimization runs. Maybe large "work units" > > such > >> > as symbols lists or entire optimization passes to minimize sync > >> > overhead? > >> > > >> > That also implies that user formula code could be optimized if we > >> > group array access together, ie use a result right after it is > >> > created, so it can be pulled out of cache, rather than wait until > >> > later in the code. I will have to try this... > >> > > >> > > >> > > >> > --- In [email protected], "Tomasz Janeczko" <groups@> > >> > wrote: > >> >> > >> >> Hello, > >> >> > >> >> I just run the same code on my relatively new notebook (Core 2 > > Duo > >> > 2GHz (T7250)) > >> >> and the loop takes less than 2ns per iteration (3x speedup). So > > it > >> > looks like the data sits entirely inside the cache. > >> >> This core 2 has 2MB of cache and thats 4 times more than on > > Athlon > >> > x2 I got. > >> >> > >> >>> If what you say is true, and one core alone fills the memory > >> >>> bandwidth, then there should be a net loss of performance while > >> >>> running two copies of ami. > >> >> > >> >> It depends on complexity of the formula and the amount of data > > per > >> > symbol > >> >> you are using. As each array element has 4 bytes, to fill 4 MB of > >> > cache > >> >> you would need 1 million array elements or 100 arrays each having > >> > 10000 elements > >> >> or 10 arrays each having 100K elements. Generally speaking people > >> > testing > >> >> on EOD data where 10 years is just 2600 bars should see speed up. > >> >> People using very very long intraday data sets may see > > degradation, > >> > but > >> >> rather unnoticeable. > >> >> > >> >> Best regards, > >> >> Tomasz Janeczko > >> >> amibroker.com > >> > > >> > > >> > ------------------------------------ > >> > > >> > Please note that this group is for discussion between users only. > >> > > >> > To get support from AmiBroker please send an e-mail directly to > >> > SUPPORT {at} amibroker.com > >> > > >> > For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG: > >> > http://www.amibroker.com/devlog/ > >> > > >> > For other support material please check also: > >> > http://www.amibroker.com/support.html > >> > Yahoo! Groups Links > >> > > >> > > >> > > >> > > > > > > > > ------------------------------------ > > > > Please note that this group is for discussion between users only. > > > > To get support from AmiBroker please send an e-mail directly to > > SUPPORT {at} amibroker.com > > > > For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG: > > http://www.amibroker.com/devlog/ > > > > For other support material please check also: > > http://www.amibroker.com/support.html > > Yahoo! Groups Links > > > > > > >
